|
(11) | EP 2 963 950 A1 |
| (12) | EUROPEAN PATENT APPLICATION |
|
|
|
|
|||||||||||||||||||||||
| (54) | Modal response compensation |
| (57) A method for compensating for a modal response of a listening room on an acoustic
output, comprising transforming a left channel signal and a right channel signal to
a mid channel signal and a side channel signal, applying a mid channel filter to the
mid channel signal to provide a filtered mid channel signal, applying a side channel
filter to the side channel signal to provide a filtered side channel signal, retransforming
the filtered mid channel signal and the filtered side channel signal to a left channel
output and a right channel output, applying the filtered left channel output to a
first loudspeaker, and applying the filtered right channel output a second loudspeaker. The resulting filtering avoids applying filtering to musical signals that do not require correction due to the interaction (or lack thereof) between the signals (as a result of their correlation coefficient) and the listening room's resonant behavior. |
Field of the invention
[0001] The present invention relates to a method for compensating for a modal response of a listening room on an acoustic output from an audio system including at least a first and a second loudspeaker.
Background of the invention
[0002] A typical listening room exhibits, as part of its acoustical characteristics, modal frequencies that are dependent on the dimensions of the room. It is common amongst room compensation algorithms in loudspeaker systems to use filters that have the reciprocal of the magnitude responses of this modal behavior. In other words, where the room mode creates an increase in the signal at a location in a listening room (due to resonating standing waves) the audio system includes a filter that reduces the signal by the same amount.
[0003] Figure 1 shows a conceptual drawing of the behavior of standing waves in one dimension of a room. Note that the lowest modal frequency (the top curve with a frequency of (1f) has an opposite polarity on either side of the room. The next modal frequency (2f) has the same polarity on either side of the room. This pattern continues - odd numbered modes have opposite polarity, even-numbered modes have similar polarity.
[0004] Figure 2 shows a simplified magnitude response of the modal response in one dimension of a listening room with a single sound source.
[0005] In current filtering systems that are intended to compensate for a listening room's acoustical behavior, loudspeakers in a system are considered to be independent sources. They are therefore measured as such using some version of the following method:
A microphone (or microphones) is used to measure the behavior of a loudspeaker in a given position or positions in a room. The calculated response (in the time domain or the frequency domain) is used to create a filter that, in some way, is the reciprocal of the room's behavior. The response of the filter may be calculated in the frequency or time domain and it may or may not be smoothed. Various techniques are currently employed in many different varieties of systems.
[0006] Each loudspeaker in a system is measured using this method, and a filter is created independently for each. Figure 3 shows a typical filtering strategy for compensation of modal behavior in a room, and figure 4 shows a typical filtering magnitude response for compensation of modal behavior in a room with a single source. Note that this is the inverse of the modal response shown in Figure 2.
[0007] A problem with current systems is that the loudspeakers are considered as independent sources. However, in a typical situation, the loudspeakers are playing correlated (or negatively correlated) signals. In fact, they are very rarely used to play uncorrelated signals when used to play typical program material such as music. Consequently, their behavior in conjunction with the listening room's acoustical behavior should not be considered independently of each other.
General disclosure of the invention
[0009] This and other objects are achieved by a method for compensating for a modal response of a listening room on an acoustic output from an audio system including at least a first and a second loudspeaker, the method comprising receiving a left channel input signal and a right channel input signal, transforming the left channel signal and the right channel signal to a mid channel signal and a side channel signal, applying a mid channel filter to the mid channel signal, to provide a filtered mid channel signal, applying a side channel filter to the side channel signal, to provide a filtered side channel signal, retransforming the filtered mid channel signal and the filtered side channel signal to a left channel output and a right channel output, applying the filtered left channel output to the first loudspeaker, and applying the filtered right channel output the second loudspeaker. The mid channel filter is created by applying correlated signals to the first and the second loudspeakers, measuring a correlated modal response of the listening room, creating the mid channel filter to have a modal response substantially equal to an inverse of the correlated modal response. The side channel filter is created by applying negatively correlated signals to the first and the second loudspeakers, measuring a negatively correlated modal response of the listening room, and creating the side channel filter to have a modal response substantially equal to an inverse of the negatively correlated modal response.
[0010] The invention is based on the realization that the measurement procedure, regardless of the method, should use the two loudspeakers as positively correlated and negatively correlated sources, instead of independent left and right sources. The resulting filters are then applied to the summed "mid" (or 'M') and difference "side" (or'S') channels formed from the right and left channels and compensated right and left channels are formed from the filtered M and S channels.
