apparatus to determine an estimate of the dimensions of a room, and, method to determine an estimate

专利摘要:
APPARATUS TO DETERMINE AN ESTIMATION OF THE DIMENSIONS OF A ROOM, AND, METHOD TO DETERMINE AN ESTIMATE OF THE DIMENSIONS OF A ROOM. An apparatus for determining an estimate of the dimensions of a room comprises a receiver (101) that provides an acoustic room response, for example, generated from acoustic measurements. A peak detector (103) detects a set of peaks in the room's acoustic response over a frequency range that has a maximum frequency of at most 400 Hz. A storage (107) comprises a set of peak profiles with dimension data of an associated room and an estimator (105) determines the estimation of the dimensions of a room from the dimension data of an associated room and a comparison of the set of peaks with the peak profiles. The estimator can perform the steps of first finding a compatible peak profile for the peak set of the peak profile set, extracting first room dimension data from storage associated with the compatible peak profile (s) (is), and determine the estimation of the dimensions of a room in response to the first dimension data of a room. Peak profiles can represent calculated Eigen frequencies. (...).
公开号:BR112015028335B1
申请号:R112015028335-7
申请日:2014-04-24
公开日:2021-02-17
发明作者:Werner Paulus Josephus De Bruijn
申请人:Koninklijke Philips N.V.；
IPC主号:

专利说明:

