audio project in mind, could it be done? need your ideas/suggestions

[M6YRU]

right, it might sound a bit complicated but here goes...
i need to record from a low audio quality mono-source for like 7 hours...
in that 7 hours i need to be able to push a button and have 10 sec before the button push and 20 sec after with the same button to be able to be pushed and time reset to finish 20 sec after the last press...
now, i need a way to have the audio seperated somehow when that button is pushed, like saved as a seperate audio file or same file but recorded on other audio channel as a stereo file perhaps...
now the 7 hour recording is not 100% needed but would be nice to have, the audio before and after the button push is what is needed.

i know i will need a usb audio dongle for mic/line input.....

what can be done? whats the simplest thing.... at the moment i been using a dictaphone but its really hard to find moments in that recording that is needed even with audio software

[M6YRU]

ok so im also on the FB group and have come up with some suggestions and problems..
its not going to be easily viable to do what i was thinking due to processor, memory limitations etc espically if its recorded as a raw file (WAV) where that could be a very huge file!

so my next idea...
ok, can the Pi3 easily act as an audio buffer?
ie, have a long delay between input vs output?
if this is the case i could easily do what i want with the dictaphone recorders, have one to record input and another one to record a switched output with VOX enabled.. so now that switch will allow audio to pass through to the recorder that will be delayed by the Pi.....

[ghans]

Why has it to be raw WAV files ? My calculations say that you need a ring buffer of 2.6 MB @ 30s with WAV, any Pi should be able to manage that easily ?

ghans

[M6YRU]

yes someone mentioned doing a ring buffer (from jack) for continious live in n out (i think) and with that i can send the selected delayed audio to the recording dictaphone...
i am also looking to do this electronically but not sure on quality of audio output...

[crofter]

hi
pi audio quality is very good

crofter

[knute]

ok so im also on the FB group and have come up with some suggestions and problems..
its not going to be easily viable to do what i was thinking due to processor, memory limitations etc espically if its recorded as a raw file (WAV) where that could be a very huge file!

You could use some compression to the audio such as uLaw or aLaw to get about a 2 to 1 ratio over direct PCM audio. GSM is about 10 to 1.

8000 samples per second, 2 bytes per sample, 7 hours is approximately 403MB of storage without any compression. You probably can't tell the difference with uLaw or aLaw compression depending on what the source is. GSM has more loss of fidelity.

Depending on your source you could sample at a lower rate and take only 1 byte samples.

I don't think you would have any problem doing this with a Pi3.