Probably a silly dynamics question

[mixsit]

Would there be a way to reduce the dynamic range of a signal by simple ratio and not include time' and threshold?

I may be missing something really basic here. But if you were to scale each sample's level value by say .5 would that make for a transparent dynamic range reduction with out the attack and release artifacts of a compressor?
I'll expand on the question if need be -or more likely if it doesn't completely get blown up and sunk right off.

 

[Bristol Posse]

If all your doing is pulling down the fader on each track by the same amount then all your doing is making the overall mix quieter.
Dynamic range is basically the difference between the loudest and the quietest. So if pre turning everything down it went from 0.0dbfsto -10dbfs and after you pulled everything down it went from -2dbfs to -12 dbfs you still have 10db dynamic range, its just at a lower over all level

That sounds right to me anyway

 

[noisewreck]

But if you were to scale each sample's level value by say .5 would that make for a transparent dynamic range reduction with out the attack and release artifacts of a compressor?

No, you would simply lower the volume by half (6db ?).

For what you are recommending, you need to have an application, that analyzes the entire audio file, and calculates the average level by taking the level of all samples into account, then you would raise the samples that fall below this point, and lower the samples that are above this point by some predefined amount to reduce dynamic range. You wouldn't want all the samples to get the same value as your average level, unless you want to get a steady DC offset

Although, this might sound mathematically pure, or clean, I am not sure it would give musical results.

 

[Waltz Mastering]

Probably the best way to get transparent dynamic range reduction with out the attack and release artifacts of a compressor is volume automation.

 

[dobro]

How come volume automation doesn't produce the artifacts that other things do?

 

[noisewreck]

How come volume automation doesn't produce the artifacts that other things do?

Because, although at times tedious, you can accurately target what needs to go down and what needs to go up.

 

[Farview]

How come volume automation doesn't produce the artifacts that other things do?

Because the transient isn't what we hear as volume, but it is what the compressor tends to react to. You will automate the volume to what your ear hears, not the random level of a peak.

 

[mixsit]

No, you would simply lower the volume by half (6db ?).

For what you are recommending, you need to have an application, that analyzes the entire audio file, and calculates the average level by taking the level of all samples into account, then you would raise the samples that fall below this point, and lower the samples that are above this point by some predefined amount to reduce dynamic range.

Yeah, not surprisingly I'm way off on the math methods.
I was thinking along the line of starting at some fairly low level, reducing everything above as would happen with a low ratio compressor.

Although, this might sound mathematically pure, or clean, I am not sure it would give musical results.

I don't know either. I've done a few things like zero' attack/extremely fast release (Sonitus plug goes there) but 2:1' in place of limiting' ratios to just lightly scale a kit for example. Lately a project on one of our three piece combo band live multi tracks as a learning tool an experiment to have no 'targeted for polish' use of any dynamics or eq, automation global only 'per song' block leveling, just to represent as close as I can our natural 'on-stage' blend and dynamics (warts or no).
In any case it strikes me that if there were no time functions involved (like you I presume done off line) why it would not be very natural sounding. Unlike a compressor that is always hunting (moving) and we juggle timing for an acceptable compromise there? I've never heard a comp not sound funky', even at very low ratio, at lowish threshold. But I've often felt I hear (at least partly) as to why.

 

[SouthSIDE Glen]

Would there be a way to reduce the dynamic range of a signal by simple ratio and not include time' and threshold?

You could do it with a graphical dynamics processor plug set to a simple ratio and a threshold set low enough to cover the entire dynamic range of the input signal.

In practice I know that Adobe and Sony platforms contain such plugs natively, and give you both look-ahead and the ability to set A/R times to 0ms. And I assume most other DAW platforms have similar native plugs. How musical the results will be is a question, but it shouldn't be too difficult to set one up and try it. Maybe a little slight of hand in the manual settings might be needed to handle stuff below the nominal bottom range of the plug (e.g. -100dBFS in Audition and -80dBFS in Sound Forge), but chances are that's going to be below the signal's noise floor anyway.

G.

 

[mixsit]

Wow. I go off line to carefully' respond, come back and bam!

I'll do some search, but in general how is the graphical dynamics processor different Glen?

And I agree automation is (often) job one -due to block' moves (relatively), and lightening the load on comps, if you're going for transparent'. This is mixing' moves. And again- pre-comp a very different goal/target than post' comp. (Choose well..
Where fader moves don't go in the case though is they are by nature fairly static, not ratio or knee driven.

 

[dobro]

Because, although at times tedious, you can accurately target what needs to go down and what needs to go up.

You started your answer with 'because', but although it answered the question for you, I'm not sure it answered it for me.

I use volume automation all the time nowadays, and avoid compression unless it's really, really necessary. Things sound more natural that way, and to my ears, for most of the mixes I do, better.

I've got a rough idea of why a compressor introduces artifacts - the algorithms can crush the stuff, but not accurately enough to sound natural, and so the more radical the setting, the more unnatural-sounding the output.

But I don't understand why a volume envelope *wouldn't* introduce artifacts. I know it doesn't queer the original sound, cuz I use it all the time with good results. But I don't understand why that is.

 

[mixsit]

How come volume automation doesn't produce the artifacts that other things do?

Look at it like this. With an automation line you have the option to target in a few interesting ways where a compressor can not even touch. (well, you can automate the compressor too but..

Where a section, phrase, or partial word for that matter, is poking out' (or to lift and emphasize ), a flat level block will raise', with out ever touching the micro.

The slope between the start level and the new level can follow to match (or scale) the ramp of the track change.
Or on a three point boost or dip (ramp up and down rather than a block') can be spread (or narrowed) to scale specifically around a peak.

The main thing is none of this is tied to a timing function, (or set threshold' compromise. To get that with a compressor you would have to have what, attack, release, ratio, threshold, four lines of automation(!) just to get close?

