Cleanest most transparent vst compressor?

[MeldaProduction]

Hmm, can I ask how did you generate this?

Additive, bulk-sines, the initial phases are the only different factor.

Well, you got me interested, I'll have to try by myself .

Btw. no crossover can generate this, simply because the phase has been changed extremely and you'd need an extremely complicated FIR to do this. But I have to admit, that I'm pretty surprised about what can phase do. Thanks!

I'm not reverb expert but I know that most are based on allpasses, I'd suspect they're the ones responsible for the reverb's diffusion quality, messing with phases.

Well, a delay is an allpass as well and when you think about convolution reverbs, they are just a combinations of thousands of delays. I think what makes something sound "reverbated" is when you alter a phase too much - you can say delay something so it becomes almost a distant echo. Not sure though.

[MeldaProduction]

addon: It should matter only in the bass because of our hearing system. Imagine 50Hz sine - it has a period of 20ms, so if you move it by 180 degrees, you shifted it in time by 10ms, and you can definitely hear that. Fortunately most filters including crossovers apply continual phase change (which this is not), so let's say we shift 50Hz by 10 ms, then we also shift 60Hz by 9 ms (I'm generating random numbers ), so it is definitely not that audible.

[tony tony chopper]

Btw. no crossover can generate this, simply because the phase has been changed extremely and you'd need an extremely complicated FIR to do this

..or a lot of instances of stuff with crossovers. Someone initially pointed out how Maximus (& your plugin & any IIR-based one) can destroy transients pretty badly when stacked a lot of times.

I've made vocoders, based on lots of very narrow bandpasses, & thus a lot of transient smearing, big problem. You wouldn't be able to make a transparent multiband compressor with like 50 bands, because they would have to be narrow enough (no point in overlapping 50 large bands), & the phasing would be horrible.

It should matter only in the bass because of our hearing system

make a very low sawtooth, it will sound like a sawtooth, lower & it will tend to a periodic click. Now just randomize the initial phases and the periodic click becomes periodic noise, it definitely matters for all freqs.

[lateks]

^ what does all this stuff mean??? lol

[aciddose]

well, in any crossover the -3db points must be far enough apart so the result sums back to the original signal.

if you had 50 bands in the range of 20khz, you'd need them to be 60db/o slopes.

[Cordelia]

^ what does all this stuff mean??? lol

That Gol knows of what he speaks, Maximus is no more or less "transparent" than other multiband compressors because they all create some issues by design. And even that "transparency" is a silly word to associate with multiband compression. The upshot is that Maximus is a very nice plugin, developed by a smart guy who cares, and it sounds great on all kinds of stuff.

[tony tony chopper]

well, in any crossover the -3db points must be far enough apart so the result sums back to the original signal.

It could normally sum up to original with very large bands as well, which would get their -3dB on each other side, you then end up with quieter, large bands. But still, I don't think it's very interesting to have 50 large bands, each of them having influence over too many neighbor ones. (at the same time, it may sound good for compression, I don't know).
(even in classic 30-bands, 1/3 oct EQs, bands are quite large afterall)

[tony tony chopper]

You'll be happy to know that we've uploaded a new Maximus (VST) install, which is now "transparent" (with 24dB slopes only) by the definition of the last page.
(it's not gonna make any difference in practice, but it's only very slightly slower. Talking about CPU usage, since we were comparing dick sizes, Maximus is also around 3.5x faster than MMultibandDynamics for 3 bands. Now your turn to work on your code ).

[MeldaProduction]

I was wondering if the main "pro-maximus-guy" is you, and you are .
Anyway don't be afraid, I'll check the speed, but we all know that it depends by more than "3-bands". Anyway how did you solve backward compatibility? (just curious)

[tony tony chopper]

but we all know that it depends by more than "3-bands"

sure but I also believe Maximus is more featured anyway. Your limiter is also 3x slower, so you need optimizations.

Anyway how did you solve backward compatibility?

