Room-informed binaural rendering

1 . A system comprising:

at least one processor; and

at least one non-transitory computer-readable medium comprising program instructions that are executable by the at least one processor such that the system is configured to:

play back first audio via first audio transducers of a playback device that is located at a first location in an acoustic environment;

capture, via one or more microphones of a headphone device while the headphone device in the acoustic environment, second audio representing playback of the first audio in the acoustic environment;

determine target data from the captured second audio, the target data comprising target room parameters;

adapt, based on the target room parameters, reference binaural rendering impulse response filters to an early reflections model representing reflections of the sound in the acoustic environment before a particular mixing time;

generate, from the target room parameters, a late reverberation model representing reverberation of the sound in the acoustic environment after the particular mixing time;

synthesize a set of binaural rendering filters comprising a direct sound model, the determined early reflections model, and the determined late reverberation model, the direct sound model based on reference head-related impulse response data;

configure a binaural renderer with the synthesized set of binaural rendering filters;

render, via the configured binaural renderer, third audio from audio input signals, wherein the rendered third audio is configured to simulate playback from virtual sources within the acoustic environment when played back via the headphone device, wherein the virtual sources include a first virtual source at the first location and one or more second virtual sources at respective second locations; and

cause the headphone device to play back the rendered third audio via second audio transducers to simulate playback from the virtual sources.

2 . The system of claim 1 , wherein the target room parameters comprise target early decay time, and wherein the program instructions that are executable by the at least one processor such that the system is configured to adapt the reference binaural rendering impulse response filters to the early reflections model comprise program instructions that are executable by the at least one processor such that the system is configured to:

modify the reference binaural rendering impulse response filters by a gain envelope that converts a reference early decay time to the target early decay time.

3 . The system of claim 2 , wherein the program instructions that are executable by the at least one processor such that the system is configured to determine the target data from the captured second audio comprise program instructions that are executable by the at least one processor such that the system is configured to:

estimate the target early decay time from energy delay curves in respective subbands of a room impulse response.

4 . The system of claim 1 , wherein the at least one non-transitory computer readable medium further comprises program instructions that are executable by the at least one processor such that the system is configured to:

select the reference binaural rendering impulse response filters to match the acoustic environment from among a plurality of filters representing different acoustic environments.

5 . The system of claim 1 , wherein the target room parameters comprise frequency-dependent late reverberation time, and wherein the program instructions that are executable by the at least one processor such that the system is configured to generate the late reverberation model comprise program instructions that are executable by the at least one processor such that the system is configured to:

generate static late reverberation that matches the frequency-dependent late reverberation time.

6 . The system of claim 1 , wherein the room parameters comprise a target mixing-time energy spectrum, and wherein the program instructions that are executable by the at least one processor such that the system is configured to synthesize the set of binaural rendering filters comprise program instructions that are executable by the at least one processor such that the system is configured to:

adjust per subband gain in the determined early reflections model from a reference mixing-time energy spectrum to match the target mixing-time energy spectrum; and

adjust per subband gain in the determined late reverberation model from a native mixing-time energy spectrum to match the target mixing-time energy spectrum.

7 . The system of claim 1 , wherein the room parameters comprise a target direct-to-early ratio representing an energy ratio between direct sound and early reflections in a room impulse response, and wherein the program instructions that are executable by the at least one processor such that the system is configured to synthesize the set of binaural rendering filters comprise program instructions that are executable by the at least one processor such that the system is configured to:

estimate an intermediate direct-to-early ratio representing an energy ratio between mean direct energy in the direct sound model and mean early reflections energy in the early reflections model;

adjust gain in the direct sound model from the intermediate direct-to-early ratio to match the target direct-to-early ratio; and

adjust gain in the determined early reflections model from the intermediate direct-to-early ratio to match the target direct-to-early ratio.

8 . The system of claim 1 , wherein the room parameters comprise a center-time of a room impulse response, and wherein the program instructions that are executable by the at least one processor such that the system is configured to synthesize the set of binaural rendering filters comprise program instructions that are executable by the at least one processor such that the system is configured to:

determine a gain that when applied to the early reflections model would cause the early reflections model to match the center-time of the room impulse response; and

apply the determined gain to the early reflections model.

9 . The system of claim 1 , wherein the program instructions that are executable by the at least one processor such that the system is configured to synthesize the set of binaural rendering filters comprise program instructions that are executable by the at least one processor such that the system is configured to:

normalize the direct sound model, the early reflections model, and the late reverberation model.

10 . The system of claim 1 , wherein the at least one non-transitory computer readable medium further comprises program instructions that are executable by the at least one processor such that the system is configured to:

send, via a network interface of the playback device, data representing the second audio;

receive, via a network interface of the playback device, the data representing the second audio; and

determine a target room impulse response from the received data representing the second audio; and

determine the target room parameters from the determined target room impulse response.

11 . The system of claim 1 , wherein the headphone device comprises a first microphone and a second microphone, wherein the captured first audio comprises left audio captured by the first microphone and right audio captured by the second microphone, and wherein the program instructions that are executable by the at least one processor such that the system is configured to determine a room impulse response comprise program instructions that are executable by the at least one processor such that the system is configured to:

before determination of the room parameters, combine a first impulse response representing the left audio with a second impulse response representing the right audio.