[0001] [001] The invention relates to the determination of an estimate of dimensions of a room and in particular, but not exclusively, to the determination of combined dimensions of length, width and height of a room. BACKGROUND OF THE INVENTION
[0002] [002] There are many applications in which it is beneficial to know the dimensions of a room. An obvious example is a sound reproduction application, in which the room has a very large influence on the sound that is experienced by the user. Knowledge of the shape and dimensions of a room provides important information that can be used to optimize sound reproduction for that particular room. For example, knowing the dimensions of a room allows you to predict important acoustic properties of the room, such as its low-frequency room (set of resonances) modes (which generate resonances at specific frequencies, resulting in an unpleasant "resounding" low sound), patterns of early reflection, reverberation time, etc. Knowing such acoustic properties allows you to process speaker signals in order to optimize the sound experience in the room. In addition, knowing the dimensions of the room allows you to provide specific recommendations for the user of a multiple speaker system on how to best configure the speaker system.
[0003] [003] In addition, apart from sound reproduction applications, there are many others that benefit from knowledge of the dimensions of a room, for example, any application in which knowledge about the user's context is used to optimize a user experience. .
[0004] [004] Although, of course, it is possible to manually measure the dimensions of a room and insert them into a device, this is uncomfortable and, often, not practical.
[0005] [005] There are visual methods that can provide some indication of the layout of a room. Typically, such methods are based on still or moving image cameras. However, although some information can be obtained through such approaches, they tend to be limited by the camera's viewing angle and are hampered by objects that block the camera's view as well as varying lighting conditions. In addition, these approaches often require additional or dedicated equipment (such as the camera) and may require specific placement of the camera, which can be inconvenient.
[0006] [006] Another possibility for estimating at least partially automated dimensions of a room is to determine estimates based on acoustic measurements in the room. This can be particularly advantageous for sound rendering applications where the audio rendering system can also comprise functionality for estimating the dimensions of the room.
[0007] [007] Several methods for estimating the acoustic dimension of a room are known, however, they tend to be sub-ideal and, in particular, tend to be uncomfortable, complex and / or inaccurate. For example, acoustic methods are known that can generate an estimate of room volume by measuring the reverberation time. However, this only results in an approximate indication of the overall size of the room (for example, small, medium, large) and does not provide estimates of the individual dimensions.
[0008] [008] For this reason, an improved approach to determining a room size would be advantageous and, in particular, an approach that would allow for greater flexibility, greater ease of operation, less complexity, less resource consumption, improved estimation accuracy would be advantageous. and / or improved performance. SUMMARY OF THE INVENTION
[0009] Consequently, the invention preferably seeks to mitigate, alleviate or eliminate one or more of the disadvantages mentioned above, individually or in any combination.
[0010] [010] In accordance with an aspect of the invention, an apparatus is provided to determine an estimate of the dimensions of a room, wherein the apparatus comprises: a receiver for providing an acoustic room response, a peak detector for detecting a set of spikes in room acoustic response in a frequency range, where the frequency range has a frequency greater than 400 Hz maximum, a storage to store a set of peak profiles with associated room dimension data, an estimator for determine the estimation of the dimensions of a room in response to the associated room dimension data and a comparison of the set of peaks with the set of peak profiles.
[0011] [011] The invention can allow an improved and / or facilitated determination of an estimate of the dimensions of a room. In particular, the approach can, in many ways, provide a more accurate estimate of the dimensions of a room and / or more detailed data. Specifically, the approach can differentiate between room dimension data in different directions and is not limited, for example, to providing an estimate of the room's size or volume. In particular, the approach allows the determination of individual estimates of one-dimensional distance and, in particular, can characterize a room by means of a plurality of such individual estimates of one-dimensional distance. For example, length and width or length, width and height can be estimated individually for a room. In this way, a substantial improvement in the characterization of a room can be achieved.
[0012] [012] The approach can allow for an estimate of a low complexity room and does not depend on complex or precise formulas to calculate the characteristics of the room from the responses of the room. The approach can automatically consider variations in room characteristics and is typically not sensitive to such acoustic variations. For example, the approach can be used in rooms with different reverb characteristics and can adapt automatically without requiring any input or knowledge of specific features, except for the room's acoustic response.
[0013] [013] The estimated room dimensions generated can be used to adapt the signal processing of an audio rendering system. The approach can therefore result in improved audio quality of the rendered sound.
[0014] [014] The inventor noted that a room's dimension can be estimated from peak characteristics in a low frequency range of a room's acoustic response. In particular, the inventor realized that the peak characteristics at low frequencies reflect specific frequencies (that is, characteristics or natural, also known as "Eigen") of the room and that they are indicative of the dimensions of a room. The specific approach can exploit such observations to provide an improved estimate of the dimensions of a room.
[0015] [015] The set of peak profiles can comprise a potentially large number of peak profiles corresponding to the acoustic responses of rooms with different dimensions and / or characteristics. Each peak profile can provide an indication of a peak distribution in a room's acoustic response. The peak profiles may reflect characteristics of the acoustic responses measured in the room. Peak profiles can theoretically reflect derived values, for example, based on Eigen frequency calculations for rooms with certain dimensions. Peak profiles can be limited to a low frequency range.
[0016] [016] The storage / memory can store a set of room-sized data for each peak profile. The dimension data of a room for a given peak profile can comprise at least one dimension value for the room corresponding to the room's acoustic response represented by the peak profile. Specifically, storage may comprise a plurality of room data sets in which each data set corresponds to a different room. Each room data set comprises an associated peak profile and room dimension data. The peak profile may reflect a peak distribution of an acoustic response in such a room. The associated room dimension data can comprise a set of one-dimensional distance values for the room. Specifically, the associated room dimension data can comprise a set of one-dimensional values for different directions (for example, width, length and height) of the room.
[0017] [017] The peak set can be compared with each peak profile in the peak profile set. The comparison may reflect how closely the peak distribution of the set of peak profiles corresponds to the peak distribution indicated by the peak profile.
[0018] [018] In some modalities, the frequency range can advantageously have a higher frequency of a maximum of 300 Hz, 200 Hz or even 100 Hz.
[0019] [019] According to an optional feature of the invention, the estimator comprises: a search engine to find at least one compatible peak profile of the peak set from the peak profile set, an extractor to extract the first dimension data from storage of a room associated with at least one compatible peak profile, and an estimation generator to determine the estimation of the dimensions of a room in response to the first dimension data of a room.
[0020] [020] This can provide an improved estimate in many scenarios and / or can allow for easier operation and implementation. In particular, this may allow an algorithm of relatively low complexity to be used to generate accurate estimates of dimensions. The compatible peak profile (or profiles) can be the profile that is considered the most compatible with the peak set. In some embodiments, the device can identify a plurality of compatible profiles and the dimension data of a room associated with such a plurality of compatible profiles can be used to generate an estimate of the dimensions of a room.
[0021] [021] According to an optional feature of the invention, each peak profile comprises a set of peak frequencies.
[0022] [022] This can provide high performance and / or easier implementation and operation. In particular, this can allow for an efficient representation of peak distribution information that is highly suitable for processing to generate room estimates.
[0023] [023] According to an optional feature of the invention, the searcher is willing to determine a distance measurement for each peak profile in response to a comparison of the peak set frequencies and the peak profile frequency set, and for select the compatible peak profile in response to distance measurements.
[0024] [024] This can provide an efficient operation and / or provide more accurate estimates. In particular, this can allow for a comparison and identification of one or more compatible profiles that are particularly suitable for room estimation. The distance measure can be, for example, an accumulated value of the differences between the individual peaks in the peak set and the peak closest to the peak profile.
[0025] [025] According to an optional feature of the invention, each peak profile comprises a set of probability values, where each probability value indicates a probability that a peak will be measured over a frequency range.
[0026] [026] This can provide high performance and / or easier implementation and operation. In particular, this can allow for an efficient representation of peak distribution information that is highly suitable for processing to estimate the dimensions of a room.
[0027] [027] According to an optional feature of the invention, the searcher is willing to determine a probability measure for each peak profile in response to the probability values of the peak profile and peak set frequencies, and to select the profile compatible peak in response to probability measures.
[0028] [028] This can provide an efficient operation and / or provide accurate estimates. In particular, this can allow a comparison and identification of one or more compatible profiles, which is particularly suitable for estimating the dimensions of a room. The probability measure can be, for example, an accumulated value of the probability values for the individual peaks in the peak set.
[0029] [029] According to an optional feature of the invention, each peak profile comprises a set of Eigen frequencies calculated for a room and the associated room dimension data comprises an indication of dimensions for the room used to calculate Eigen frequencies.
[0030] [030] This can provide effective room estimation dimensions and can, in many scenarios, facilitate implementation. For example, this can avoid or reduce the need for uncomfortable and resource-intensive measurements and data collection in order to popularize storage. In many scenarios, improved results can be obtained as noise and errors associated with measured room acoustic responses can be avoided for stored peak profiles.
[0031] [031] According to an optional feature of the invention, the associated room dimension, for at least part of the peak profiles, comprises at least one-dimensional value and the estimator is willing to generate the estimate of the dimensions of a room to understand at least one-dimensional value.
[0032] [032] The approach is not restricted to an estimate of overall size or volume, as is typical for many prior art approaches. Instead, individual one-dimensional measures can be generated. In some embodiments, estimating the dimensions of a room may, for example, comprise one, two or three one-dimensional values in length. For example, estimating the dimensions of a room can comprise individual estimates of length and width or estimates of length, width and height.
[0033] [033] According to an optional feature of the invention, the estimator is willing to generate a compatibility indication for the peak set and each peak profile in the profile set, and to generate the dimension estimate through a weighted combination of the associated room dimension data, where the weight of the associated room dimension data of a first peak profile in the set of peak profiles depends on the compatibility indication of the first peak profile.
[0034] [034] The feature can provide an improved estimate in many scenarios. In particular, the resource can provide more accurate estimates in many scenarios by providing an average or more flexible combination of contributions of the characteristics associated with different candidate rooms. In particular, the approach can reduce noise sensitivity.
[0035] [035] According to an optional feature of the invention, the associated room dimension, for at least part of the peak profiles, comprises a plurality of one-dimensional values corresponding to different directions, and the estimator is willing to determine indications of average compatibility for a subset of directions in response to a calculation of the average of compatibility indications for peak profiles in at least one direction that is not in the subset, and to determine unidirectional dimensions estimates of a room for the subset directions in response to indications of medium compatibility.
[0036] [036] The feature can provide an improved estimate in many scenarios.
[0037] [037] According to an optional feature of the invention, the receiver is arranged to receive a plurality of room acoustic responses, which correspond to different positions of at least one of a sound source and a microphone, and the apparatus is arranged to perform a combination for the plurality of room acoustic responses.
[0038] [038] This may allow more accurate estimates to be provided. In particular, the approach can provide additional data that allows for an improved estimate. In particular, the approach may allow for the measurement and detection of a greater number of Eigen frequencies in the room.
[0039] [039] In some modalities, the calculation of the average can be a calculation of the average (low-pass filtering) of room acoustic responses for different positions.
[0040] [040] According to an optional feature of the invention, the combination includes at least one of the calculation of the mean of the comparisons between sets of peaks corresponding to different room acoustic responses and the set of peak profiles and a calculation of the mean of estimates room dimensions determined for different room acoustic responses.
[0041] [041] This can provide improved estimates in many scenarios.
[0042] [042] According to an optional feature of the invention, the combination comprises generating the set of peaks by including peaks from more than one of the plurality of room acoustic responses.
[0043] [043] This can provide improved estimates in many scenarios.
