MeldaProduction

ZentralmassivSound · Post by **ZentralmassivSound** » Sun Aug 23, 2015 11:19 am

Hi Vojtech,

when using upward compression in MDynamics, you can't see the gain reduction meter because it is fixed to a negative scale and does not show positive values. Can this be fixed? If the user activates upward compression, the graph scale should just flip to a positive scale so the gain reduction can be displayed. No need to turn it upside down, just leave zero at the top and change the bottom to +infinity instead of -infinity and plot the GR as is.

I think I asked this already some time ago... anyway, asking again

MeldaProduction · Post by **MeldaProduction** » Sun Aug 23, 2015 11:25 am

Hi there

,

yes, you asked already, I denied it, explained and here you are again

. Ok, so nope this won't happen. Besides it would actually be technically complicated:

1) It's not what you want!

Upwards compression is wrong, period

. It amplifies noise, doesn't give you any control about what happens to silent signals if you want to avoid it etc...

2) There's much easier and better way: Just use normal upwards compression, set the threshold very low, and then use Range parameter!

. That way you control where the compression starts and where it ends, GR meters will work, there's no noise amplification (unless you set the threshold too low)...

ZentralmassivSound · Post by **ZentralmassivSound** » Sun Aug 23, 2015 11:59 am

MeldaProduction wrote:Hi there ,

yes, you asked already, I denied it, explained and here you are again . Ok, so nope this won't happen. Besides it would actually be technically complicated:

I don't believe it's complicated

Actually you just need to flip the sign of the scale and plot GR as it is, no?

1) It's not what you want! Upwards compression is wrong, period . It amplifies noise, doesn't give you any control about what happens to silent signals if you want to avoid it etc...

2) There's much easier and better way: Just use normal upwards compression, set the threshold very low, and then use Range parameter! . That way you control where the compression starts and where it ends, GR meters will work, there's no noise amplification (unless you set the threshold too low)...

But you can use the range parameter in upward compression as well to avoid boosting noise, right? Upward Compression has different sound characteristics, which I like in some cases, and I would like to use it (thanks your plugin offers this!). I would just like to have the GR plot, then it would be perfect

ZentralmassivSound · Post by **ZentralmassivSound** » Sun Aug 23, 2015 12:41 pm

Hm so what I liked about the upward compression is e.g. that you can have a snappy attack on a note and at the same time boost the reverb tail after that. I wasn't able to get this result in one regular downward compressors, I needed two of the to get it.

However the upward compression actually is a little problematic, the GR seems to rise when no signal is there, and it creates a huge attack when a signal comes in. This can be controlled better with the downward compression I agree.

frizzbee · Post by **frizzbee** » Sun Aug 23, 2015 9:41 pm

ZentralmassivSound wrote:But you can use the range parameter in upward compression as well to avoid boosting noise, right? Upward Compression has different sound characteristics, which I like in some cases, and I would like to use it.

Loved to use it too. Some time ago I had an hardware comp/expander where you could do upward/downward compression and upward/downward expansion. Its an different approach to get the sound you want and theres nothing wrong about it

_al_ · Post by **_al_** » Sun Aug 23, 2015 10:02 pm

i was also talking to vojtech about this through email last year. was emailing about a bug or something.
anyway, even back then, he wasnt happy about me using compression like that, even though i explained why i needed to do it like that.
anyway here's the last couple of emails:
----
me:
i tried that range thing. got to be honest i never even knew what that was for, and it seems to do the same as what im doing, yep

so thanks for showing me that. it will get some use on some things.
the only thing i dont like about that technique: you know that point i put at 38db? the one at the top of the curve? well, the way i work, i like double click just below/left of that point, and drag the new point to the bottom, to create a quick gate, and listen to the area i dont want compressed.
doing that, its easy to get the point where you can only hear the snare, and maybe a little bit of hihat whatever, and then just delete the temp gate point, feel safe knowing that if i change the volume of a sound, i only have to get and doubleclick for a new gate and re-adjust.
also, i dont really have any noise floor that i dont enjoy, lol. its electronic music, and all in digital, so i enjoy the effect when i get it right.

thanks again for all the help, and have a nice day

----

vojtech:
Hi Alex,

Very well

.
I know what you mean about the points. Here's the thing - if you wanna be creative, just play with points until it sound right. But if you some specific surgical operation, which is what i thought you are after, then the processors will be the best bet, because these shapes are designed precisely for a clean audio processing, where even 0.00001 matters.

