VSE Sound Strip volume in Decibels (and Speaker)
Right now the volume is a float value which is somewhat arbitrary when it comes to audio volume, it would be very hand if it was in decibels.
This would also apple to the Speaker object.
Right now the volume is a float value which is somewhat arbitrary when it comes to audio volume, it would be very hand if it was in decibels.
This would also apple to the Speaker object.
6 Comments
That's simply not how sound works, you don't set some decibel value and your audio magically gets there. For various reasons, the first one being that every sound is a variation of waves, not a constant value, it varies through time as well, and each different input is yet another level.
How we manage levels within any sound mastering is by using factors. We simply multiply the input sound by a value, which itself can be varied through time as we wish, and we also might use some filters to fine tune the volume in many ways, may it be compressors, EQ, limiters, expanders, normalizers, and so on.
What could be useful to Blender would be to have audio monitoring, at least a master dBFS would be mandatory. So the sound master can "see" what he sends out and adjust to its needs.
Decibels are not an absolute metric; they're always relative to something. They can be relative to the softest sound you can still hear, or relative to the loudest volume the system can produce. The latter is what some amplifiers use, and as a result there are only negative decibels on the volume knob. But then again that's already what the float value indicates: a percentage between "silent" and "the loudest possible".
Which dB scale are you proposing, with which reference level?
The current standard would be to have at least a dBFS monitor somewhere in the video editors to get a measurement of the output electric signals.
{attachment 2020-09-24_21-37-54png}
But even dBFS isn't always the best thing to measure, we could also think about LUFS ...
* @-L0Lock- * This add-on includes a VU-meter for the VSE: github.com/snuq/VSEQF
* @ tintwotin * that's a start
* @-L0Lock- * Apparently, is it very complicated to implement the stuff for doing this in audaspace, which is the library used by Blender for audio, so the way it is done in this add-on is properly all we're going to get.