[0011] The resulting filtering is more efficient, as it avoids applying filtering to musical signals that do not require correction due to the interaction (or lack thereof) between the signals (as a result of their correlation coefficient) and the listening room's resonant behavior.
[0012] The method consists of performing the loudspeaker measurements using the same stimuli and computational method as desired and as is currently in use with existing systems. Consequently, the actual implementation of the filters themselves (both how they are created and how they are implemented) is independent of this invention. This invention can be applied to any prior art (or future) filtering method that is designed for a single loudspeaker in a multi-loudspeaker configuration.
[0013] The difference lies in the presentation of the stimulus during the measurement procedure. First, the loudspeaker+room response of both loudspeakers playing correlated signals is measured and a filter is created using the method of the user's choice. This filter is referred to as the Filter for Positively Correlated Signals, Fpos. Then, the loudspeaker+room response of both loudspeakers playing negatively correlated signals is measured and a filter is created using the method of the user's choice. This filter is referred to as the Filter for Negatively Correlated Signals, Fneg.
[0014] One additional advantage of this system is that the same filtering is applied to the correlated signals in both loudspeakers. Thus, there is no impact on phantom images panned to be localized directly between the two loudspeakers. Similarly, negatively correlated signals are also filtered similarly, thus maintaining their negative correlation and consequently their perceived spatial impression at the listening position.
Brief description of the drawings
[0015] The present invention will be described in more detail with reference to the appended drawings, showing currently preferred embodiments of the invention.
Figure 1 shows a conceptual drawing of the behavior of standing waves in one dimension of a room.
Figure 2 shows a simplified magnitude response of the modal response in one dimension of a listening room with a single sound source.
Figure 3 shows a typical filtering strategy for compensation of modal behavior in a room.
Figure 4 shows a typical filtering magnitude response for compensation of modal behavior in a room with a single source.
Figure 5 shows the behavior of standing waves in one dimension of a room being activated by two loudspeakers playing positively correlated signals.
Figure 6 shows a simplified magnitude response of the modal response in one dimension of a listening room with two sound sources playing correlated signals.
Figure 7 shows an example of the desired filtering magnitude response for compensation of the modal response in Figure 6.
Figure 8 shows the behavior of standing waves in one dimension of a room being activated by two loudspeakers that are playing negatively correlated signals.
Figure 9 shows a simplified magnitude response of the modal response in one dimension of a listening room with two sound sources playing negatively correlated signals.
Figure 10 shows an example of the desired magnitude response for compensation of the modal response in figure 9.
Figure 11 shows a block diagram of an implementation according to a first embodiment of the present invention.
Figure 12 shows a block diagram of an implementation according to a second embodiment of the present invention.
Detailed description of preferred embodiments
[0016] When two loudspeakers are placed in opposite sides of one dimension of a listening room and are playing correlated signals, they have the same polarity at all times. This directly conflicts with the odd-numbered modal frequencies in the room, which, as is shown in Figure 1, are opposite in polarity on either side of the room. Consequently, these modes are not activated in the room, since they are being actively cancelled by the signals from the two loudspeakers. This is illustrated in figures 5- 6. Figure 7 shows the desired filter magnitude response required to compensate for the modal response in figure 6. Note that this is the inverse of the modal response in Figure 6.
[0017] Simultaneously, the negatively correlated signals in the two loudspeakers actively cancel the even-numbered modal frequencies which have similar polarities on either side of the room. The even-numbers modes are not activated, since the loudspeakers are negatively correlated and the modal behavior from the locations of the loudspeakers has the same polarities. This is illustrated in figures 8-9. Figure 10 shows the desired filter magnitude response required to compensate for the modal response in figure 9. Note that this is the inverse of the modal response in Figure 9.
Main embodiment
[0018] In a first embodiment, illustrated in figure 11, the filters Fpos and Fneg are created and applied in a central unit 2 of an audio system 1, with access to both left and right audio channels (L, R). This is referred to as a "central" system. Typically, but not necessarily, the central unit 2 is the same unit that receives and amplifies the audio source signal. The central unit is connected to a left speaker 3 and a right speaker 4.
[0019] The central unit 2 here includes circuitry 10 for creating, storing and applying the filters Fpos 11 and Fneg 12. The central unit 2 further has a L/R-to-M/S converter 13, the summed or "M" output being connected to the Fpos filter 11 and the difference or "S" output being connected to the Fneg filter 12. The filtered outputs Mfilter, Sfilter of the filters 11, 12 are each connected to a M/S-toR converter 14 and a M/S-to-L converter 15, respectively. The outputs Lout, Rout from these converters 14, 15 are connected to the left and right speakers 3, 4, respectively. It is noted that the converters 14, 15 may be integrated as one M/S-to-L/R converter. The unit 2 is also connected to at least one microphone 9, which provides input to the circuitry 10.