[0044] [044] According to an optional feature of the invention, the estimator may be willing to select a subset of the set of peak profiles for comparison in response to a user input that indicates a room dimension value.
[0045] [045] This can improve the dimensions of room estimation and / or ease of operation.
[0046] [046] According to an optional feature of the invention, the estimator is willing to weight at least one of different peaks in the set of peaks and different peaks of a peak profile differently when performing the comparison.
[0047] [047] This can improve the room estimation dimensions and / or ease of operation.
[0048] [048] In accordance with one aspect of the invention, a method is provided to determine an estimate of the dimensions of a room, the method comprising: providing a room acoustic response, detecting a set of peaks in the room acoustic response in a room frequency range, where the frequency range has a maximum frequency of at most 400 Hz, provide a set of peak profiles with associated room dimension data, and determine the estimation of the dimensions of a room in response to data from dimension of an associated room and a comparison of the set of peaks with the set of peak profiles.
[0049] [049] These and other aspects, features and advantages of the invention will be evident from and elucidated with reference to the modality (s) described later in this document. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] [050] The modalities of the invention will be described, by way of example only, with reference to the drawings, in which:
[0051] [051] Figure 1 illustrates an example of elements of an apparatus to determine an estimate of the dimensions of a room according to some modalities of the invention;
[0052] [052] Figure 2 illustrates an example of elements of an estimator for an appliance to determine an estimate of the dimensions of a room according to some modalities of the invention;
[0053] [053] Figure 3 illustrates an example of Eigen frequencies and room acoustic response peaks; and
[0054] [054] Figure 4 illustrates an example of Eigen frequencies and room acoustic response peaks. DETAILED DESCRIPTION OF SOME MODALITIES OF THE INVENTION
[0055] [055] Figure 1 illustrates an example of an apparatus for determining an estimate of the dimensions of a room such as, specifically, a set of one-dimensional room dimensions, for example, a room's width, length and height.
[0056] [056] The device estimates the dimensions of a room based on a low frequency acoustic response measured from the room in one or more positions. The device detects peak frequencies in the measured room acoustic response (s) and compares the detected peak frequencies with a set of peak profiles stored in a database in which each peak profile corresponds to the acoustic response of a low frequency room for a given room. Thus, each peak profile is associated with a set of dimensions of a room corresponding to the room for which the peak profile was determined. The system generates estimates for the room in question based on comparisons and associated room dimension data. For example, the peak profile that is most closely compatible with the set of detected peaks is identified, and the dimension data of a room stored for that identified peak profile is extracted and used as dimension estimates of the room in question. Thus, it can be considered that the dimensions of the room in question correspond to the dimensions stored for the peak profile that is most closely compatible with the detected peaks.
[0057] [057] The approach explores the fact that the sound field in any room is accumulated from a discrete set of Eigen modes (normal modes), which are the solutions to the acoustic wave equation for that room and that they can be used to estimate room dimensions. Each of the Eigen modes has a corresponding Eigen frequency (also often called a modal or natural frequency). For a rectangular room with rigid walls, such Eigen frequencies can be provided by: (see, for example, Heinrich Kuttruff, Room Acoustics (Second Edition), Applied Science Publishers, 1979, ISBN-10: 0853348138, ISBN-13: 9780853348139:
[0058] [058] The total room acoustic response can be modeled using the well-known "modal decomposition model", which calculates the total response p at a receiver position ⃗r and frequency ω, for a given position of the sound source ⃗r 0:
[0059] [059] For a rectangular room, the functions of Eigen pn in equation (2) are provided by:
[0060] [060] In a typical room there are a multitude of Eigen modes and, in fact, individual Eigen modes are, in general, considered inseparable and efficiently form a "continuum", in which individual Eigen modes cannot be distinguished in the room frequency response. However, the current approach exploits the fact that at low frequencies, typically, the Eigen frequencies are sufficiently separated to allow the identification of individual Eigen frequencies. The approach further explores the fact that the characteristics of such individual Eigen frequencies can be used as a fingerprint or signature for a room response and, therefore, the room in question, and that they can be used to identify rooms with similar room dimensions. In addition, the system can exploit the fact that identification can be used to provide individual one-dimensional (length) estimates for different directions in the room. Specifically, individual estimates for room length, width and height can be generated based on the peak signature for the room. Thus, in contrast to conventional approaches, which typically can only provide estimates of the overall size or volume of the room, the current approach can provide estimates of individual one-dimensional characteristics.
[0061] [061] In fact, the approach explores the fact that each room size (specifically in terms of length, width and height) corresponds to a more or less exclusive set of Eigen frequencies. Two rooms that share one or two dimensions (for example, have the same length and / or height) will have some Eigen frequencies in common in their respective sets, however, they will also have some that are different.
[0062] [062] The apparatus in Figure 1 comprises a receiver 101 that receives / provides a room acoustic response. The room acoustic response is a measured room acoustic response and can be, specifically, a room acoustic response measured by the apparatus in Figure 1. The room acoustic response can specifically represent the frequency domain transfer function between a speaker that emits an acoustic test signal and a microphone that receives an acoustic test signal in a room. This frequency domain transfer function can be obtained directly from measuring the amplitude at the microphone position to the individual narrowband test signals, or from measuring a time domain pulse response and transforming it for the frequency domain, or can be obtained in any other way known to one skilled in the art.
[0063] [063] It will be understood that many different approaches and techniques are known for determining a room's acoustic response including, for example, logarithmic scans, MLS signals, etc. In some embodiments, the receiver 101 may include an audio output to provide a test audio signal to an external speaker and a microphone input to receive a microphone signal from an external microphone. The receiver 101 can therefore generate and output a test signal and analyze the resulting microphone signal to generate a room acoustic response according to any suitable approach.
[0064] [064] In other modalities, the room acoustic response can, for example, be provided from an external unit or retrieved from a remote or local storage.
[0065] [065] The receiver 101 is coupled to a peak detector 103 that is willing to detect a set of peaks in the room's acoustic response in a frequency range that has a maximum frequency of at most 400 Hz and in many modalities at most 300 Hz, 200 Hz or, in some modalities, even 100 Hz.
[0066] [066] It will be understood that many different peak detection techniques and algorithms are known and that any suitable technique can be used without departing from the invention. For example, the peak detector can first apply low-pass filtration and then detect all local highs below 400 Hz. The peak set can then be generated as the frequencies of such local highs. In other examples, a fixed number of peak frequencies can be selected, such as the frequencies of the largest, for example, ten or twenty peaks.
[0067] [067] The set of peaks, therefore, provides an application of parameters of characteristics of the room's acoustic response and, therefore, the room and can be seen as a signature or fingerprint of the room's acoustic response. However, the peak set is not just an application of room acoustic response parameters, but instead is specifically selected to provide a strong correlation with the room's Eigen modes and Eigen frequencies and is therefore selected to have a particularly strong correlation with the physical dimensions of the room. For brevity and clarity, the set of peaks (peak frequencies) extracted from the room's acoustic response will be called the peak signature (from the room or room acoustic response).
[0068] [068] Peak detector 103 is coupled to an estimator 105 which is additionally coupled to a memory or storage 107 that stores a database of peak profiles and associated room dimension data.
[0069] [069] Specifically, the database comprises a plurality of data sets in which each data set comprises a peak profile that corresponds to a room acoustic response for a different room (real or virtual / calculated), along with dimensions of a room for that room. In this way, each data set comprises the associated peak profile and room dimension data. The peak profile provides a representation of peaks (and, specifically, peak frequencies) in a room response and the associated room dimension data provides an indication of the dimensions of that room. The associated room dimension data can, for example, indicate the length, width and height of the room that has an acoustic response represented by the peak profile. Thus, the peak profile can also be considered a signature or fingerprint of the acoustic response of a room.
[0070] [070] The database may, in some modalities, also include a representation of the room acoustic response, however, it will typically include only the parameterized representation provided by the peak profile. A specific advantage of the current approach is that only a peak profile needs to be stored, rather than the room's own acoustic response. This can substantially reduce storage requirements, as well as computational requirements.
[0071] [071] Typically, the database will include data sets for a wide range of different rooms. This can include data sets for a large number of rooms with different dimensions, but it also includes different data sets for rooms with the same dimensions. For example, data sets can be stored for rooms of the same dimension, but with different acoustic characteristics, such as for an empty echoic room, a room with the same dimensions, however, furnished and less echoic, etc.
[0072] [072] In addition to the peak profile, the data set for each room's acoustic response comprises room-size data for the room that corresponds to the room's acoustic response. The dimension data of a room can specifically comprise one-dimensional distances, such as the width, length and height of the room.
[0073] [073] In some modalities, the data set may include additional information for each room. For example, in some modalities, each data set can also contain additional acoustic information, such as an indication of how reverberating the room is (an indication that can, for example, be reflected in the amplitudes of the peaks). Such additional information can be used, for example, by an audio rendering system based on the generated estimates.
[0074] [074] Estimator 105 is willing to estimate the dimensions of a room based on a comparison of the peak set with the set of peak profiles and the associated room dimension data stored for the peak profiles. In this way, the estimator 105 can compare the peak signature of the room in question with the stored peak profiles and evaluate the dimension data of a room for the stored peak profiles based on such comparison. The room dimension data for the peak profiles, which are closely compatible with the peak signature, can be weighted above the room dimension data for the peak profiles that are compatible with the peak signature of the room in less ideally.
[0075] [075] An example of how the estimator 105 can generate an estimate of the dimensions of a room will be described below. In the example, an efficient search method is used to determine the most likely room size that corresponds to the set of detected peak frequencies, that is, the peak signature. The search method is based on the comparison of the detected peak / spectrum signature with the peak profiles of the database and the use of an appropriate error criterion to determine a set of compatible peak profiles / comfortable candidates. The estimation of the dimensions of a room is then generated from the dimensions of a room stored from such compatible peak profiles / candidate rooms. Specifically, a single compatible peak profile / compatible room can be identified and the room estimate can be defined as the stored room dimensions for that peak profile.
[0076] [076] It will be understood that in other modalities the estimator 105 can use other approaches to generate an estimate of the dimensions of a room based on comparisons and stored room dimension data.
[0077] [077] Figure 2 illustrates an example of elements of the estimator 105 according to some modalities.
[0078] [078] In the example, the estimator 105 comprises a searcher 201 that is coupled to the storage 107 that comprises the database. The search engine 201 is willing to find a compatible peak profile in the database for the peak signature generated from the room's acoustic response.
[0079] [079] The search engine 201 is coupled to an extractor 203 that is also coupled to storage 107 and which is willing to extract from the database the dimension data of a room associated with the compatible peak profile. Thus, when the search engine 201 identifies one or more compatible profiles in the database, the extractor 203 will extract the dimension data of a room stored in such data sets.
[0080] [080] Extractor 203 is coupled to an estimator generator 205 that is willing to estimate the dimensions of a room in response to extracted room dimension data.
[0081] [081] The approach will be described with reference to the specific example in which the peak signature is represented by a set of peak frequencies in the low frequency range of the room acoustic response. In this way, the peak signature can simply be represented by a set of frequency values that are the peak frequencies detected by the peak detector 103.
[0082] [082] Similarly, each peak profile comprises a set of peak frequencies. In this way, each data set comprises a set of frequency values that correspond to peaks in the determined room acoustic response (regardless of whether it is measured or calculated) for the room represented by the data set.
[0083] [083] In this way, both the peak signature and the peak profiles characterize the corresponding room acoustic response by various frequency values that correspond to peaks in a low frequency range. In an ideal case with no noise, these frequencies correspond to the room's Eigen frequencies and therefore provide a particularly efficient basis for finding compatible room acoustics that correspond to compatible rooms.
[0084] [084] In the example, finder 201 determines a distance measurement for each peak profile based on a comparison of the frequencies of the peak signature and the frequencies of the individual peak profile. For example, for each peak signature frequency value, searcher 201 can determine the frequency offset to the frequency value closest to the peak profile. This can be repeated for all peak signature frequencies, and frequency offsets or errors can be added together to provide an overall distance measurement (or error) for the peak profile. In this way, the search engine 201 can generate a single distance measurement for each peak profile.
