Ideally, I'd like to find a VST plugin that would automatically compress the living (excrement) out of the audio dynamic range of a given source. I would do this during one or another stage of authoring, so I don't need this done in real time. In fact, it very much seems to me that a two-pass compression operation would be perfect.
I'd very strongly prefer to avoid the "why?" debates in this thread, which I find extremely tiresome. Suffice to say that as a night owl, 90% of the time I watch anything on TV, I'm at serious risk of waking people up and causing unacceptable levels of aggravation for myself and others. It's bad enough that I have to constantly adjust the volume levels every few seconds while watching live or streaming TV (adjustments that frequently come too late), but that's utterly unacceptable for content I author myself!
My usual plan would be to apply such a VST filter to the audio source within TMPGEnc Video Mastering Works 5, but I could also use the filter in Goldwave or Audacity (or whatever) and mux the adjusted audio back in later.
Now, there are quite a few audio dynamic range compression VST filters out there (I've already examined roughly 20 of them), but all of those I've found are far from automatic and instead require several tricky manual settings, settings that can take on arbitrary and virtually infinitely variable values, essentially requiring innumerable hours of trial and error tweaking! Well, (intercourse) that!!!
I even found this: Intelligent dynamic range compression, from the Queen Mary School of Electronic Engineering and Computer Science, that's alleged to be just such an automatic range compression VST plugin. But is it? Is it really? Hell, no! The plugin interface has all the usual manual adjustments that all the other compression plugins have, but if you enable the "automatic" buttons, the plugin does absolutely nothing to the audio waveform! That is to say, the amplitudes of every bit of the processed waveform remain unchanged (which, since it's a university student's freely provided software, is unsupported).
What I'm looking for is a truly automatic plugin with one and only one adjustment: something like "compression level", from "light" to "compress the living (excrement) out of it". Again, ideally it would be a two-pass operation, which could easily avoid unpleasant artifacts like "pumping" and "breathing" and so on. And I don't mind paying a reasonable amount for it (ideally no more than about $50-75 USD).
So the question is: Does such a VST plugin exist? Or any other Windows-compatible software? Anything at all?
eta: And don't bother telling me about "normalization"; it's nowhere near good enough for me.
I would have thought that this would be a much more common question here, but aside from a few old threads that don't directly address my question, there's surprisingly few threads about it here...
+ Reply to Thread
Results 1 to 16 of 16
Last edited by EmmB; 26th May 2014 at 01:34.
There is a VST plug-in that emulates the SSL 4000 recording console compressors that really sounds good.
I've seen it demonstrated ( I should say heard ) and sounded like the real console compressor (analog) that
I have used for commercial CD mastering. I'm not sure if it comes individually but I know it comes bundled
in a package. The bad news the bundle cost about $600 (US). But I was impressed by it.
Last edited by Dougster; 26th May 2014 at 01:40.
In any case, compressors need to be tweaked for each case since the source levels and dynamics also vary.
So finding a one setting fits all automatic dynamic range control can become damn near impossible. You also need to
consider what the intended playback environment conditions are too. A lot of problems people have with bad compression
artifacts can be to improper attack and release rates, improper input sensitivity and compression ratios. You get the idea.
Hope this helps.
Last edited by Dougster; 26th May 2014 at 01:56.
I thought most TVs would have audio dynamic range compression built in these days. Mine has a "night mode" which compresses the audio pretty hard, although I tend not to use it for audio.
If you don't mind decoding via DirectShow, ffdshow's audio decoder will happily load these WinAmp plugins, and it appears there's a VST version of the first one, which is very simple to use.
As I use my PC as my media player, I encode the audio "as-is" and compress it on playback via ffdshow and a Winamp plugin. If memory serves me correctly VLC has an audio compressor, but if they're not an option......
I'd be trying to find a way to decode and compress at the same time if it was me. That way you could simply re-encode and not need to convert it to wave or import it into an editing program first, apply compression, export it, maybe re-ecode it again......
Last edited by hello_hello; 26th May 2014 at 02:50.
Not only do most TVs have this (Night Mode), but almost ALL HT Receivers do.
Or you could do the simple, low-tech alternative: Use HEADPHONES! (I'm doing that very thing with a wireless pair right now).
Would save you a WHOLE LOT of wasted time & energy & effort.
Thanks again, Dougster!
You're absolutely correct that nearly all (or arguably just plain all) software compressors require significant tweaking if high quality output is desired.
But first, I don't care all that much about the resulting audio quality, since I won't be playing anything loud enough to hear any flaws. Second, over the years and decades I've tried precisely such tweaking -- adjusting this a bit higher and that a bit lower, then this a bit lower and that a bit higher, ad infinitum -- until the point where I craved a few months rest in a psychiatric hospital!
Now, just imagining such tweaking is enough to bring on the beginnings of an anxiety attack!
Modern computer hardware and software have staggeringly powerful capabilities. My main computer is a powerhouse: an overclocked Intel i7-4770K with 32GB G.Skill 2400MHz Trident X RAM and a nVidia GeForce GTX 750 with 512 CUDA cores (which I got for video rendering etc, since I don't play games). Why the bloody hell can't such a computer analyze the audio and select the appropriate compression settings automatically? Let it take 20 or 200 passes at 99% CPU utilization if it needs to, but for Bob's sake, do it automatically!
As an aside, I've been a systems and real-time software designer/developer for going on 40 years now, so I'm not so ignorant as to believe such a piece of software would be easy to create. Oh, no. (In my callow youth I foolishly tried to develop a software tool with FFTs and other such primitive algorithms to first digitize recordings from vinyl discs (using PDP-11's connected to ancient ADCs) and then reduce loud pops and clicks from the data. Talk about being ludicrously beyond my tiny little brain's abilities! I had to have been a low-grade idiot to even try!)
