KVR Audio

Decebal · Post by **Decebal** » Mon Jul 28, 2014 1:15 am

I think I must be missing something here but I don't see the end-game difference between these two things.

According to http://www.mastering101.nakedelearning. ... Itemid=168

Upward compression is parallel compression and maintains transients while downward compression smushes it all. But from my understanding that is all wrong.

My understanding is that:
Downard Compression: Lowering the loud parts (transients) of audio to decrease the dynamic range.
Upward Compression: Raising the quiet parts (body) of audio to decrease the dynamic range.
Parallel compression: Combining your audio signal with a heavily compressed (any type, but I assume typically downward compressed) version of that audio.

So if you just make the peak value the same for the downwards compressed audio as it is for the upwards compressed audio, I don't see how the two will be in any way shape or form different? Seems to be just two different ways of achieving the same thing, where upward compression seems more efficient if you want to keep the same peaks and not have to deal with makeup gain. Am I missing something?

cryophonik · Post by **cryophonik** » Mon Jul 28, 2014 2:29 am

Yes, you've got it right. Upward compression decreases dynamic range by increasing the part of the signal that falls below the threshold, thereby increasing the overall loudness by increasing the level of the quiet parts. So, it's not the same thing as parallel compression, but the end result can be similar.

Decebal · Post by **Decebal** » Mon Jul 28, 2014 3:07 am

Right so how is it any different than downward compression in the long run?

robertszalapski · Post by **robertszalapski** » Mon Jul 28, 2014 1:36 pm

The volume envelope is different.

Decebal · Post by **Decebal** » Mon Jul 28, 2014 2:29 pm

Can you elaborate on that?

No_Use · Post by **No_Use** » Mon Jul 28, 2014 2:36 pm

older, a bit similar thread

http://www.kvraudio.com/forum/viewtopic ... 50&start=1

I've asked there about the difference between parallel and upward compression a few posts down.

robertszalapski · Post by **robertszalapski** » Mon Jul 28, 2014 3:16 pm

Decebal wrote:Can you elaborate on that?

Do both, then flip the phase on one. Listen to the difference. Try your best to get it to match.

cryophonik · Post by **cryophonik** » Mon Jul 28, 2014 3:46 pm

Decebal wrote:Right so how is it any different than downward compression in the long run?

Downward compression: the part of the signal ABOVE the threshold is DECREASED. Dynamic range is decreased (i.e., compressed) by reducing only the louder parts of the audio. The quiet parts (i.e., the part of the signal below the threshold) are not affected. The end result is an overall reduction in loudness and increased headroom due to lowered peaks in the audio signal.

Upward compression: the part of the signal BELOW the threshold in INCREASED. Dynamic range is decreased (i.e., compressed) by increasing only the quieter parts of the audio. The louder parts (i.e., the part of the signal above the threshold) are not affected. The end result is an overall increase in loudness and raising of the noise floor. Headroom is not really affected because the peaks are still the same.

One thing to remember is that compression (and expansion) refers to the dynamic range, not the volume or loudness (although loudness can be positively or negatively affected). I think that's a common source of confusion to many people. Dynamic range is the difference between the quietest and loudest parts of your audio. You can decrease this range (i.e., reduce the difference) by reducing the peaks or by increasing the valleys. Expansion is the opposite - it increases (i.e., expands) the dynamic range by increasing the peaks (upward expansion) or by reducing the quiet parts (downward expansion).

Decebal · Post by **Decebal** » Mon Jul 28, 2014 4:17 pm

cryophonik wrote:
Decebal wrote:Right so how is it any different than downward compression in the long run?
Downward compression: the part of the signal ABOVE the threshold is DECREASED. Dynamic range is decreased (i.e., compressed) by reducing only the louder parts of the audio. The quiet parts (i.e., the part of the signal below the threshold) are not affected. The end result is an overall reduction in loudness and increased headroom due to lowered peaks in the audio signal.

Upward compression: the part of the signal BELOW the threshold in INCREASED. Dynamic range is decreased (i.e., compressed) by increasing only the quieter parts of the audio. The louder parts (i.e., the part of the signal above the threshold) are not affected. The end result is an overall increase in loudness and raising of the noise floor. Headroom is not really affected because the peaks are still the same.

One thing to remember is that compression (and expansion) refers to the dynamic range, not the volume or loudness (although loudness can be positively or negatively affected). I think that's a common source of confusion to many people. Dynamic range is the difference between the quietest and loudest parts of your audio. You can decrease this range (i.e., reduce the difference) by reducing the peaks or by increasing the valleys. Expansion is the opposite - it increases (i.e., expands) the dynamic range by increasing the peaks (upward expansion) or by reducing the quiet parts (downward expansion).

Ok so you repeated exactly what I said in my initial post, thanks for the clarification

...So based on the lack of answers I take it there is no difference between the two if you can change the peak volume. That brings up the question why even use downwards compression if you intend on maintaining the same peaks, auto makeup gain seems pointless if I can just do upward compression.
And yes I know what compression refers to.

robertszalapski · Post by **robertszalapski** » Mon Jul 28, 2014 4:58 pm

Ummm... it's not that hard to understand. There is a difference between the two. You were already told this. Upward compression won't level out peaks the same way. If the upward compression doesn't touch signals above, say, -12 dBFS, a -6 dBFS peak will be unaffected and a -3 dBFS peak will be unaffected. There will be a 3 dB difference in the peaks.

If you are using a normal compressor with a 3:1 ratio and a really fast attack with a threshold of, say, -12 dBFS, a -6 dBFS peak will be compressed to -10 dBFS and a -3 dBFS peak will be compressed to -9 dBFS. There will be a 1 dB difference in the peaks.

The volume envelope is going to be different. I told you this already.

Decebal · Post by **Decebal** » Mon Jul 28, 2014 5:45 pm

Alright I'd like to kil this thread, this clearly isn't going to go anywhere

robertszalapski · Post by **robertszalapski** » Mon Jul 28, 2014 5:51 pm

What makes you say that? Is there anything unclear about the specific example I just posted?

BDeep · Post by **BDeep** » Mon Jul 28, 2014 7:40 pm

Decebal wrote:Alright I'd like to kil this thread, this clearly isn't going to go anywhere

People here trying to help and this is your reaction?

robertszalapski explained it perfectly. What's ironic is that his explanation is supported by the graphs in the article you posted.

Decebal wrote:According to http://www.mastering101.nakedelearning. ... Itemid=168

Upward compression is parallel compression and maintains transients while downward compression smushes it all. But from my understanding that is all wrong.

If you look at the waveforms, that's the difference between upward and downward compression.

realmarco · Post by **realmarco** » Thu Jul 31, 2014 2:14 am

as far as i know downward compression is what any old compressor does.

that is attenuate anything that breaks the "threshold" volume, the ratio is for how much, the attack is by how quickly it reacts and release is how quickly it goes back to normal. Most compressor have "make up" gain knobs so that u can boost the signal since its been lowered by the ratio

a long attack is what lets a snare do the "Smack". with a long release the next Smack will sound different from the first

Upward compression is usually called an "expander" its the opposite of what a compressor does, that is it renders Louder anything that is Below the threshold level.The ratio determine by how much.
the attack and release do that same-ish job as a regular compressor

realmarco · Post by **realmarco** » Thu Jul 31, 2014 2:21 am

here is a better explanation from SAE(big american chain of colleges for this stufff)

http://www.sae.edu/reference_material/a ... ession.htm

Downward & Upward Compression