[0085] [085] The search engine 201 can then proceed to identify the resulting peak profile in the lowest distance measure, and that peak profile can be selected as the compatible profile and fed to the 203 extractor that starts to extract the dimension data of a room in the corresponding data set. The data is then fed into the estimation generator 205 which generates the dimension estimate, for example, simply by using the directly retrieved room dimension data.
[0086] [086] In many scenarios, it may be advantageous that at least some of the peak profiles are calculated Eigen profiles rather than based on the measured acoustic responses of the room. Thus, in some modalities, at least part of the peak profiles can comprise a set of Eigen frequencies calculated for a room. The associated room dimension data can comprise an indication of the dimensions of the room used to calculate Eigen frequencies. This can provide a more accurate comparison in many scenarios, as it can reduce the impact of noise, measurement uncertainty, etc. In fact, it can be considered that the comparison seeks to compare the underlying Eigen frequency modes at low frequencies and that peak detection and processing can be seen as an estimate of such modes. Therefore, directly representing the underlying Eigen frequency modes instead of their estimates in peak profiles can provide an improved estimate.
[0087] [087] In the following, the specific example will be described in more detail.
[0088] [088] In the example, the first phase of the search procedure is to define a search space for possible room sizes. The search algorithm in the example is based on searching a database of possible candidates of the size of the room that is most likely to correspond to the set of detected peaks. This means that the scope of the database can, or often needs to, be restricted to a certain range of room sizes. This is done by defining a minimum and maximum size for each room dimension. In addition, for each dimension, a step size is defined that will determine the minimum difference in size that can be detected for that dimension. Having defined the range and step size for each dimension, a discrete room search space is defined, which will be considered candidates by the algorithm.
[0089] [089] The next step is to calculate the set of discrete Eigen frequencies that correspond to each of the rooms in the search space, for example, using equation (1). Since only Eigen frequencies in the low frequency range (where Eigen frequencies can be detected individually) are of interest, only Eigen frequencies up to a certain frequency need to be included. For example, for a typical sized living room, it may be sufficient to include Eigen frequencies up to about 100 Hz, of which there may be, typically, about 20.
[0090] [090] The resulting group of comfortable candidates and their corresponding sets of Eigen frequencies form the database that is searched by the algorithm.
[0091] [091] It should be noted that in the present description, the generation of the database is done as part of the search. However, it should be noted that in many modalities, the database can be populated before the search. For example, it can be pre-calculated during a design and / or manufacturing stage and stored as a look-up table on the device. It will also be understood that the database can be dynamically modified, for example, by adding additional peak profiles when deemed appropriate (for example, when a search is carried out in a new range of possible dimensions).
[0092] [092] In the example, search engine 201 calculates an error or distance measure for each of the candidate rooms in the search space (that is, for each peak profile), using a distance metric that reflects the difference between the set of peak frequencies detected and the set of Eigen frequencies of the candidate room (that is, between the frequencies of the peak signature and the peak profile).
[0093] [093] A specific implementation occurs as follows: For a candidate room l (out of a total of L rooms) and a detected peak frequency m (out of a total of M frequencies) from the measured room acoustic response of the room a to be estimated, it was found that the Eigen frequency n of the set (with a total amount of N) that corresponds to the peak / room profile l has the shortest distance (in frequency) dim to the detected peak m. This is done for all M peak frequencies detected from the peak signature and the resulting M distances dlm are added together to provide the general error measurement Dl for the peak / room profile l. This distance measure is calculated for all L peak profiles / candidate rooms. The peak profile and, therefore, the room candidate, which has the general error measurement Dl, is now completed to have the greatest probability of corresponding to the actual room geometry, since it has the best compatibility with the frequencies peak measurements.
[0094] [094] The error measurement can be, for example, calculated as:
[0095] [095] In this example, the distance measurement includes two weight factors w1 and W2 that can be used to refine the search. Weight factors can be, in many modalities, simply set to a value of 1.
[0096] [096] The peak profile in the lowest distance measure can be identified and the dimensions of a room stored for that peak profile / room candidate can be used as the estimation of the dimensions of a room for the room in question. For example, the width, length and height stored for the compatible peak profile can be produced as an estimate of the width, length and height of the room in question.
[0097] [097] Alternatively, to conclude directly that the only room candidate with a lower overall value of the distance measurement is the correct one, other strategies may be advantageous. While for simulated rooms it is true that the candidate room that corresponds to the modeled room does, in fact, have the lowest overall error value, this may not always be the case for real-life rooms. It can happen, for several reasons, that there are multiple (groups of) candidate candidates that are very different in geometry, but that have similar general distance values. This can result in the incorrect identification of the most suitable room candidate if the algorithm simply selects the candidate room with the lowest calculated distance measure.
[0098] [098] In some embodiments, the search engine 201 can, for example, identify a plurality of compatible peak profiles (for example, all those for which the distance measurement is below a certain threshold), and the extractor 203 can extract the room dimension data for all compatible peak profiles. The estimate generator 205 can then combine the extracted data to generate the dimension estimate. For example, an estimate of the length of a room can be generated by averaging stored lengths for compatible peak profiles, an estimate of width can be generated by averaging stored widths for compatible peak profiles, and a height estimate can be generated by averaging the stored heights for compatible peak profiles.
[0099] [099] In some modalities, the calculation of the average can be a calculation of the weighted average, in which the weight of each extracted dimension is weighted depending on the distance measurement generated for the corresponding peak profile.
[0100] [0100] Another approach that has been proven to improve the reliability of estimates in various scenarios, in which room estimates include a plurality of dimensional values corresponding to different directions (for example, width, length and height), is to determine indications of average compatibility (such as a distance measure, an error measure, or a probability measure) for a subset of directions in response to an average calculation (for example, a weighted average calculation) of compatibility indications for peak profiles in one or more directions that are not in the subset. The dimension estimate for the directions in the subset can then be determined using the mean compatibility indications.
[0101] [0101] The approach can calculate an average, for example, of distance measurements for different peak profiles in individual directions. For example, by identifying which peak profile provides the best compatibility for determining the width and length of a room, the distance measurement can be generated by averaging all peak profiles that are the same width and length, that is, averaging distance measurements for all possible room heights. In particular, if the peak profiles are provided with different measurements in three dimensions, a three-dimensional distance measurement matrix will be generated by calculating the distance measurement for each peak profile. In this case, all distance measures can be added or averaged in relation to a 3rd dimension of the matrix, thus generating a two-dimensional matrix of distance measures. The lowest distance measure can then be selected in such a two-dimensional matrix. Essentially, this can be considered to reduce the estimate to a two-dimensional estimate. In some modalities, the reduction can be for a single dimension, that is, the average can also be calculated in the second dimension. The approach can be applied in parallel to the different dimensions, thus allowing a three-dimensional estimate to be generated. It has been found, experimentally, that the approach provides improved performance in many scenarios.
[0102] [0102] In some embodiments, peak profiles may comprise a set of probability values where each probability value indicates a probability of a peak that is measured over a frequency range. Thus, instead of indicating the frequencies of the Eigen modes, the low frequency range can be divided into relatively small frequency ranges with a probability value that is provided for each interval. This can provide a more diffuse or widespread representation of the underlying Eigen modes and this can provide, in many scenarios, a more robust search.
[0103] [0103] In such modalities, the search engine 201 can, instead of generating a distance measure for each peak profile, determine a probability measure indicative of the probability that the frequencies of the peak signature result from a room that corresponds to that represented by the individual peak profile.
[0104] [0104] The probability value can be calculated, for example, for each peak signature frequency, by extracting the probability value for the corresponding frequency range and then multiplying the extracted probability values. Estimator 105 can now use the same approaches as described above, however, using the probability measure instead of the distance measure.
[0105] [0105] Thus, in some modalities, the estimator 105 can use a probability-based search approach, instead of an error measurement based on a distance metric. In this case, the database, instead of understanding peak profiles represented by a set of Eigen frequencies that correspond to each candidate room in the search space, can comprise peak profiles represented by a "probability vector" that, for each one of the K frequency ranges, contains the probability that a peak within that frequency range will be detected in the measured room response of a room that corresponds to the candidate room. Such probabilities can again be based mainly on the set of Eigen frequencies that correspond to each room candidate according to equation (1). A probability distribution can be applied to Eigen frequencies which can include, for example, aspects related to the frequency resolution of the measurement configuration, small differences between the "theoretical" Eigen frequencies and the real life of the room due to small deviations between the model theoretical room (which can be an empty rectangular box with rigid walls) and the real room situation (which may not be strictly rectangular and include all types of objects and a little absorption).
[0106] [0106] From the probabilities of the individual detected peaks, a general probability is calculated that provides an indication of the probability that the detected peaks correspond to a given candidate room. This general probability calculation can treat peak probabilities as independent (suggesting that individual probabilities can simply be multiplied), or can take into account some interdependence (for example, the presence of a peak at frequency f increases the probability of detecting a peak in multiples of the frequency, too).
[0107] [0107] A specific advantage of the approach is that it can provide estimates that do not simply reflect the overall size or volume of the room, but instead allow individual dimensions to be estimated and, in particular, allow the individual length, width and height are estimated.
[0108] [0108] In the specific examples above, the estimator 105 performs, specifically, a search to identify a subset and, possibly, only one among the peak profiles. The estimate is then generated from the selected subset. However, it will be understood that it is not essential for a search to be performed or for a subset to be selected.
[0109] [0109] For example, in some modalities, the estimator 105 can generate the estimates by combining the dimension data of a room from the different data sets, where the weight of the dimension data of an individual room depends on how much the profile corresponding peak is compatible with the peak signature.
[0110] [0110] Specifically, the estimator 105 can generate a compatibility indication for each peak profile. For example, the distance measurement described earlier can be used as an indication of compatibility. The distance measure can then be used, for example, to calculate the weighted average of the dimensions of a room. For example, an estimate of the width of a room can be generated by calculating the weighted average of all the width measurements stored with the peak profiles, where the weight for each width is determined by measuring the distance. Typically, weights can be a nonlinear function of the distance measurement, so that high distance measurements will result in weights that are substantially zero.
[0111] [0111] In some modalities, the estimation of the dimensions of a room may be based on a plurality of room acoustic responses and, specifically, it may be based on acoustic responses measured from the room to different positions of the sound source and / or the microphone . For example, a test signal can be rendered by a speaker and a microphone can capture the signal with the room acoustic response that is generated from it. The microphone (or speaker) can then be moved and the measurement repeated, to result in a new room acoustic response. The approach can be repeated more often.
[0112] [0112] The plurality of room acoustic responses can be used to generate an estimate of the dimensions of a room and, consequently, a combination of data for the different room acoustic responses is necessary at some stage in the processing.
[0113] [0113] In some modalities, the combination can be an average (low-pass filtration).
[0114] [0114] The averaging may have already been done, for example, by averaging room acoustic responses to generate an average room acoustic response that is then used in the same way, as previously described for a single room acoustic response. In this way, peak detection can be applied to the average room acoustic response and the resulting peak signature can be used.
[0115] [0115] However, in other modalities, the calculation of the average can be carried out at other stages of processing. For example, peak detection can be applied individually to each room's acoustic response to generate a plurality of peak signature. A comparison can then be made for each of them and averaged the result of the comparisons. For example, a distance measure can be calculated for each peak signature and for each peak profile. The resulting distance measurements can then be averaged for each peak profile and the compatible peak profile can be selected as the one with the lowest average distance measurement.