 

[dobro]

Nope. Don't get it. But never mind. I know it works, even if I don't know why it works. For the time being, I'm gonna think of it in terms of simplicity - volume automation is so simple - it only does one thing, whereas compression has at least four things going on, like you say.

 

[flatfinger]

How come volume automation doesn't produce the artifacts that other things do?

compressors do wave-shaping (distortion) if the attack or release settings are not set in a sympathetic relationship with the source material . You usally don't do volume automation in any where near as high of a resolution to do that . With a , compressor, If you have a fast attack , then you are attenuating the volume in milliseconds .

Drawing volume automation should result in a compression that would be akin to an opto unit ( known for slow , two stage releases) or a compressor that uses a RMS detection scheme. In other words ; gently.

Another reason volume automation is cool is because allot of sources are not staying in a nice little zone for the whole length of the track so the settings that were working in the verse are too much when the chorus gets jacked up . The aformentioned automation on say just the threshold ( bringing it up a few db during the louder section ; then back when it quiets down ) can often suffice . Sometimes just setting it up whilst looping the loudest part of the song and then automating it to lower the threshold down during the softer parts will do the trick.

The point is that unless your doing a track that just drones along at the same energy level , then you might need to get "dynamic" with your dynamics processors .

Some harkin back to the times when you had a bunch of folks , hovering over a analog desk ,(no automation at all ) making all the fader moves , turning sends off and on ect . More of a preformance art unto itself , and some argue better mixes as a result.

 

[SouthSIDE Glen]

I'll do some search, but in general how is the graphical dynamics processor different Glen?

I assume you are asking how it differs from a standard compressor? It's different in that it's not trying to emulate a standard compressor, and therefore does not limit itself as to what it can do to signal dynamics. A "compressor", as we commonly call them here, is technically specifically one particular kind of compressor - a broadband downward compressor.

With a dynamics processor, you can downward compress, upward compress, downward expand, upward expand, use multiple compression/expansion ratios at different signal levels, select the specific frequency bandwidth it works upon, yada yada yada. If it has to do with dynamics, it'll play with it somehow. And because it's not trying to emulate hardware, it doesn't limit itself (much) when it comes to the main settings of A/R times, threshold and ratio; if you cam punch in a number, it'll take it (within some reason, of course).

Here's a screen shot of the one from Adobe Audition 1.5, just as an example, but all of them I've seen have a similar format (with one exception, below) of having the graphical line on X/Y coordinates where the X scale represents input volume, and the Y scale the output volume. In this format a compression ratio of 1:1 is a solid 45° slope. By changing the angle of the slope or putting bumps or valleys in it, one can set up a custom compression/expansion curve with/without makeup gain as one wishes. The example shown has a 2:1 compression on everything up to -10dBFS, where it then hard-limits:

The one I prefer to use myself - and I often use this as my go-to neutral compressor, is the "Dynam-izer" plug from Roger Nichols Digital. The interface is far superior (IMHO) to the standard one above, and I've always like the surgical neutrality of the RND stuff. I carry a lot of the RND stuff with me wherever I go.

EDIT; YIKES! I just went to link to the RND site and it looks like they shut down operations as of just last week. It makes me wonder, if anybody is interested in what I consider to be the most underrated line of VST plugs on the planet, you might want to check the dealers and see if they're blowing out their inventories on this stuff.

G.

 

[miroslav]

Yes...that is weird the way the Roger Nichols Digital website shut down...?

 

[SouthSIDE Glen]

Yes...that is weird the way the Roger Nichols Digital website shut down...?

I'm feeling deja vu all over again. It was just a couple of years ago that the exact same thing happened to a company called Elemental Audio, and a similar question popped up on this forum.

Well, it turned out that Elemental Audio was bought by Roger Nichols. he re-named and re-packaged the excellent Elemental Audio plugs and thus the "Roger Nichols Digital" line of plugs was born. I wonder if something similar is happening here.

And, just for full disclosure, technically it's the older Elemental Audio versions of the plugs that I own and use, but the difference between those and the original RNDs was/is cosmetic only.

G.

 

[flatfinger]

Does'nt the RND use a I-lok now ?? If they went kaputz then they may have taken there authorization servers off line too .Lets hope EA can get the productline back and recover. IF not.....


Thank this gentleman for being a better A.E. than buisnessman!!!!!!

 

[SouthSIDE Glen]

Thank this gentleman for being a better A.E. than buisnessman!!!!!!

I'm afraid that I have to agree with that. He mis-stepped right out of the gate when as soon as he bought the EA stuff he immediately bumped the prices of each plug some (if I remember right) $50 each minimum. There was screaming on the boards, except from those who were either friends of his or who were afraid of the politics of pissing him off . He softened the pricing and put together some bundles after that, but I think a lot of people never came back after that.

It's a shame, because the software itself is great. I'm hoping someone else buys the company or at least the codebase and adds the product line to their own. Then again, I suppose I don't really care, because I already got my stuff .

G.

 

[dobro]

compressors do wave-shaping (distortion) if the attack or release settings are not set in a sympathetic relationship with the source material . You usally don't do volume automation in any where near as high of a resolution to do that . With a , compressor, If you have a fast attack , then you are attenuating the volume in milliseconds .

Drawing volume automation should result in a compression that would be akin to an opto unit ( known for slow , two stage releases) or a compressor that uses a RMS detection scheme. In other words ; gently.

Got it, thanks.

So, it would be possible to draw volume automation in a way that would sound like a compressor that's really working? Just get down to the sample level, in other words?

One thing this thread has done for me is to make me realize I don't have the first idea about how my software works for even a simple thing like volume/level. Each sample must have a value that translates into volume level. Yes?