Is there a way to select entire track but amplify only regions below a certain DB threshold?

I’ve been looking for a way to automatically amplify the quiet parts to reach the highest peak of that wav format. This would save a lot of time of manual volume automation in vocals where there’s very quiet singing (think billie eilish whispers type) and loud singing.

Other “automated” solutions like slapping a huge limiter (like izotope maximizer or whatever)/compressor on the track to bring the peaks down to the level of the whispers so that the whispers sound “loud” doesn’t work because it completely messes up the actual loud parts (they sound super compressed).

Is there anything out there (whether it’s a cubase effect or not) that does what I’m describing? Amplifies all parts below a certain DB level to let’s say the highest DB found in the track or the highest possible DB before clipping? And when I say parts I don’t mean simply amplify each 5 second phrase until the peak of that phrase reaches the real peak of the track. I mean milisecond by milisecond the wav is evaluated against the highest peak so that the entire wave file looks like a completely flat line touching the top of the db limit (kinda like what happens if you use loudness normalization at 0.0 LUFS - but without unnecessarily amplifying the already loud parts and causing ugly clipping obviously)

What you are looking for is compression…
Waves MV2 is usually good for that. But maybe Waves Vocal Rider might be suite that too.

@cubace I specifically mentioned that compression is exactly NOT what I’m looking for and explained why. Already tried tools like MV2. Yes, they bring all levels to the same level with the right configuration, but the parts that were loud in the first place get compressed as hell to prevent clipping. What I’m looking for should leave loud parts untouched. It should only bring up the quieter parts.

The Vocal Rider suggestion you made was closer to what I’m looking for, but unfortunately still does not allow enough customization to equalize the volume of parts that have a huge DB difference (i’m talking about 25dB difference between the quietest and the near peak part). And even on the “fast” setting, the slider takes way too long to go to the original default position.

Too bad. Vocal rider only do 24 dB. I think you might look in to preproduction software like Izotope RX. And millisecond evaluations is fast for compressors. 1 ms is equal one period of 1kHz.

@cubace I need something that will work live as well so “noise removal” solutions (typically highly demanding) probably will mess up with latency.

I actually tried the bass rider version of the waves plugin and it moves sliders faster since it was designed for a bass. Starting to get somewhere by lining up more than 1 bass rider on the same track

Have you looked into upward compression? Squasher might be helpful. OTT by Xfer has less latency and is for free.
However, both are quite drastically changing the sound.

Fabfilter Pro-MB can do upward compression.