Surfing YouTube and other video sites, it’s obvious that the audio associated with the various clips varies a lot. In particular, the volume can be anything from barely audible to obnoxious.
It occurred to me that since this audio is at heart a digital signal, it should be possible for a PC app to normalize it, so what you hear is of a more or less constant volume (that you can select).
I also would like something like that. I have my audio levels on my PC, media player, and speakers at a nice median level, so having to adjust one or the other for a one-off clip on YouTube is inconvenient. I’d rather see something separate I can adjust just for those, sort of like an additional “boost volume” feature.
I wondered if it was possible for someone to write something like an extension for Firefox or Greasemonkey, but it may be beyond their reach.
I think the algorithm would have to be fairly intelligent as far as what’s wanted and unwanted audio. Some videos have unwanted background noise, that would be amplified to a normal level. Some videos would have shouting or explosions, that would be quieted to a normal level. A loud/quiet/loud type song would suffer by being the same volume level throughout.
Perhaps the entire audio track could be taken into consideration and the volume level of the entire track adjusted a set amount, but web videos and therefore the audio aren’t all loaded into memory before they start playing so the algorithm would have incomplete data.
Incomplete, sure - but most audio tracks are fairly consistent throughout, so even a short sample should allow software to do a good job.
In the uncommon case where the track later becomes much quieter or louder, further adjustment is possible - or the user can do this manually (which is what happens now - rather often).
It’s probably dynamic range compression you’re looking for. Wiki link
There are links to various plug-ins at the bottom of that article. Don’t know if it will solve your problem with YouTube, but it might be a starting point.