Transfer Expansion Curves Between Multibands ???

[SouthSIDE Glen]

no...because all numbers here are in the negative; & these are all truly expansion curves.

Not necessarily. Remember, "expansion" means expansion in the dynamic range of the signal. In your example curve, Slope 1 indicates a ratioed increase in the volume of your signals *below threshold B*. This indicates an upward compression - a compression of the range below a threshold up towards the threshold - not an expansion.

See the attached graphic for a different way of looking at it (via the Neodynum multi-level compressor). In that graph, the red-tabbed line in the middle at -16.6dBFS corresponds to "threshold B" on your graph (threshold A is the bottom of the graph). The left side of the graph indicates input values and the right side indicates output values. The yellowish part below that indicates the upward compression towards that threshold as indicated by "Slope 1" on your graph, whereas the reddish part above represents the "standard" downward compression of "Slope 2".

And this may be where some of your translation problems may be coming in. If your second compressor is not capable of upwards compression - many are not - then what you may actually be doing is downward compressing or downward expanding the lower half of your audio, which will give you entirely different results.

Oh and BTW, translating to a multiband compressor in the true sense of "multiple frequency bands" will not work unless they are capable of upward compression, and unless they are capable of setting each band for full spectrum (100% overalpping). What you really need is something like Neodynum (now called "Dynaamizer", by Roger Nichols Software). This lets you compress/expand based upon up to 4 different threshold levels, compressing or expanding in volume level bands rather than by frequency.

one could say that for every 16.6dB's of input, 1.2dB's of expansion occurs. That's the accepted premise, right? Ratio is always relative to *above* the threshold.

Ratio is measured by what happens above the threshold, that's correct. However, what you are getting are the numbers for Slope 2, as that point is the threshold for slope 2. That is a slope of 13.83, and therefore a downward compression of 13.83:1. (let's round off to 14:1 just for sake of discussion.)

And now that I think about it, I believe - in this case, anyway - that the slope of Slope 1 can automatically be calculated as the inverse of Slope 2, because their sum has to add up to 1:1; the overall curve starting and ending at the same place as a 1:1 line. That indicates a slope of ~1:14, which looks about right.

(Note that if the overall curve began or ended anywhere else than the corners, this would not necessarily be true. Also, if it went from corner to corner, but there were more than two slopes - if there were a third threshold and a third slope, for example - then the sum of all three slopes would have to add up to 1:1.)

[I was also going to ask you if you flipped your original ratio math...since were talking expansion here...& since the threshold is normally relative to output, w/ compression...but that's another story for another day.]

Well, yes and no. I guess it depends upon your definition of "flipped". Note that the slope for Slope 1 is 1:14. That appears to be expansion. But because it's adding gain only to the part of the signal below threshold 2, it's actually an upward compression in this case. So while there is a gain in volume coming out for every dB going in, one would still refer to it as "an upward compression of 14:1". It gets a bit confusing, I know; I'm not sure how to notate it otherwise.

G.

P.S. Keith, I said HE was missing (i.e. overlooking) the information that was available to him, not that the information was not available at all. That's three swings and three strikes for you in this thread. Hit the bench, Casey.

 

[kylen]

There are no ratio & threshold settings on the MultiComp plug-in. Instead, at the hard knee of the curve (which Spectral Designs calls the 'breakpoint'), 3 specs are displayed on mouseover...input, output & gain. The input & output are displayed as negative values.

On one of my favorite curves, for the LF band, reads as follows:

input -16.6dB
output -1.2dB
gain 15.4dB.

Since I'm not familiar w/ expansion, these numbers are throwing me...but I want to be able to plug them into another mastering multiband...which has expansion capability...but which has the typical ratio & threshold controls.

Hi Folks,
If you want another fun toy to play with Mark4man, Tracktion3 has a few fun compressors that will let you set multi-point compression/expansion using input/output transfer curve values, some are single-band some are multi-band. It'll give you negative compression if you want it...

Here's the manual:
http://www.mackie.com/products/tracktion3/pdf/T3_Plugin_Guide.pdf

Look at the 3-band stereo compressor and see if you like it. I've been tooling around with it for a while now and it sounds pretty good to me. I think you can get Tracktion3 for about $99 and the mastering plugs come with it - only usable in T3 unfortunately.

Youse guys are havin too much fun with this one!

Take Care,
Kyle

 

[Robert D]

Unfortunately, Keith...or should I say Dr. House without the Vicodin and without the genius...
G.

I've got to give the point to Glen, just for that line. My car got broken into yesterday, my wallet, cellphone, and video camera taken, and I haven't smiled for about 20 hours....... till I read that gem.
On the subject; it's an impressive little number game, but I don't think it has much to do with music. Different compressors sound totally different with the same attack/decay/threshold/ratio/knee settings. If I were comparing different MBCs, my criteria would simply be which one worked best for me and my brain, to arrive at the most musical sounding result. I'd be less concerned about the actual transfer function.

