MAutoVolume reaction time?

[D2sX9ek8w3]

I tried MAutoVolume today and came away with the impression that it was more like and effect than a production tool.

I tried many settings and never got it too react slow enough to provide the sort of results I would strive for if I was hand editing an envelope.

I tried using look ahead at 5%, 10%, 50% and 100%.

I tried Sensitivity from 0% through 100% and felt like anything more than 50% produced a tremolo effect, but anything less than 70% didn't alter the sound levels enough to appreciate that the sound level was being averaged.

I tried the Speed at 0%, 10%, 50% and 100%. It seemed like 0% wasn't slow enough for what I wanted to do, so all the rest of the settings seemed to fast.

I was hoping for something that would use its look ahead capability to act smartly and see that one measure was loud and another was quiet and then even them out with a smooth level adjust, but instead what I experienced was a very fast acting and constantly changing level that sounded like a tremolo effect.

I don't think this tool will do what I hoped it would do, but I'm wondering if anyone has some ideas about how to use it to work slowly and stably on longer passages.

Thanks.

[MeldaProduction]

Well, it depends on the signal of course and hard to give you any answer without having the actual audio. Anyway the forthcoming update will have a Hold switch, which implements a true hold feature, which along with the look-ahead could do exactly what you want. So please just wait for the update and then let me know if it helped.

[D2sX9ek8w3]

Thanks, I will try it again when you add the hold feature.

[alexis1]

The melda vid sounded like it worked well!, To the OP mike_mccue, was your source audio more dynamic than the ones shown in the video?

[alexis1]

Addendum to above:

OK, I tried the demo, and I couldn't get it to work like in the video either. I see what the OP mike_mccune is saying, out of the box there is a tremolo that I agree can't be used for much of anything but creative effects. Like he did, I tried it under all kinds of conditions of sensitivity and speed, plus toggling the new "Hold" and the "Ultra Slow" buttons, but it remained a no-go because of the tremolo.

(As an aside, I'm guessing an updated manual is on the way, but mine doesn't describe "Ultra Slow" ... and it mentions both a "Peak Hold" and a "True Hold in advanced detector settings if available", but doesn't distinguish between them or say if the "Hold" function in the "General" setting (not the "advanced detector settings") is the same thing. I imagine this will be explained with the next manual version!).

HOWEVER I tracked the problem down to the fact that the automation is being written by MAutoVolume out of sync with the audio. It was always earlier than the audio. By manually dragging all the automation points en mass to the right position, all the tremolo was eliminated . It did a pretty good job of smoothing out the volume after that. I haven't used any other products like this (Quiet Art Wave Rider, Waves Vocal Rider), but I imagine there is some manual tweaking of automation needed with all of them.



Even when I hit the plugin hard (like putting in test values of 48 dB range, 100% sensitivity, 100% speed, and "Hold" off), as long as I shifted the automation points to match the audio (and this was needed every time) it worked as expected .

OP, mike_mccue, are you still around? Can you see if maybe the reason you had tremolo was similar to mine, that the automation points were out of sync with the audio?

Mr. Melda, does this help at all, I hope it helps for troubleshooting!

Thanks -



i5, 16GB, W7, Cubase 7.5.20. 44.1kHz, 512 buffer

[MeldaProduction]

Well, automation recording is a problem on it's own when latency is expected... Different hosts work differently, but there are often problems with them.

In general I recommend NOT to use the automation in MAutoVolume! It has been added just because customers ask for it as it is in the Waves' thing, but I really regret listening to the customers about this... It sucks, it always will, and it IS useless. If you need automation, then just automate output gain for example. But automating the gain performed by the plugin is very wrong, the response is very slow, depends on the host's interpretation & optimization, maybe different everytime you play it... it's just bad... For maximum audio quality, do NOT use it!

[alexis1]

Well, automation recording is a problem on it's own when latency is expected... Different hosts work differently, but there are often problems with them.

In general I recommend NOT to use the automation in MAutoVolume! It has been added just because customers ask for it as it is in the Waves' thing, but I really regret listening to the customers about this... It sucks, it always will, and it IS useless. If you need automation, then just automate output gain for example. But automating the gain performed by the plugin is very wrong, the response is very slow, depends on the host's interpretation & optimization, maybe different everytime you play it... it's just bad... For maximum audio quality, do NOT use it!

NOW you tell me, after hours of troubleshooting!

Beyond that ... what you say about the automation is very interesting, thank you. I just played together the automated exported output with the "direct" (non-automated) output, same settings on MAutoVolume. You are right, they don't null, but it is just barely. No phasing either when they are played together. And just to confirm no coloration by the plug-in, no phasing when the processed audio is played back with the original.

Importantly - no tremolo in any situation, even without having the "Hold" parameter on, and even with 100% speed sensitivity, i.e., pushing the plug-in hard. Just to clarify, the only time I ever had tremolo was when I didn't correct the timing of the automation by shifting it to the right to match it to the audio.

That's too bad about the automation not being generated consistently by the plug-in. But, I am OK with working without automation, very happy with the sonic results of simply exporting (i.e., recording) the processed vocal to another track, works as expected!

Question - all my processed audio comes back with about 20dB less gain than the source. Is there some setting in MAutoVolume that is responsible for this?

Thanks -

[MeldaProduction]

Hehe yeah .

About the automation - the problem is not caused by the plugin (and often not even a host), it's just that automation was never designed this way... it's meant for user changes, usually very slow ones. The plugin normally works sample-perfectly, so it would need one automation value for each sample, that's impossible with most plugin interfaces and impossible to handle by the hosts anyway. Many of them even employ lots of optimizations, such as removing points that seem "useless" etc. The grain curve should be really more like a dedicated audio signal, not automation curve. Eventually there are so many factors, that it just won't end up well... it's a stupid idea from the beginning, don't know why I agreed with it .

20dB - I'm not sure what you mean, but I would say that your audio is already quite high in level. MAutoVolume has some predefined loudness level to which it tries to level the input. I'd suggest watching the gain graph and if it looks like it is too much above 0dB all the time, then you should probably use the input gain parameter to bring it all down.

[alexis1]

Hehe yeah .

About the automation - the problem is not caused by the plugin (and often not even a host), it's just that automation was never designed this way... it's meant for user changes, usually very slow ones. The plugin normally works sample-perfectly, so it would need one automation value for each sample, that's impossible with most plugin interfaces and impossible to handle by the hosts anyway. Many of them even employ lots of optimizations, such as removing points that seem "useless" etc. The grain curve should be really more like a dedicated audio signal, not automation curve. Eventually there are so many factors, that it just won't end up well... it's a stupid idea from the beginning, don't know why I agreed with it .

Understood, I guess it doesn't matter as it seems that MAutoVolume works great without automation R/W. BTW, what did you think of the automation curves generated with two of the other plug-ins out there that do this, Quiet Arts Wave Rider, and Waves Vocal Rider? I am so glad I held off on buying them, and found yours instead!

20dB - I'm not sure what you mean, but I would say that your audio is already quite high in level. MAutoVolume has some predefined loudness level to which it tries to level the input. I'd suggest watching the gain graph and if it looks like it is too much above 0dB all the time, then you should probably use the input gain parameter to bring it all down.

Yeah, it was an old random too-hot sample, will check it out using audio with more reasonable levels.

[MeldaProduction]

I didn't test that in the other 2 plugins, I only receive lots of messages from people going from Waves to us claiming it sounds better . In generally I don't think it could be better in any way, simply because it will always be slow. And also since each host represents the automation different way, there would be inherent differences...