[0116] [0116] In some modalities, the average can be calculated from room dimension estimates determined for different room acoustic responses. For example, the approaches described above for a single room acoustic response can be applied to each room acoustic response of the plurality of room acoustic responses. The resulting room dimension estimates can then be averaged to generate the exit estimates. For example, all length estimates can be averaged to generate a single length estimate.
[0117] [0117] Thus, in many modalities, the estimate can advantageously be based on measurements of a plurality of room acoustic responses with the use of multiple source and / or receiver positions for the measurements.
[0118] [0118] In some modalities, the combination may comprise the generation of a peak signature by including peaks from a plurality of room acoustic responses.
[0119] [0119] As can be understood from equations (2) and (3), the extent to which each of the room's Eigen modes is excited depends on the position of the sound source. Similarly, the resulting sound pressure due to each Eigen mode depends on the position of the receiver. Consequently, the combination of the source position and the listening location determines which of the room's low-frequency modes will actually be visible in the measured room response. For some combinations, it is possible to detect all relevant modes from a single measurement, while for others only a few peaks will be detected from a single measurement.
[0120] [0120] By combining the detected peaks resulting from the peak detection applied to room acoustic responses for two or more source and / or receiver positions, the results of the method can be improved. This achieves a double improvement. First, by combining the peaks from multiple source / receiver combinations, a more complete set of peaks can be obtained, since one position will provide peaks that are invisible in another position and vice versa. Combining the peaks from multiple source / receiver position combinations will reduce the possibility of misidentification due to an incomplete set of peaks that is available for the search algorithm. Second, the peaks that are detected in the response from more than one source / receiver position combination can be considered more reliable than the peaks that are detected in only one of the responses. One way to explore this is to feed the total set of combined peaks, so as to include duplications that occur in multiple source / receiver position combinations, in the search algorithm that thereby efficiently assigns more weight to the peaks that are detected multiple times.
[0121] [0121] In principle, it does not matter whether multiple source positions, multiple receiver positions or both are used, since the source and receiver positions are interchangeable (acoustic reciprocity principle).
[0122] [0122] In some modalities, the system may have a microphone in a fixed position, for example, integrated with some main unit of the system and multiple speakers. In this case, the number of responses that can be obtained is the same for different speakers.
[0123] [0123] In other modalities, a microphone can be integrated into each speaker enclosure and measurements can be made between all pairs of speakers and microphones or a subset of all such pairs. For a system with N speakers, there are N (N + 1) / 2 pairs in total, including those with speaker and microphone in the same enclosure.
[0124] [0124] A particularly interesting case is that when the source and / or the receiver are positioned in a corner. From equation (3), it can be seen that, in theory, a source that is placed in a corner should excite all room modes with maximum intensity. Similarly, a microphone that is placed in a corner must produce a signal that shows all the modal frequencies that are excited by the source. This means that if the use case allows the source and / or the receiver to be placed in a corner, this should enable a very complete set of peaks to be obtained only from a few measurements. In the special case where both the source and the receiver are placed in one corner (in the same corner or in different corners), a single measurement should, in principle, provide the complete set of Eigen frequencies.
[0125] [0125] In some embodiments, the estimator 105 may be willing to weight different peaks in the set of peaks differently. For example, peaks can be weighted differently, depending on the peak frequency or depending on the peak amplitude, or based on a measure of reliability generated by peak detection. This can be done, for example, by varying the weights w1 in equation (4) depending on such factors.
[0126] [0126] In some embodiments, the estimator 105 may be willing to weight different peaks of a peak profile differently when making the comparison, such as, for example, when determining the distance measurement. For example, the distance measurement may be weighted differently depending on the type of room. This can be done, for example, by varying the weights w2 in equation (4).
[0127] [0127] In fact, although the measure of distance from equation (4) yields good estimates in many scenarios without changing the weights w1 and W2 (ie, where w1 = 1 and W2 = 1), the results may, in others scenarios, be significantly improved using variable weights.
[0128] [0128] The effect of variable weight w1 is to place more emphasis on some detected peaks than on others. A reasonable option is to make wl a frequency function, where the weight decreases to increase the frequency. The logic behind this is that, since the density of Eigen frequencies increases, to increase the frequency, the higher peak frequencies are less characteristic of a specific room size than the lower peak frequencies.
[0129] [0129] The effect of the W2 weight is to attach more importance to certain Eigen frequencies in the set that correspond to a certain candidate room than to others. As with w1, this could be done a frequency function in a similar way. Another possibility is to make W2 dependent on the type of mode.
[0130] [0130] In a three-dimensional room, there are three different types of modes: axial, tangential and oblique. Axial modes are modes in which pressure variations are oriented across a single dimension, while pressure is constant across the other two dimensions. Tangential modes have pressure variations in two dimensions, while oblique modes involve all three dimensions. An analysis of the distribution of the different types of modes in a room and their prominence in the room's acoustic response (modeled or measured) reveals that axial modes and low-order tangential modes are more likely to stand out clearly in the room's acoustic response. than oblique modes or tangential modes of a higher order. As a result, a peak that is detected in a measured room response is more likely to correspond to a low-order axial or tangential mode than another type of mode. This can be explained by assigning different W2 weights to the different types of modes, so that low-order axial and tangential modes have a greater chance of being selected as the greater compatibility of a detected peak frequency than other types of mode. .
[0131] [0131] In some embodiments, estimator 105 is willing to select a subset of the set of peak profiles to use in response to a user input that indicates a room dimension value. In this way, a range of room candidates / peak profiles can be selected and only these will be used in the comparison and search. The user can, for example, indicate an approximate dimension and the system can generate a range that includes that dimension. In other modalities, the user can directly specify the range, for example, the length, width and height of possible candidate rooms.
[0132] [0132] In some modalities, the external information that can be useful to identify the correct dimensions of a room can be used by the estimator 105 and, specifically, can be used to make the search more reliable. External information can consist of "physical" data, such as known positions of speakers and / or microphones in relation to each other, knowing that one or more dimensions of the room are in a certain range, acoustic parameters of the room, such as the reverberation time, etc.
[0133] [0133] External information may, in some modalities, consist of "heuristic" data, such as statistics on how common a room dimension ratio is in real life.
[0134] [0134] External information can be used to select a subset of peak profiles, thus reducing the space search for candidate room sizes to only those that match the external information. This will not only reduce the chance of misidentification, but will also speed up the search procedure.
[0135] [0135] Alternatively or additionally, external information can be used to differentiate between compatible peak profiles resulting from a search. This can be particularly appropriate when peak profiles are found that correspond to very different candidate rooms, however, which may have similar distance or probability measures.
[0136] [0136] The inclusion of external information will be particularly beneficial in cases where the set of measured peaks is considerably incomplete (for example, because only one position measurement was used, or the measurement positions were such that only a few peaks were found) present in the answers), in which case room size ambiguity may occur.
[0137] [0137] Simulations were made to evaluate the approach. In the simulations, the acoustic response of the room and the peak profiles were molded using the modal decomposition model according to equations (2) and (3).
[0138] [0138] The modeled room had dimensions of 7.4 x 4.2 x 2.8 m and an absorption coefficient α = 0.3 (which corresponds to approximately δ = 2 0 in equation (2)). Figure 3 illustrates a room acoustic response of the room modeled in four randomly selected positions in the room. Eigen frequencies that correspond to axial and tangential modes in the room are indicated by vertical lines. The execution of a simple peak detection algorithm returned the peak frequencies indicated by the circles. As can be seen, each of the four simulated responses includes only a subset of the complete set of Eigen frequencies. However, all the peaks that have been detected belong to the set and the four subsets are, in fact, partially complementary.
[0139] [0139] The detected peaks were inserted in a search algorithm using the distance measure of equation (4) in which both weights w1 and w2 are statistically defined as 1 and where the search is limited to the profiles of peak corresponding to the dimensions of 2.5 m to 8 m for the first two dimensions and between 2 and 4 m for the third dimension. The pitch size for the peak profiles was defined as 0.1 m in all dimensions. The lowest overall distance measure was found for a geometry of 7.4 x 4.3 x 3.6 m, so two of the three dimensions are accurate up to 0.1 m. However, if the distance measure is averaged over the third dimension, the lowest distance measure will be found for 7.4 and 4.2 m for the first two dimensions. Calculating the average of the distance measure over the second dimension results in the lowest distance measure for 7.4 and 2.8 m for the first and third dimensions, respectively. In this way, in this case, accurate estimates are generated for all dimensions.
[0140] [0140] The approach was also assessed through practical experiments. The measurements were taken in several real rooms. As an example, a room 7.4 m long and 5.7 m wide, and furnished to be representative of a typical living room, with a carpet and curtains, several large chests of drawers against the walls, a large dining table and chairs and many other large and small objects were used to test the approach.
[0141] [0141] Measurements were made with four full-range speakers distributed arbitrarily in the room. Each speaker was equipped with a microphone. Logarithmic scan measurements were performed between each speaker to generate room acoustic responses. Figure 4 illustrates an example of a room response measured after a mild smoothing with a Gaussian window. The peaks were detected from that smoothed response. The detected peaks are indicated in Figure 4 by circles. Also indicated by vertical lines are the axial and tangential modal theoretical frequencies that correspond to the actual room geometry. It can be seen that most of the detected peaks correspond, in fact, to the theoretically expected Eigen frequencies.
[0142] [0142] The detected peak frequencies were fed into the search algorithm, which returned the correct room dimensions with an accuracy of 0.1 m (the step size used in the search algorithm). Similarly, results were obtained for other rooms.
[0143] [0143] It will be understood that for clarity the above description described the modalities of the invention with reference to different circuits, units and functional processors. However, it will be evident that any suitable distribution of functionality between the different functional circuits, units or processors can be used without deviating from the invention. For example, the illustrated functionality to be performed by separate processors or controllers can be performed by the same processor or controller. Therefore, references to specific functional units or circuits should be considered only as references to appropriate means of providing the described functionality and not as indicative of a physical structure or a strict logical or physical organization.
[0144] [0144] The invention can be implemented in any suitable form, including hardware, software, firmware or any combination thereof. The invention can optionally be implemented, at least partially, as computer software running on one or more data processors and / or digital signal processors. The elements and components of an embodiment of the invention can be physically, functionally and logically implemented in any suitable way. In fact, the functionality can be implemented in a single unit, in a plurality of units or as part of other functional units. In this way, the invention can be implemented in a single unit or it can be physically and functionally distributed among different units, circuits and processors.
[0145] [0145] Although the present invention has been described in conjunction with some modalities, it is not intended to be limited to the specific form presented here. Instead, the scope of the present invention is limited only by the appended claims. In addition, although it may appear that a given resource is described in conjunction with specific modalities, the person skilled in the art will recognize that various features of the described modalities can be combined according to the invention. In the claims, the term "comprising" does not exclude the presence of other elements or stages.
[0146] [0146] Furthermore, although individually mentioned, a plurality of means, elements, circuits or method steps can be implemented, for example, by a single circuit, unit or processor. In addition, although individual resources may be included in different claims, they can be advantageously combined, and their inclusion in different claims does not imply that a combination of resources is not feasible and / or advantageous. The inclusion of an appeal in a category of claims also does not imply a limitation to that category, but instead indicates that the appeal is equally applicable to categories of other claims, as appropriate. In addition, the order of the resources in the claims does not imply any specific order in which the resources need to be worked on, and in particular, the order of the individual steps in a method claim does not imply that the steps need to be performed in that order. Instead, the steps can be performed in any suitable order. In addition, singular references do not exclude a plurality. Thus, references to "one / a", "one / one", "first / a", "second / a", etc., do not exclude a plurality. The reference signs in the claims are provided for illustrative purposes only and should not be construed as limiting the scope of the claims in any way.

