Guidance on Audio Filters/Loudness

[gdpodesta]

How does one use loudness and/or filters to obtain a video that has the very best best audio possible, with almost the worst video?

I realize that its' an odd question, but we send some of our videos for transcription, and obtaining the cleanest audio is the only real concern.
Googling this for ffmpeg gives me all sorts of overall quality answers that always include video quality discussion, so I'm confused about how to focus only on the audio, and then, how to accomplish it with FFAStrans.

Any guidance is appreciated. Suggestions on what our video production staff could do with Adobe Premiere and Adobe Media Encoder would also be appreciated - I'm a software guy, and can't speak their language too well, never mind offer suggestions, lol.

[admin]

Hi gdpodesta,

Yes, it's a bit odd question Could you please elaborate what you mean with best possible audio? In regards to loudness conforming all it does is to make sure audio is corrected to the desired output volume. It should have very little influence on audio quality, if that is your concern. On the other hand, with best audio do you actually mean fixing/restoring bad audio? If so FFAStrans has no inbuilt filters to do this.

But again, if you could give some more details on what you expect or wish for it's easier to help.

-steinar

[gdpodesta]

Ah, my admitted encoding ignorance is showing, lol. The videos are submitted to a transcription service, which appies AI speech-to-text technology to extract transcriptions and captions. The audio quality/clarity/consistency is what I'm trying to optimize, and keep video quality at a minimum (for uploading time). The videos are of meetings that last 4-6 hours, which result in gigantic files if encoded for typical video streaming. We could submit a video with 100x100 video resolution and get the same result as submitting a 4k video, as long as the audio was the same.

I was hoping that FFAStrans had some including filtering, etc. that I simply don't understand, but hoped that the encoding expertise here might be available as well Additionally, I can't determine what overall compression settings would also inadvertly affect audio quality. My thinking is to sett everything to the highest possible quality, video AND audio, but the video resolution to something ridiculously small.

[seePyou]

I hope this doesn't count as resurrecting a zombie thread, but I see no answer here.

So, gdpodesta, why not just choose a very low resolution for the video, a very fast transcode option, and a low video bitrate? Then leave all the sound options CD quality. Stereo (if that applies, although if the source is a conferance mic, I guess you only have a mono track?) and then depending on which audio compatibility issue with your AI service, select the codec and give it ample data rate. If MP3 then go for 192kbps, for example.

[emcodem]

@seePyou hehe we don't complain about thread resurrections
Honestly i didnt really understand the question as the best audio is the original. Probably all he wanted to do is to use PCM "encoding"?