But again: With today's powerful computers and software -- after all, Turing-level AI and the alleged "singularity" are allegedly "right around the corner" (pphht) -- why should I have to tweak all those settings manually? Forget it! All I should need to specify is just how much compression I want. Let the software calculate good values for threshold, attack and release/decay timings, etc, etc, and adjust them dynamically as the waveform varies.
Isn't that the whole point of having computers?
But seriously, Dougster, please don't think I'm lashing out at you. I'm actually very grateful for your replies. It just seems to me that by the second decade of the 21'st century, this problem should not only have been solved already, but the solution should be cheap...
You're right about those options being simple, Scott. But I've used both and I'm not at all happy with the results.
Yes, all my receivers have a "night mode" or the equivalent. But holy (excrement) do they suck! The dynamic range compression they provide is grossly inadequate, such that I still have to constantly ride the volume control and even then I still end up waking people up while using it.
As for phones, I have all kinds, from cheap earbuds to Stax electrostatic headphones. The best combination of quality and comfort I've found are my Grado SR80's. But holy (excrement) do they drive me crazy after wearing them for more than about 20-30 minutes! My ears start getting overly warm and start hurting about then and my frustration rises to the point where I simply must tear them off. And that's with just earbuds or the most comfortable phones I've got! I could never be an audio engineer, alas.
Nope, neither option is anywhere close to good enough for me...
Try Limiter No6. One of the presets should work for you. I use the Master_3 preset for quick and dirty comp.
@EmmaB, the problem with "automating via computer" and "why can't computers figure this out" is the fact that, while a computer can analyze the EXCREMENT out of a signal, it has 3 weaknesses:
1. It is very literal and cannot handle the wide variations of unpredictable signals that are truly available (or even common?) in the real world (and we haven't advanced THAT far in AI & patternmatching, yet)
2. It cannot read our minds
3. It cannot account for the wide & often unpredictable variability/subjectivity of human tastes & perceptions all within one algorithm/setting
There are Dynamic Range compressors that have a set pivot point (usually in the exact center of the 16bit data sample space), work on the WHOLE signal (all levels=infin. low threshold), are instantaneous (both attack & release), unity gain, and have simple, universal 2:1 compression ratios. dBX is one (based on it's common analog box).
They are not hard to find, but have much more limited use in this post-analog world (where it isn't necessary for transmission/storage-sake to compress DR). Your biggest problem might be finding one that doesn't cost $$ and/or isn't built as a proprietary plugin to a $$ DAW (e.g. ProTools).
This might be what you're looking for. I don't recommend that, but then I'm not in your exact predicament, nor know whether you'll like what you get or not. I also still think there are headphone types that you may not have tried that DON'T bother your ears, but I'm not you so I couldn't guarantee that.
Last edited by Cornucopia; 26th May 2014 at 11:24.
Now this is kind of a GENERAL technique I used in the pro audio and mastering industry (about a decade or so ago).
Assuming you have a compressor unit with more than just ratio and threshold. And I'm using a CD mastering technique here.
And I am also assuming you want a kind of a 'leveler' setting. Generally you want a long attack time (100 to 180 ms ish) +/-
and a short release rate (but not too short) 10 to 25 ms ish, +/-. Set the ratio for about 8 to 1 (+/-). Then set the
threshold to achieve about a 2 to 6 dB of GAIN REDUCTION. Then adjust the output for the desired level. Remember
that I'm using a Rock/Pop well mixed music source to begin with and a client/artist in the who is very happy with what he or
she is hearing. Just one example here. I was not trying to 'limit' it but 'level' it. The real technique to using dynamic range
control is the attack / hold / release rates, something most people overlook. Limiting uses short attack short release
and a higher threshold setting.
Additional note here, when 'leveling', the unit is in constant gain reduction mode (as shown by the indicator) as for
'limiting', GR is only on the peaks and I'm generally using a higher ratio ( maybe 20 to 1 give or take ).
Cornucopia makes a great case in the previous post there.
Last edited by Dougster; 26th May 2014 at 12:32.
compressor as such (same with the Loudmax plugin) because it doesn't actually compress the loud bits..... rather it increases the volume of the quiet parts. I guess the end result is somewhat the same, although I have the amplification time set to the minimum (10ms). I've found a setting which seems to work well for me (I'm not constantly reaching for the volume control) and I rarely hear it working, but you could set it to "compress" more by increasing the full/maximum amplification amounts a little.
This was encoded using the same settings I use on playback. The first is a soundtrack downmixed to stereo and normalised. The second was encoded via DirectShow. I used ffdshow to decode, it's volume filter enabled with a volume reduction of 6db applied (to prevent clipping and give the compressor plugin more room to work), it's mixer filter downmixing to stereo (with the centre channel boosted a bit), and finally ffdshow's Winamp filter running the RockSteady plugin.
Last edited by hello_hello; 20th Jun 2014 at 12:47.
The subject of compression came up again in another thread, so I decided to make some short samples. All downmixed to stereo. "Uncompressed" vs "compressed with the RockSteady plugin" (same settings as in the pic above) vs "compressed with Levelator".
For anyone who stumbles across this thread and might be interested, here's a link: https://forum.videohelp.com/threads/364978-Use-a-different-player-to-Preview-in-Auto-Go...=1#post2332264
If anyone has any comments regarding the samples I'd be interested, but it might be a good idea to post them here, as the thread I've linked to wasn't started with the topic of audio compression in mind.