权利要求:
Claims (15)
[0001]
APPLIANCE TO DETERMINE AN ESTIMATION OF THE DIMENSIONS OF A ROOM, the appliance being characterized by comprising: a receiver (101) for providing a room acoustic response; a peak detector (103) for detecting a set of peaks in the room's acoustic response in a frequency range, the frequency range having a frequency of at most 400 Hz maximum; a storage (107) for storing a set of peak profiles with associated room dimension data; an estimator (105) to determine the estimation of the dimensions of a room in response to the associated room dimension data and a comparison of the set of peaks with the set of peak profiles.
[0002]
APPARATUS, according to claim 1, characterized by the estimator (105) comprising: a searcher (201) to find at least one peak profile compatible for the peak set from the peak profile set; an extractor (203) for extracting first room dimension data from storage associated with at least one compatible peak profile; and an estimation generator (205) to determine the estimation of the dimensions of a room in response to the first dimension data of a room.
[0003]
APPLIANCE according to claim 2, characterized in that each peak profile comprises a set of peak frequencies.
[0004]
APPARATUS, according to claim 3, characterized in that the searcher (201) is arranged to determine a distance measurement for each peak profile in response to a comparison of the frequencies of the peak set with the frequency set of the peak profile, and to select the compatible peak profile in response to distance measurements.
[0005]
APPARATUS, according to claim 2, characterized in that each peak profile comprises a set of probability values, where each probability value indicates a probability that a peak will be measured in a frequency range.
[0006]
APPARATUS, according to claim 5, characterized by the searcher (201) being arranged to determine a probability measure for each peak profile in response to the probability values of the peak profile and peak set frequencies, and to select the profile compatible peak in response to probability measures.
[0007]
APPARATUS, according to claim 1, characterized in that each peak profile comprises a set of Eigen frequencies calculated for a room and the associated room dimension data comprise an indication of dimensions for the room used to calculate Eigen frequencies.
[0008]
APPARATUS, according to claim 1, characterized by the associated room size, for at least some of the peak profiles, comprising at least one one-dimensional value, and the estimator (105) being arranged to generate the estimation of the dimensions of a room for understand at least one-dimensional value.
[0009]
APPARATUS, according to claim 1, characterized in that the estimator (105) is arranged to generate an indication of compatibility for the set of peaks and each peak profile of the set of profiles, and to generate the dimension estimate by means of a combination weighted of the associated room dimension data, where the weight of the associated room dimension data of a first peak profile from the set of peak profiles depends on the compatibility indication of the first peak profile.
[0010]
APPARATUS, according to claim 1, characterized by the associated room size, for at least some of the peak profiles, comprising a plurality of one-dimensional values corresponding to different directions, and by the estimator being arranged to determine indications of average compatibility for a subset from directions in response to an average of compatibility indications for peak profiles in at least one direction that is not in the subset, and to determine unidirectional dimensions estimates of a room to the subset directions in response to average compatibility indications.
[0011]
APPLIANCE, according to claim 1, characterized in that the receiver (101) is arranged to receive a plurality of room acoustic responses corresponding to different positions of at least one of a sound source and a microphone, and the apparatus is arranged to perform a combination for the plurality of room acoustic responses.
[0012]
APPARATUS, according to claim 11, characterized in that the combination includes at least one among a calculation of the mean of comparisons between sets of peaks that correspond to different acoustic responses of the room with the set of peak profiles and a calculation of the average of estimates of room dimensions determined for different room acoustic responses.
[0013]
APPARATUS, according to claim 11, characterized by the combination comprising the generation of the set of peaks through the inclusion of peaks from more than one among the plurality of room acoustic responses.
[0014]
APPARATUS, according to claim 1, characterized in that the estimator (105) is arranged to weight at least one among different peaks in the set of peaks and different peaks of a peak profile differently when performing the comparison.
[0015]
METHOD FOR DETERMINING AN ESTIMATION OF THE DIMENSIONS OF A ROOM, the method being characterized by understanding: provide a room acoustic response; detect a set of peaks in the room's acoustic response in a frequency range, in which the frequency range has a maximum frequency of at most 400 Hz; provide a set of peak profiles with associated room dimension data; and determine the estimation of the dimensions of a room in response to the associated room dimension data and a comparison of the set of peaks with the set of peak profiles.