12 . The system of claim 1 , wherein the playback device comprises a high-definition multimedia interface (HDMI) port, and wherein the at least one non-transitory computer readable medium further comprises program instructions that are executable by the at least one processor such that the system is configured to:

receive, via the HDMI port from a television, the audio input signals.

13 . A playback device comprising:

a network interface;

at least one processor; and

at least one non-transitory computer-readable medium comprising program instructions that are executable by the at least one processor such that the playback device is configured to:

play back first audio via first audio transducers of the playback device while located at a first location in an acoustic environment;

cause a headphone device to capture, via one or more microphones while the headphone device in the acoustic environment, second audio representing playback of the first audio in the acoustic environment;

determine target data from the captured second audio, the target data comprising target room parameters;

adapt, based on the target room parameters, reference binaural rendering impulse response filters to an early reflections model representing reflections of the sound in the acoustic environment before a particular mixing time;

generate, from the target room parameters, a late reverberation model representing reverberation of the sound in the acoustic environment after the particular mixing time;

synthesize a set of binaural rendering filters comprising a direct sound model, the determined early reflections model, and the determined late reverberation model, the direct sound model based on reference head-related impulse response data;

configure a binaural renderer with the synthesized set of binaural rendering filters;

render, via the configured binaural renderer, third audio from audio input signals, wherein the rendered third audio is configured to simulate playback from virtual sources within the acoustic environment when played back via the headphone device, wherein the virtual sources include a first virtual source at the first location and one or more second virtual sources at respective second locations; and

cause the headphone device to play back the rendered third audio via second audio transducers to simulate playback from the virtual sources.

14 . The playback device of claim 13 , wherein the target room parameters comprise target early decay time, and wherein the program instructions that are executable by the at least one processor such that the playback device is configured to adapt the reference binaural rendering impulse response filters to the early reflections model comprise program instructions that are executable by the at least one processor such that the playback device is configured to:

modify the reference binaural rendering impulse response filters by a gain envelope that converts a reference early decay time to the target early decay time.

15 . The playback device of claim 14 , wherein the program instructions that are executable by the at least one processor such that the playback device is configured to determine the target data from the captured second audio comprise program instructions that are executable by the at least one processor such that the playback device is configured to:

estimate the target early decay time from energy delay curves in respective subbands of a room impulse response.

16 . The playback device of claim 13 , wherein the at least one non-transitory computer readable medium further comprises program instructions that are executable by the at least one processor such that the playback device is configured to:

select the reference binaural rendering impulse response filters to match the acoustic environment from among a plurality of filters representing different acoustic environments.

17 . The playback device of claim 13 , wherein the target room parameters comprise frequency-dependent late reverberation time, and wherein the program instructions that are executable by the at least one processor such that the playback device is configured to generate the late reverberation model comprise program instructions that are executable by the at least one processor such that the playback device is configured to:

generate static late reverberation that matches the frequency-dependent late reverberation time.

18 . The playback device of claim 13 , wherein the room parameters comprise a target mixing-time energy spectrum, and wherein the program instructions that are executable by the at least one processor such that the playback device is configured to synthesize the set of binaural rendering filters comprise program instructions that are executable by the at least one processor such that the playback device is configured to:

adjust per subband gain in the determined early reflections model from a reference mixing-time energy spectrum to match the target mixing-time energy spectrum; and

adjust per subband gain in the determined late reverberation model from a native mixing-time energy spectrum to match the target mixing-time energy spectrum.

19 . The playback device of claim 13 , wherein the room parameters comprise a center-time of a room impulse response, and wherein the program instructions that are executable by the at least one processor such that the playback device is configured to synthesize the set of binaural rendering filters comprise program instructions that are executable by the at least one processor such that the playback device is configured to:

determine a gain that when applied to the early reflections model would cause the early reflections model to match the center-time of the room impulse response; and

apply the determined gain to the early reflections model.

20 . At least one non-transitory computer-readable medium comprising program instructions that are executable by at least one processor such that a system is configured to:

play back first audio via first audio transducers of a playback device that is located at a first location in an acoustic environment;

capture, via one or more microphones of a headphone device while the headphone device in the acoustic environment, second audio representing playback of the first audio in the acoustic environment;

determine target data from the captured second audio, the target data comprising target room parameters;

adapt, based on the target room parameters, reference binaural rendering impulse response filters to an early reflections model representing reflections of the sound in the acoustic environment before a particular mixing time;

generate, from the target room parameters, a late reverberation model representing reverberation of the sound in the acoustic environment after the particular mixing time;

synthesize a set of binaural rendering filters comprising a direct sound model, the determined early reflections model, and the determined late reverberation model, the direct sound model based on reference head-related impulse response data;

configure a binaural renderer with the synthesized set of binaural rendering filters;

render, via the configured binaural renderer, third audio from audio input signals, wherein the rendered third audio is configured to simulate playback from virtual sources within the acoustic environment when played back via the headphone device, wherein the virtual sources include a first virtual source at the first location and one or more second virtual sources at respective second locations; and

cause the headphone device to play back the rendered third audio via second audio transducers to simulate playback from the virtual sources.