Any good VST Compressor for Youtube volume normalizing?

[carrieres]

maybe this :
https://www.hornetplugins.com/plugins/h ... ormalizer/

[mikehende]

If you're not willing to work out how these things work and how to get the best from them using that learned knowledge, I'm not sure how other people can help. Why would any default preset be the instant fix? I've never experienced that, ever.

I'll say it here: no single plugin will be able to do this for you in the perfect way you imagine. None. You will always get distortions, peaks where the volume management system didn't kick in quick enough, or clamps down too hard and ruins the explosion or bang or something.

Well, here's the thing. Since there are volume softwares which normalize mp3's and without having to do any settings adjustments, I was expecting the same with youtube audio and I am surprised that is not the case.

[jules99]

Perfect compressor: https://mixingnightaudio.com/lolcomp/

[CinningBao]

If you're not willing to work out how these things work and how to get the best from them using that learned knowledge, I'm not sure how other people can help. Why would any default preset be the instant fix? I've never experienced that, ever.

I'll say it here: no single plugin will be able to do this for you in the perfect way you imagine. None. You will always get distortions, peaks where the volume management system didn't kick in quick enough, or clamps down too hard and ruins the explosion or bang or something.

Well, here's the thing. Since there are volume softwares which normalize mp3's and without having to do any settings adjustments, I was expecting the same with youtube audio and I am surprised that is not the case.

I think you slightly misunderstand how these volumes are calculated.

Software which normalises MP3's or WAVs or whatever has the opportunity to scan the file for the loudest sample (as in samples per second, where you'll have at least 44,100 of them). So if the loudest sample is 0.8, the program multiplies the volume by 1.25 to get that 0.8 sample at 1. Normalisation is a relatively easy 'offline' operation .

If you're running live audio into a volume managing function (a compressor, say) it only has the amount of time the envelope is listening for to adjust the volume for you. Envelope detectors (the bit in the compressor which is listening to the volume of the input) will only be able to get an average for, at most, 20 seconds? Most compressors have a window of somewhere between 0ms and a second or 2 (2000ms). The compressor will only be managing the volume of the last, at most, 2 seconds of audio. And even then, the compressor will only be responding to the big energy jumps, not to fast transients like drums or vocal plosives.

Now, if there was a way to acquire the LUFS level of a yt file without playing it, you could extract that and just boost/cut your audio to compensate. But I don't think that data is available, so that's not possible.

It sounds like you're slowly realising your original plan is unreachable, without extra data from the source material. It's basically to do with how volumes are calculated 'in the moment / real-time' and 'offline' (could probably find better terminology but they'll suit for here). Offline can do loads of things real-time operations can't, including scanning files for energetic envelopes, and looking for particular digital information, like the loudest sample. In some ways, offline processing can be much better than real-time because the tools can understand the whole material, where as real-time tools which are 'listening' only understand the last few seconds,at most, of what happened.

But, again, none of that really helps you find the solution you're looking for because, well, I don't think it exists in the way you want it. APU Loudness Compressor and the Hornet equivalent Normalizer do basically do what you want (to target all the sound at a certain volume) but they have their limits because they work in real-time.

Does that make any sense?

[mikehende]

Starting to make things clearer 'far as how com/limiters work', thanks for explaining.

There is an mp3 software Breakaway Audio Enhancer which works great with normalizing mp3's 'on the fly' without having to do anything, I have been using it since years now for volume levelling when running my mp3 playlists [outside of Reaper] with my software mp3 players. But doing some testing on a laptop yesterday running a playlist of youtube music videos I was surprised that the levelling doesn't work with the BAE.

See, my thinking is audio is audio no matter the source of it why I am having trouble understanding why the problem with audio from youtube music videos compared to regular mp3's.

[thecontrolcentre]

You could try ADPTR Audio Streamliner. Not sure if it works in real time ...

https://youtu.be/HOeHrCPe2jc

[CinningBao]

Starting to make things clearer 'far as how com/limiters work', thanks for explaining.

There is an mp3 software Breakaway Audio Enhancer which works great with normalizing mp3's 'on the fly' without having to do anything, I have been using it since years now for volume levelling when running my mp3 playlists [outside of Reaper] with my software mp3 players. But doing some testing on a laptop yesterday running a playlist of youtube music videos I was surprised that the levelling doesn't work with the BAE.

See, my thinking is audio is audio no matter the source of it why I am having trouble understanding why the problem with audio from youtube music videos compared to regular mp3's.

I've had a look at the BAE product and although I can't tell if you load files into it or just stream through the algorithm, it's actually doing similar things to what radio stations do to characterise and 'brand' their output. Radio stations tend to use hardware like https://www.orban.com/radio-audio-processors to keep the volumes 'good' and the sonic fingerprint consistent, and they do it with a load of multi-band compressor, limiters, clippers, sub-harmonics generation and more. Sometimes up to 8-9 bands of compressing and limiting. And they have a load of auto-gaining stuff keeping the levels steady.

https://www.thimeo.com/stereo-tool/

There is a company which make a software processor which does a pretty good job of emulating that hardware experience in-the-box so it might be worth looking at that.

https://www.thimeo.com/stereo-tool/buy/

The basic license, with tools which _may_ get you where you want, is £35 and you can add features as you think you might need them.

Maybe give it a shot.

[mikehende]

Thanks and yes you are correct BAE is used mainly for Broadcasting type situations. When you install it you route your pc's headphone out through it so all continuous playlists audio is processed through it on the fly. It's mainly used when running music playlists through broadcast, DJ and software mp3 players.

Seems you are on the right track as it specifically mentions 'youtube':

Feeding Stereo Tool from other software on your computer
If you want to feed Stereo Tool from a program that doesn't support audio processing plugins, or for example when you just want to send all the audio (eg. YouTube video's) through Stereo Tool,

I had tried Stereo Tool years ago but for Reaper and their Support is great. I will ask around on their forum and can give it a try again this time outside of Reaper to see if it can work for this purpose, will report back here.

[camsr]

If the audio was delayed, it could be look-ahead compressed to avoid spurious compressor transients from being generated. But that audio would be delayed against the video, so for the correct result the video would need delay too.

Youtube publishing amplitude stats would be very useful, cutting out much redundant work.

[mikehende]

which does a pretty good job of emulating that hardware experience in-the-box so it might be worth looking at that.

https://www.thimeo.com/stereo-tool/buy/

The basic license, with tools which _may_ get you where you want, is £35 and you can add features as you think you might need them.

Maybe give it a shot.

So I have been experimenting with ST with the help of some guys on their forum but 2 problems, latency is a big issue as the voice does not sync with the lip movements and it does nothing to help with the volume normalizing. When I had asked if ST's upgraded features which is very expensive would solve these 2 problems I was told this by an experienced ST user:

"No, the license will help with quality and with the right preset you will have even volume without distortion, but the latency/sync, on the other hand, will always be there on any video and that the latency is most likely caused by having to use the VB cables."

So bottom line is it might be best to simply ditch this and any other problem videos from my playlists to avoid all of these problems. If there was a way to save the the video at a specified volume level this would have solved this problem for the most part.

[Funky40]

Honestly, I think APU Loudness Compressor can do what you need, .....

at least to some degree.....
i´d also pointed towards the APU.

there is also the playfair audio dynamic grading.
It has more nodes.

bought mine during a sale.
https://playfair-audio.com/

It could be that it might help to run several such devices behind each other, with painstakingly setting up the settings. stage by stage.