类似技术:

---

公开号 | 公开日 | 专利标题

---

BR112015028335B1|2021-02-17|apparatus to determine an estimate of the dimensions of a room, and, method to determine an estimate of the dimensions of a room

---

Lawson et al.2015|On velocity gradient dynamics and turbulent structure

---

Li et al.2013|Two single-shot methods for locating multiple electromagnetic scatterers

---

US9256933B2|2016-02-09|System for determining flow properties of a blood vessel

---

KR101909544B1|2018-10-18|Apparatus and method for plane detection

---

Rogers et al.2003|Using statistical image models for objective evaluation of spot detection in two‐dimensional gels

---

JP4422662B2|2010-02-24|Sound source position / sound receiving position estimation method, apparatus thereof, program thereof, and recording medium thereof

---

CN105403878B|2018-07-10|A kind of Ocean Acoustic Tomography method based on delay sensitive core

---

Popescu2011|Nonparametric signal detectability evaluation using an exponential transformation of the FROC curve

---

Salvati et al.2018|Sensitivity-based region selection in the steered response power algorithm

---

TWI686180B|2020-03-01|Sports information acquisition method and device

---

US10404524B2|2019-09-03|Resource and metric ranking by differential analysis

---

BR112019011281A2|2019-12-31|dive wave lighting using migrated collections

---

CN104680118B|2018-06-15|A kind of face character detection model generation method and system

---

JP2004349846A|2004-12-09|Outlier detecting method

---

Tholin-Chittenden et al.2018|Automatic parameter selection of image reconstruction algorithms for planar array capacitive imaging

---

Müller et al.2007|Evaluation of left atrial size in patients with atrial arrhythmias: comparison of standard 2D versus real time 3D echocardiography

---

Hernández et al.2014|Overestimation and underestimation biases in photon mapping with non-constant kernels

---

Vogiatzis2013|Influence study on hyper-graphs

---

KR20200048883A|2020-05-08|Method and system for searching generating position of sound generated because of interstinal disease

---

US20170371418A1|2017-12-28|Method for recognizing multiple user actions on basis of sound information

---

JP2009207764A|2009-09-17|Visceral fat amount estimation system

---

Kai-ming et al.2011|Positioning ability comparison research on several matched-field processing methods with increasing white noise

---

GB2465888A|2010-06-09|Locating a physiological feature in image data using an intensity projection line

