I've been meaning to ask, have you tried applying a "strip silence" kind of process on a vox recording and then batch processing the resulting clips to normalise them to a certain RMS or LUFS value? This restricts any level changes to between the vocal phrases but levelling like this might not give ideal results... Maybe with enough fine-tuning of the normalisation process this could become a fully automated process with good results? Just an idea...


Alistair

I haven't, but I suspect I wouldn't actually want it fully normalized. That's why the plugin I proposed has a percentage match slider... then the other control for region sizing takes care of any overall sensitivity issues. Without those controls, I think you're right that it wouldn't work very well. Proposed plug also matches by peak, rms, or LUFS instead of peak. It's a guess, but don't think I'd use it in peak.

I've done very little with batch processing as I'm a create as I go type, and offline processes don't generally lend themselves well to that workflow. That being said, I consider getting vocals right to be very important, so ultimately whatever yields the best result wins.

Yeah I'm not really talking about external batch processing, that would kill my workflow too. In Cubase you can select a bunch of clips and apply the same processing to them, including plugins. There is no RMS normalise in Cubase (just peak) so I haven't tested this yet. I would have to test with one of the LUFS normalisers or another plugin that allows RMS or other types of normalisation over a whole clip. Maybe Melda have something that can do the job...

In Cubase you can build Macro's and "Logical Processes" so it shouldn't be too hard to make this whole process a one key command shortcut that separates the recording into clips using the "strip silence" process and then opens a "normalising plugin" and applies it to all the clips. (I've created similar one key shortcuts for other multi-step processes I do often). Maybe the DAW you use has similar functions?

Alistair

I haven't come across a batch function like that... then again, I haven't been looking for it.

It's worth a shot, though I suspect without controls for region size and matching strength there will still be quite a bit of manual adjustment required.

Too swamped to dig in at the moment. I stumbled across something 3 weeks ago nearly by accident that has me working around the clock ever since to develop into a new type of plugin. I should get back to vocals a week or so after completing that.

Finally recorded some new vocals yesterday, and looking to move quickly getting this track wrapped up.

Down to the wire getting my (extended) taxes squared away, so it may be a day or so before I can move on to leveling, but it's next up on the agenda.

Anyone used this process yet for vocal leveling?

You know, now that I'm thinking about it again, perhaps a tool like the MaxxVolume or MV2, might help here?? Perhaps Sonnox Inflator??? Not sure if this is the correct direction for you either.

MaxxVolume - Volume Leveler Plugin | Waves
Sonnox | Oxford Inflator v3

I've got some result with Sound Radix Powair (leveler only, I desactivate the compressor module) and have yet to try Nectar Auto Level Mode :
Automatically Set Vocal Levels with Nectar 3’s Auto Level Mode

Someone has compared these 2 plugs with Waves Rider and Melda for vocal auto leveling ?

I'm looking for an auto leveler too as manual editing is not an options for me (4 hours long interviews)... and I want consistent levels before hitting any compressor.

Philippe

I don't have Nectar 3 yet. Still not sure if I'll get that upgrade. I haver so many choices for processing now about the only thing I don't have is that dynamically adjusting eq. Even the auto-leveller could, in theory, be replaced with Vocal Rider.

So, anyway, I haven't managed to compare the two. I will say that Nectar's ALM seems to function quite well and transparently, but offers you very little in control, whereas Vocal Rider is much more customisable. Which is both a good and bad thing. because it not only offers you more ways to get a signal sounding rough, but more ways to get it wrong too!! Haha!!

I also haven't tried Nectar, so can't help with comparison.
However frankly I wasn't so impressed by PowAir personally, when I tested it some time ago. There was some initial sensation, but after few comparisons, even after some tweaking, it never beat my usual chains from separate plugins or normal procedures. But its always material dependent of course and I didn't feed it with something, where it really shines .

With regards to gain riding and processing of longform material like your mentioned interviews, I was very impressed by ease of use and results from TB Dynaride. It has dedicated profiles for detection algorithm like speech, vocals, bass etc. and there is also switchable look-ahead to overcome possible pops. It produces pretty good results out of box with some sensible setup for target level, range and noise floor. Like with similar tools, you can use it just as straightforward effect processor or write automation to your DAW for manual correction if you need that.
In combination with some other sensibly set quicker compressor, limiter or de-esser according to your material and needs.. you can achieve quite good results IMO/IME.

If you happen use Reaper as a DAW, there are also other options for those tasks at their script collection. There are some pretty good ones, which can do offline analysis of selected clips according to riding parameters and write clip based volume automation faster than real-time. Then you can again tweak it manually and combine it with other effects as necessary.

Michal

I have reently tested Powair on vocals. In theory it shuld be excellent because it uses LUFS (LKFS) to measure loudness. In practice it simply didn't work as well as Drum Leveller on vocals. Maybe it's just that I have learned how to use DL to its maximum output. IDK.

What's your settings on Drum Leveler?

PSP's new Twin-L is great for levelling (opto and VCA modes also), followed by Waves L2 for peaks.

Any limiting I do is in parallel to preserve some movement in the vocals. But I’ve gotten to the point where the parallel track and the “clean” track are the same level going to my vocal level aux. I feel like it’s the best of both worlds. Soft clipping vocals sounds better than limiting to me if it has to be directly on the insert.

Try TimP's new Opto 32 clipper - damn good on vocals. The updated compression is spot on also.

I’m dying to try his stuff but I can’t justify buying N4 right now. Yeah have my eye on dyn one and a few others first.

Twin-L is quickly becoming one of my favorite PSP plugins and dynamics processors for this exact reason. I don't use L2 though. I have started using things like analog channel ac101 or other clean preamp style plugins to transparently handle peaks followed by twin-l for leveling and I not only get quick results but also get dynamic range control with consistent character and very few audible time based compression artifacts. Twin-L works so well on drums that I've started using it during sound design. They are working on a zero latency series of plugins which I'm impatiently waiting for.

Yes

I want powair to work for me and I’ve got it close on a couple of vox tracks, but, never predictably. I agree with switching off the comp

You're one if the few people I've heard say it's not amazing. I've always felt that it couldn't be as good as I heard. Haven't tested it myself still but I'm curious now. Going to have to test it tomorrow.

I don’t find that it’s transparent... which is why I bought it... it’s not bad or anything... but I’m not finding that it’s doing something that I can’t get done with some old go-to’s

Pretty much what I figured. I always find our couldn't be better than Pro-C2 or Compassion. Going to check it out later

Accusonus Era Voice Leveler doesn't work correctly. It detects aahh and ohhh as tremolo.