[0021] During an initiation phase, the circuitry 10 controls the audio system to output correlated signals on each loudspeaker 3, 4. The resulting audio signals are received by the microphone(s) 9, and supplied to the circuitry 10. The circuitry then determines a first room response, referred to as a room response for correlated signals. The first room response is used to define and create the filter 11. Then, the circuitry controls the audio system to output negatively correlated signals on each loudspeaker 3, 4. Again, the resulting audio signals are received by the microphone(s) 9, and supplied to the circuitry 10. The circuitry then determines a second room response, referred to as a room response for negatively correlated signals. This second room response is used to define and create the filter 12.
[0022] It is noted that the determination of room response can be performed using conventional measurement techniques. Also, the filters can be created using conventional techniques.
[0023] When used with room compensation systems that employ a target magnitude response, it should be remembered that the target responses of the Positively Correlated signals and the Negatively Correlated signals should be different.
[0024] During playback, the left and right channel input signals L, R are supplied to the converter 13, where the M and S channel signals M, S are formed according to M = L + R, S = L - R. The mid channel signal M is then supplied to the filter 11, while the side channel signal S is supplied to the filter 12. The filtered outputs from each filter 11, 12 are each connected to the converters 14, 15, where left and right channel output signals Lout, Rout can be calculated according to Lout = 0.5x(Mfilter + Sfilter), Rout = 0.5x(Mfilter - Sfilter). The left and right channel output signals Lout, Rout are applied to the left and right speakers 3, 4 respectively, where they are applied to electro-acoustic transducers.
[0025] In an alternative embodiment, illustrated in figure 12, the filters Fpos and Fneg are created and applied in each speaker 103, 104 of an audio system 101. This is referred to as a "distributed" system. Also in this case, a central unit 2 has access to both left and right audio channels (L, R), and is connected to a left speaker 103 and a right speaker 104.
[0026] In this case, however, each loudspeaker 103, 104 will have a set of the filters Fpos 111 and Fneg 112, and also includes circuitry 110a, 110b for creating, storing and applying the filters Fpos 111 and Fneg 112. Each loudspeaker further includes a R/L-to-M/S converter 113, the M output being connected to the Fpos filter 111 and the S output being connected to the Fneg filter 112. Each loudspeaker 103, 104 is also connected to at least one microphone 109a, 109b, which provides input to the circuitry 110a, 110b.
[0027] The left speaker further includes a M/S-to-L converter 114, connected to the filtered outputs Mfilter, Sfilter of the filters 111, 112, and a transducer connected to the output Lout from the converter 114.
[0028] The right speaker further includes a M/S-to-R converter 115, connected to the filtered outputs Mfilter, Sfilter of the filters 111, 112, and a transducer connected to the output Rout from the converter 115.
[0029] The operation of the system according to this embodiment is similar to that described in relation to figure 11, except for where the various steps take place.
[0030] During an initiation phase, the circuitry 110a, 110b first interact to control the audio system to output correlated signals on each loudspeaker 103, 104. The resulting audio signals are received at the central unit by the microphone(s) 109, and supplied to each circuitry 110a, 110b. Each circuitry 110a, 110b then determines a first room response, referred to as a room response for correlated signals. The first room response is used to define and create the filter 111 in each circuitry 110a, 110b, independently of each other.
[0031] Secondly, the circuitry 110a, 110b controls the audio system to output negatively correlated signals on each loudspeaker 103, 104. Again, the resulting audio signals are received at the central unit by the microphone(s) 109, and supplied to each circuitry 110a, 110b. Each circuitry 110a, 110b then determines a second room response, referred to as a room response for negatively correlated signals. This second room response is used to define and create the filter 112 in each circuitry 110a, 110b, independently of each other.
[0032] During playback, the left and right channel input signals L, R are supplied to each loudspeaker 103, 104, where the converters 113 generate mid and side channel signals M, S according to M = L + R, S = L - R. The M channel signal M is then supplied to the filters 111, while the S channel signal S is supplied to the filters 112. The filtered outputs from each filter 111, 112 are each connected to the converters 114, 115, where left and right channel output signals Lout, Rout can be calculated according to Lout = 0.5x(Mfilter + Sfilter), Rout = 0.5x(Mfilter - Sfilter). The left and right channel output signals Lout, Rout are applied to the transducers in the left and right speakers 103, 104 respectively.