---

JP2015055843A|2015-03-23|Number-of-signal-source estimation device, number-of-signal-source estimation method, and program

同族专利:

---

公开号 | 公开日

---

EP2997327B1|2016-12-07|

---

RU2015153383A3|2018-03-22|

---

RU2655703C2|2018-05-29|

---

CN105229414A|2016-01-06|

---

JP5998306B2|2016-09-28|

---

BR112015028335A2|2017-07-25|

---

CN105229414B|2018-11-23|

---

US20160061597A1|2016-03-03|

---

EP2997327A1|2016-03-23|

---

JP2016524693A|2016-08-18|

---

WO2014183970A1|2014-11-20|

---

RU2015153383A|2017-06-19|

---

US9909863B2|2018-03-06|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

---

US4470048A|1982-03-29|1984-09-04|Sperry Corporation|Range profile target classifier|

---

US4890266A|1987-04-22|1989-12-26|Federal Industries Industrial Group Inc.|Acoustic range finding system|

---

JPH07146127A|1993-11-24|1995-06-06|Nippon Telegr & Teleph Corp <Ntt>|Apparatus for measuring moving object|

---

JP2004101343A|2002-09-09|2004-04-02|Hosiden Corp|Distance measuring machine|

---

WO2004097350A2|2003-04-28|2004-11-11|The Board Of Trustees Of The University Of Illinois|Room volume and room dimension estimation|

---

JP3834660B2|2004-02-18|2006-10-18|国立大学法人鳥取大学|Crack detection device for structures|

---

JP2005337848A|2004-05-26|2005-12-08|Matsushita Electric Ind Co Ltd|Ultrasonic distance measuring device|

---

EP1863190B1|2005-03-09|2016-11-02|Omron Corporation|Distance measuring apparatus, distance measuring method, reflector and communication system|

---

WO2007028094A1|2005-09-02|2007-03-08|Harman International Industries, Incorporated|Self-calibrating loudspeaker|

---

JP2007309838A|2006-05-19|2007-11-29|Non-Destructive Inspection Co Ltd|Conduit inspection method and conduit inspection device used therefor|

---

JP4470048B2|2006-05-20|2010-06-02|船井電機株式会社|Information recording and playback equipment, electronic equipment|

---

US8194874B2|2007-05-22|2012-06-05|Polk Audio, Inc.|In-room acoustic magnitude response smoothing via summation of correction signals|

---

DE102007031677B4|2007-07-06|2010-05-20|Sda Software Design Ahnert Gmbh|Method and apparatus for determining a room acoustic impulse response in the time domain|

---

US8396241B2|2008-05-21|2013-03-12|Nittobo Acoustic Engineering Co., Ltd.|Reflector structure, sound field adjusting method, columnar reflector structure, room, program, and various acoustic room designing system|

---

JP5451188B2|2009-06-02|2014-03-26|キヤノン株式会社|Standing wave detection device and control method thereof|

---

EP2572212B1|2010-05-20|2013-12-25|Koninklijke Philips N.V.|Distance estimation using sound signals|

---

JP5857071B2|2011-01-05|2016-02-10|コーニンクレッカ フィリップス エヌ ヴェＫｏｎｉｎｋｌｉｊｋｅ Ｐｈｉｌｉｐｓ Ｎ．Ｖ．|Audio system and operation method thereof|

---

US9031268B2|2011-05-09|2015-05-12|Dts, Inc.|Room characterization and correction for multi-channel audio|ES2543710T3|2002-04-04|2015-08-21|Zoetis Belgium S.A.|Immunostimulatory oligonucleotides containing G and U|

---

US9084058B2|2011-12-29|2015-07-14|Sonos, Inc.|Sound field calibration using listener localization|

---

WO2016172593A1|2015-04-24|2016-10-27|Sonos, Inc.|Playback device calibration user interfaces|

---

US10664224B2|2015-04-24|2020-05-26|Sonos, Inc.|Speaker calibration user interface|

---

US9106192B2|2012-06-28|2015-08-11|Sonos, Inc.|System and method for device playback calibration|

---

US9219460B2|2014-03-17|2015-12-22|Sonos, Inc.|Audio settings based on environment|

---

US9264839B2|2014-03-17|2016-02-16|Sonos, Inc.|Playback device configuration based on proximity detection|

---

US9706323B2|2014-09-09|2017-07-11|Sonos, Inc.|Playback device calibration|

---

EP3531714B1|2015-09-17|2022-02-23|Sonos Inc.|Facilitating calibration of an audio playback device|

---

US9952825B2|2014-09-09|2018-04-24|Sonos, Inc.|Audio processing algorithms|

---

DE102015106114B4|2015-04-21|2017-10-26|D & B Audiotechnik Gmbh|METHOD AND DEVICE FOR POSITION DETECTION OF SPEAKER BOXES OF A SPEAKER BOX ARRANGEMENT|

---

US9538305B2|2015-07-28|2017-01-03|Sonos, Inc.|Calibration error conditions|

---

US10416306B2|2015-08-17|2019-09-17|Texas Instruments Incorporated|Methods and apparatus to measure and analyze vibration signatures|

---

US9693165B2|2015-09-17|2017-06-27|Sonos, Inc.|Validation of audio calibration using multi-dimensional motion check|

---

US9743207B1|2016-01-18|2017-08-22|Sonos, Inc.|Calibration using multiple recording devices|

---

US10003899B2|2016-01-25|2018-06-19|Sonos, Inc.|Calibration with particular locations|

---

US11106423B2|2016-01-25|2021-08-31|Sonos, Inc.|Evaluating calibration of a playback device|

---

US9864574B2|2016-04-01|2018-01-09|Sonos, Inc.|Playback device calibration based on representation spectral characteristics|

---

US9860662B2|2016-04-01|2018-01-02|Sonos, Inc.|Updating playback device configuration information based on calibration data|

---

US9763018B1|2016-04-12|2017-09-12|Sonos, Inc.|Calibration of audio playback devices|

---

US9794710B1|2016-07-15|2017-10-17|Sonos, Inc.|Spatial audio correction|

---

US10372406B2|2016-07-22|2019-08-06|Sonos, Inc.|Calibration interface|

---

US10459684B2|2016-08-05|2019-10-29|Sonos, Inc.|Calibration of a playback device based on an estimated frequency response|

---

US11172320B1|2017-05-31|2021-11-09|Apple Inc.|Spatial impulse response synthesis|

---

EP3422737A1|2017-06-28|2019-01-02|Vestel Elektronik Sanayi ve Ticaret A.S.|Audio device and method for generating a flat acoustic response|

---

KR102334070B1|2018-01-18|2021-12-03|삼성전자주식회사|Electric apparatus and method for control thereof|

---

US11206484B2|2018-08-28|2021-12-21|Sonos, Inc.|Passive speaker authentication|

---

US10299061B1|2018-08-28|2019-05-21|Sonos, Inc.|Playback device calibration|

---

US10734965B1|2019-08-12|2020-08-04|Sonos, Inc.|Audio calibration of a portable playback device|

法律状态:
2018-11-13| B06F| Objections, documents and/or translations needed after an examination request according art. 34 industrial property law|

---

2020-04-14| B06U| Preliminary requirement: requests with searches performed by other patent offices: suspension of the patent application procedure|

---

2020-12-08| B09A| Decision: intention to grant|

---

2021-02-17| B16A| Patent or certificate of addition of invention granted|Free format text: PRAZO DE VALIDADE: 20 (VINTE) ANOS CONTADOS A PARTIR DE 24/04/2014, OBSERVADAS AS CONDICOES LEGAIS. |

优先权:

---

申请号 | 申请日 | 专利标题

---

EP13167971.4|2013-05-16|

---

EP13167971|2013-05-16|

---

PCT/EP2014/058362|WO2014183970A1|2013-05-16|2014-04-24|Determination of a room dimension estimate.|

---