Cheers!
Vojtech

so i guess it aint gonna happen

ZentralmassivSound · Post by **ZentralmassivSound** » Sun Aug 23, 2015 10:36 pm

My request is only about the display, no new feature or anything, so I don't know whether we talk about the same thing?

_al_ · Post by **_al_** » Sun Aug 23, 2015 10:56 pm

thats all my request was iirc.
vojtechs first response on here, was almost identical to the response i got by email

ZentralmassivSound · Post by **ZentralmassivSound** » Mon Aug 24, 2015 8:43 am

_al_ wrote:thats all my request was iirc.
vojtechs first response on here, was almost identical to the response i got by email

Ah I see. Well I thought it would just be flipping the sign of the GR plot if upward compression is activated, sounds like no big deal at all, but maybe it's more difficult to integrate with the rest of the code, I don't know.

MeldaProduction · Post by **MeldaProduction** » Mon Aug 24, 2015 6:13 pm

It's actually FAR more complicated than just flipping a GR plot. Start by something telling the analyzer to analyze something else (just this transaction is complicated in such a complex system), then you have to represent the data different may, detect maximums instead of minimums etc...

Yep, as always, things that seem simple to humans are complex for computers and vice versa

. Thanks god for that, it makes us "not computers"

.

Anyway I think the whole terminology about upwards/downwards is highly overcomplicated and overrated. The downwards compression just can be completely replaced by upwards compression. There's no advantage hidden there (unless you really want to amplify noise

).

Just a little test question for you folks : So with downwards compression of say 1:2, threshold 0dB, you expect the processor to increase level from -20dB to -10dB, -40dB to -20dB etc... So what happens to silence? Huh?

_al_ · Post by **_al_** » Tue Aug 25, 2015 5:22 am

no, i draw my own curve, like that picture i emailed you (obviously you wont remember), so zero stays at zero, and has an upside down exponential curve from the zero node, up to whereever i want the comp to stop, and let transients through untouched.
that was the 38db node in the above example (the pic i emailed you)
thats the sort of thing i cant do with the range slider

MeldaProduction · Post by **MeldaProduction** » Tue Aug 25, 2015 9:01 am

Aaaaha, sorry, I missed that one

. Still I'd prefer using the classic modes - upward compression with range. Less fiddly imho.

ZentralmassivSound · Post by **ZentralmassivSound** » Tue Aug 25, 2015 5:59 pm

Well upward compression is just a regular compression method. The range parameter is important, yes, not to have infinite gain boost on silence, but otherwise it should work normally and it should also display normally.

This tutorial shows a little about it's use:

In fact they don't act so differently. I thought about it in theory and I concluded that they even can act identically, but I am not sure. I wanted to check in MDynamics.

First thing, it really reacts strangely to the numbers you dial in. It seems to have a different actual threshold from what you set as threshold, because it is somehow put into calculation together with range. So it's difficult for me currently to find the right values.

Second thing I saw, attack and release phases seem to work in opposite ways, i.e. in downward compression you have sharpened attacks on the loud parts (when the compressor kicks in over the attack time), and smoothed attacks on the quiter parts, while with upward compression you have smoothed attacks on the loud parts and added attack on the quiet parts.

Have to think about that more... up to now I just used it once in a while by ear, not really trying to understand. However I do want to understand now.

edit: apparently I am misinterpreting scales in the graphs... will follow up

MeldaProduction · Post by **MeldaProduction** » Tue Aug 25, 2015 7:28 pm

Well, they are the same actually. They just differ in the shape and if you want to use upward compression you need to solve the silence problem anyway... Well, just use normal compression with low threshold and range

.

ZentralmassivSound · Post by **ZentralmassivSound** » Sun Aug 30, 2015 4:11 pm

MeldaProduction wrote:Well, they are the same actually. They just differ in the shape and if you want to use upward compression you need to solve the silence problem anyway... Well, just use normal compression with low threshold and range .

Ha, I was confused about the "upward" knob. Actually it depends on the ratio what you actually have:

regular mode (upward knob NOT enabled)
ratio > 1 => downward compression
ratio < 1 => upward expansion

"upward" mode (button enabled)
ratio > 1 => downward expansion
ratio < 1 => upward compression

Don't know if it is common to flip the ratio when changing to upward mode?

Anyway with the correct settings I could verify that upward and downward compression indeed yield the same result, just a different approach to handle it. So I think one doesn't actually need the upward mode.

(Though it's still strange to offer a feature and not offer the respective visualization imho

)

MeldaProduction

KVR Audio

Upward Compression GR display