[0033] The person skilled in the art realizes that the present invention by no means is limited to the preferred embodiments described above. On the contrary, many modifications and variations are possible within the scope of the appended claims.
[0034] For example, the correlation of the measurements (both positive and negative) can be performed by summing and subtracting the time measurements (impulse responses) of the two loudspeakers. This can be done instead of playing the signals from both loudspeakers simultaneously. Both methods are equivalent.
[0035] Although this method is described for a two-channel system considering only one dimension of a listening room, it is applicable to systems with more than two loudspeakers and rooms with more than one dimension. The system can be applied to any pair of loudspeakers in a loudspeaker system that contains more than two loudspeakers.
[0036] In systems where a gain and delay can be applied to the loudspeaker outputs to compensate for incorrect listener placement, these compensation elements should be applied to the Left and Right input signals before the M/S filtering takes place. In this way, the effects of the mismatched gains and delays on the room's acoustical behavior will be compensated for in the filtering. Placing the gains and delays after the filter blocks will result in an incorrect effect. One recommended implementation is shown in Figure 13, showing the correct placement of the delay and gain used for listener placement compensation. Note that the Gain and Delay blocks are placed before the M/S Filtering stages.
receiving a left channel input signal and a right channel input signal;
transforming said left channel signal and said right channel signal to a mid channel signal and a side channel signal;
applying a mid channel filter (Fpos) to said mid channel signal to provide a filtered mid channel signal;
applying a side channel filter (Fneg) to said side channel signal to provide a filtered side channel signal;
retransforming said filtered mid channel signal and said filtered side channel signal to a left channel output and a right channel output;
applying said filtered left channel output to said first loudspeaker; and applying said filtered right channel output said second loudspeaker;
wherein said mid channel filter is created by:
applying correlated signals to said first and said second loudspeakers,
measuring a correlated modal response of said listening room,
creating said mid channel filter to have a modal response substantially equal to an inverse of said correlated modal response; and
wherein said side channel filter is created by:
applying negatively correlated signals to said first and said second loudspeakers,
measuring a negatively correlated modal response of said listening room, and
creating said side channel filter to have a modal response substantially equal to an inverse of said negatively correlated modal response.
measuring a correlated modal response of said listening room,
creating said mid channel filter to have a modal response substantially equal to an inverse of said correlated modal response,
measuring a negatively correlated modal response of said listening room, and
creating said mid channel filter to have a modal response substantially equal to an inverse of said negatively correlated modal response,
are performed by microphone(s) (9) and circuitry (10) in a central unit (2) connected to said first and second loudspeakers (3, 4).
measuring a correlated modal response of said listening room, and
creating said mid channel filter to have a modal response substantially equal to an inverse of said correlated modal response,
are performed by microphone(s) (109a) and circuitry (110a) in or connected to said first loudspeaker (103),
and wherein the steps of :
measuring a negatively correlated modal response of said listening room, and
creating said mid channel filter to have a modal response substantially equal to an inverse of said negatively correlated modal response,
are performed by microphone(s) (109b) and circuitry (110b) in or connected to said second loudspeaker (104).
a first and a second loudspeaker (3, 4; 103, 104);
circuitry (10) for applying, in a first sequence, correlated signals to said first and said second loudspeakers, and, in a second sequence, negatively correlated signals to said first and said second loudspeakers,
at least one microphone (109; 109a, 109b) for measuring, in said first sequence, a correlated modal response of said listening room, and, in said second sequence, a negatively correlated modal response of said listening room,
circuitry (10) for creating a mid channel filter (Fpos) with a modal response substantially equal to an inverse of said correlated modal response; and creating a side channel filter (Fneg) with a modal response substantially equal to an inverse of said negatively correlated modal response;
L/R-to-M/S converting circuitry (13; 113) for transforming a left channel signal and a right channel signal to a mid channel signal and a side channel signal;
circuitry (10; 110a, 110b) for applying said mid channel filter to said mid channel signal to provide a filtered side channel signal, and for applying a side channel filter to said side channel signal to provide a filtered side channel signal;
L/R-to-M/S converting circuitry (14, 15; 114, 115) for retransforming said filtered mid channel signal and said filtered side channel signal to a left channel output and a right channel output,
wherein said filtered left channel output is applied to said first loudspeaker (3; 103); and said filtered right channel output is applied to said second loudspeaker (4; 104);