You mean adding the extra allpass to make it "transparent" (& actually mess with the phases even more)? I thought about adding a switch, but it makes no audible difference in practice. If anyone can hear a difference other than by not compressing & setting up a tiny band, I'll make it optional.

[MeldaProduction]

Hehe Tony, you are a big joker .

[Nokenoku]

And you are someone, who won't sell many plugins with this attitude.

I was already disturbed by the insolence of your "free" plugins (which should be called DEMO versions, because that's what they are), but now coming in this thread and bashing other developers is even worse.

[MeldaProduction]

And you are someone, who won't sell many plugins with this attitude.

I was already disturbed by the insolence of your "free" plugins (which should be called DEMO versions, because that's what they are), but now coming in this thread and bashing other developers is even worse.

Are you kidding? They are free, thousands people use it for free. They can be improved, that's all. Get some information before you start accusing!


And ok, I didn't want to discuss the previous thing, because it's not going anywhere, but ok. So why joker

ad1) I'd really like to know, where does he get the conclusion that Maximus has more features than MMultiBandDynamics. There will definitely be something we don't have and I'm sure, that for every feature MMultiBandDynamics does not have I will find some, that Maximus doesn't have.
Not mentioning that this is a completely ridiculous argument, since claiming that set of features is the only factor, that affects performance, is simply crazy.
So obviously this was an attempt to get some revenge for being accused, that Maximus had a buggy crossover, which it has/had. And because of that I didn't want to discuss it anymore. It simply makes no sense here.

ad2) Last 3 pages of this thread tony was trying to persuade everyone, that phase really matters (to justify the fact, the Maximus equalizes the audio because of the buggy crossover...). Then he said, he released new version, which changes the phase a lot (but I hope the frequency response is now flat), and there is no backward compatibility, since it probably isn't audible . Sorry, but he simply negated everything he stated before. Not mentioning, that -4dB (approx) dips, that where present there before are now not there, so all your projects will now sound different .

Nevermind enough of it, it's out of the topic anyway...

[HanafiH]

Caruthers, I don't mind you laughing at my compressor design, just don't ever call me 'Ygor' again.

[tony tony chopper]

ad1) I'd really like to know, where does he get the conclusion that Maximus has more features than MMultiBandDynamics.

Your plugin doesn't seem to feature saturation (& stereo sep but let's say this is not this plugin's normal job & mine is more a maximizer) and seems to have less control on the release curves (one release time & a couple of presets?).
It also doesn't show time vs compression feedback (which takes CPU to buffer when GUI is visible). It may not affect the sound, but it gets easier when you see what the compression is doing on a graph.

But the main reason I should be free to claim it's more featured than yours, is that I saw you pimping your plugins in all kinds of threads, like saying "MMultiBandDynamics is currently the most versatile". Well that's one claim, the most versatile, of ALL multiband compressors out there, is yours!

Not mentioning that this is a completely ridiculous argument, since claiming that set of features is the only factor, that affects performance, is simply crazy.

I didn't say that, I implied that bad programming was the other factor

Last 3 pages of this thread tony was trying to persuade everyone, that phase really matters

I don't try to persuade, I would like people to download your supposedly transparent multiband compressor (& you can't have anything against people downloading your plugin?), pass a simple sawtooth through it, and hear by themselves if it really is transparent. Quoting yourself, "use your ears".

Then he said, he released new version, which changes the phase a lot

Not a lot, since the added allpass is tuned on the high band (as you know), thus a higher freq, less dramatic. But yes, if you set the first 2 bands tiny on the lower part, it will be audible.

since it probably isn't audible

It's not, but feel free to show us in practical use, and then I can add a switch pretty easily.
The effect of a multiband compressor that does something is A LOT higher than the dephasing we're talking about. But you were the one making claims about transparency, not me - Maximus is not transparent, it messes phases just like yours & other short-latency multiband compressors, which is still neglectable compared to the constant auto-EQing & dynamics smashing a compressor does, as that's its job.