Is there a way to normalize across sentences? For the example file, there are 3 sentences separated by gaps (effective silence). However, the middle sentence is at a lower volume. For reference, the picture is in “Waveform (DB)” mode). When I use the “RMS Normalize” effect on the entire clip, it doesn’t help in this case. The issue is that someone can be listening and the volume changes between sentences.
If you only have a few sentences you could manually select each one,
including with as little* silence as possible at the ends,
then RMS normalize each one individually.
[* a ~100ms of silence either end is close enough ]
If you have a lot of audio to process you’ll need a tool which detects the level & adjusts it automatically:
e.g. a compressor, or a leveller.
Thanks a bunch! This seemed to work OK in small scale tests.
If I have already applied a RMS normalize (then a limiter) to meet ACX standards, can I then apply the level speed, RMS normalize, then limiter (the last two would be repeated). Or would this have the potential to mess the audio up? The issue is that I applied the RMS normalize and limiter, then did a whole lot of editing, but now I want to apply the Level Speech. After that effect, it seems I need to do the RMS normalize again (but maybe not the limiter again).
can I then apply the level speed, RMS normalize, then limiter (the last two would be repeated). Or would this have the potential to mess the audio up?
RMS Normalize is simply a volume adjustment so it’s pretty-safe (and reversible) as long as you don’t push the peaks into clipping. Limiting pushes the peaks & RMS level closer together so it’s a little more “dangerous” and generally it’s irreversible.
And, I’d “be careful” with sentence-by-sentence adjustments because sudden jumps in volume, including sudden jumps in background noise, can be distracting and unnatural.
Gotcha, yeah the objective would be to apply this process (the to the entire audio file (20-50 min each) and to about 15 of them (its for a course). I want to have it sound consistent all the way through. The Speech Level effect should help with that. I just wish I knew about that before I applied the limiter and did a whole bunch of edits (100 hours worth so no going back now). But since there is not clipping, maybe I can get away with applying the Speech Level now still.
Regarding background noise, there is basically none. I built a really great sound booth and sound room and have pretty good equipment. You can see from the attached file that the sentences are very clean. I can put the headphones at max and play the “silent” section (which is not true silence) and not hear anything.