Filtering an audio signal for a non-real-time recipient

1. An apparatus comprising:

a processor;

a memory that stores code executable by the processor to:

detect communication of an audio signal to a first primary speaker

filter the audio signal with a speaker identification filter to preserve a first primary speaker audio signal;

detect a change from the first primary speaker to a second primary speaker from one or more of a change in relative amplitudes of the first primary speaker audio signal and a second primary speaker audio signal, a motion, and a specified change phrase; and

modify the speaker identification audio filter to a preserve the second primary speaker audio signal.

2. The apparatus of claim 1 , wherein the code is further executable by the processor to:

detect communication of the audio signal to a non-real-time recipient, wherein the non-real-time recipient is one of a voicemail account, a messaging system, a speech recognition system, and an automated phone menu; and

filter the audio signal with a high-latency audio filter in response to detecting the non-real-time recipient, wherein a combined latency of the high-latency audio filter and a path to the non-real-time recipient is less than a latency threshold.

3. The apparatus of claim 2 , wherein the speaker identification audio filter that preserves one or more primary speaker audio signals and removes other audio signals.

4. A method comprising:

detecting, by use of a processor, communication of an audio signal to a first primary speaker

filtering the audio signal with a speaker identification filter to preserve a first primary speaker audio signal;

detecting a change from the first primary speaker to a second primary speaker from one or more of a change in relative amplitudes of the first primary speaker audio signal and a second primary speaker audio signal, a motion, and a specified change phrase; and

modifying the speaker identification audio filter to a preserve the second primary speaker audio signal.

5. The method of claim 4 , wherein the method further comprises:

detecting communication of the audio signal to a non-real-time recipient, wherein the non-real-time recipient is one of a voicemail account, a messaging system, a speech recognition system, and an automated phone menu; and

filtering the audio signal with a high-latency audio filter in response to detecting the non-real-time recipient, wherein a combined latency of the high-latency audio filter and a path to the non-real-time recipient is less than a latency threshold.

6. The method of claim 5 , wherein the low-latency audio filter has a latency of less than 50 milliseconds.

7. The method of claim 5 , wherein the non-real-time recipient is detected from a specified automation phrase.

8. The method of claim 5 , wherein the speaker identification audio filter that preserves one or more primary speaker audio signals and removes other audio signals.

9. The method of claim 8 , wherein the one or more primary speaker audio signals are identified from one or more of phonemes, prosodic attributes, and frequency profiles.

10. The method of claim 5 , wherein the high-latency audio filter has a latency of greater than 200 milliseconds for a voicemail account, messaging system, and a speech recognition system.

11. The method of claim 5 , wherein the high-latency audio filter has a latency in the range of 50-200 milliseconds for an automated phone menu.

12. A program product comprising a non-transitory computer readable storage medium storing code executable by a processor to perform:

detecting communication of an audio signal to a first primary speaker

filtering the audio signal with a speaker identification filter to preserve a first primary speaker audio signal;

detecting a change from the first primary speaker to a second primary speaker from one or more of a change in relative amplitudes of the first primary speaker audio signal and a second primary speaker audio signal, a motion, and a specified change phrase; and

modifying the speaker identification audio filter to a preserve the second primary speaker audio signal.

13. The program product of claim 12 , wherein the processor further performs:

detecting communication of the audio signal to a non-real-time recipient, wherein the non-real-time recipient is one of a voicemail account, a messaging system, a speech recognition system, and an automated phone menu; and

filtering the audio signal with a high-latency audio filter in response to detecting the non-real-time recipient, wherein a combined latency of the high-latency audio filter and a path to the non-real-time recipient is less than a latency threshold.