 

[flatfinger]

I've got to give the point to Glen, just for that line. My car got broken into yesterday, my wallet, cellphone, and video camera taken, and I haven't smiled for about 20 hours....... till I read that gem.
On the subject; it's an impressive little number game, but I don't think it has much to do with music. Different compressors sound totally different with the same attack/decay/threshold/ratio/knee settings. If I were comparing different MBCs, my criteria would simply be which one worked best for me and my brain, to arrive at the most musical sounding result. I'd be less concerned about the actual transfer function.

Sorry about your misfortune! hang in there , agree with your sentiments about function vs, form,
ONE MORE POST AND YOUR THERE!! 10K!!!!!!!!!!!!

 

[SouthSIDE Glen]

My car got broken into yesterday, my wallet, cellphone, and video camera taken, and I haven't smiled for about 20 hours....... till I read that gem.

Ugh, sorry to hear that, Rob. Glad I could give you a little smile, anyway. I'd give you some more rep, but I just got through taking you up to 9999 a couple of days ago, and this stingy BBS wants me to give some more to others first.

And I do agree with your point about no two compressors sounding the same. I just thought that this was a very interesting question (certainly better than yet another "how do I make my CD-R sound as loud as launching the Space Shuttle from my back seat" question), a fun one to figure out, and one which touches upon all sorts of interesting information, including yours.

BTW, your unfortunate situation reminds me of that old story about the home recordist who locked his keys in the car. He had to break the windows to get his wife out. [RIMSHOT]

G.

 

[kylen]

If I were comparing different MBCs, my criteria would simply be which one worked best for me and my brain, to arrive at the most musical sounding result. I'd be less concerned about the actual transfer function.

@Robert D...Good point - I'm sure that's true. I think the transfer function is of such great interest only because that's the main 'lame' user interface for this MB comp. The one I mentioned in T3 has similarities to this but goes a few steps further and provides ratio info - which is what we'd be talking about if Mark's comp had a better UI.

What's the main UI on comps? Threshold, ratio, attack/rel, gain(s)...I like gain reduction LEDs personally but they're not necessary. Without that kind of UI then we're stuck in the crazy transfer curve discussion. But like other folks have mentioned transfer curve doesn't account for dynamic stuff like detector, att/rel, makeup gain stage of an individual comp affecting the sound. I kind of agree - you can get a course tweak copying settings maybe but the rest has to be tweaked and finetuned by ear...just a thought.

 

[SouthSIDE Glen]

What's the main UI on comps? Threshold, ratio, attack/rel, gain(s)...I like gain reduction LEDs personally but they're not necessary. Without that kind of UI then we're stuck in the crazy transfer curve discussion.

I agree with you completly, when it comes to simple compressors (with one exception, which I'll get to in a moment.)

The thing is, what Mark is talking about with that right side graph on his plug is not really a compressor, it's a dynamics processor. That's like the difference between a volume control and a parametric equalizer. Standard single-threshold downward compression is only one small function that such a plug can perform. The irony is, that almost every waveform editor comes standard with one of these things, yet almost no one uses them...mostly because they simply don't understand them, and because, frankly, they are not easy to master.

I'll admit that these things are not easy to use effectively, but once you get the hang of them they are like a high-skill, high-specialty tool; when the task calls for it, there's no better tool in the world.

(I suppose I shouldn't complain they don't get used much. The exact same description could be applied to an MBC, and I bitch about those things get overused, misued and abused all the time )

For me the ultimate dynamics processor and UI style be the I/O graph that Neodynum/Dynamizer uses (along with it's extra associated controls, including most of the basic controls and meters, not shown in my example.) It sounds like maybe the one in T3 may be similar, I don't know, I haven't seen it. But it so well graphically displays both compression and expansion from input to output in a very (IMHO) intuitive fashon, that I'd love to see that become a new standard for dynamics processing layout instead of the X/Y graph. and personally I think even standard compressors could use such a graphical display as well.

G.

 

[kylen]

For me the ultimate dynamics processor and UI style be the I/O graph that Neodynum/Dynamizer uses (along with it's extra associated controls, including most of the basic controls and meters, not shown in my example.) It sounds like maybe the one in T3 may be similar, I don't know, I haven't seen it. But it so well graphically displays both compression and expansion from input to output in a very (IMHO) intuitive fashon, that I'd love to see that become a new standard for dynamics processing layout instead of the X/Y graph. and personally I think even standard compressors could use such a graphical display as well.
G.

True - Neodynium is a good one - I need to use it more. There's a fun little video demo on RN's web site that shows a thing or 2 about it. That is a really different interface kinda tricky till ya get used to it (which I'm not yet).

The thing I like about the Mackie Tracktion 3 comps (actually dynamics processors like you mentioned) is that they're also multi-node - but each node can be an upward/downward compressor or expander (isn't Neodynium downward only ? - I forget). I'm really getting into the inflating sound of a little upward compression (or is it expansion?) somewhere around -30db to -20db relatively speaking. Now doing that in a multi-band setting is something I think I'm liking - mostly just to handle the bass and mids dynamics differently, and so I can leave the highs alone.

 

[SouthSIDE Glen]

The thing I like about the Mackie Tracktion 3 comps (actually dynamics processors like you mentioned) is that they're also multi-node - but each node can be an upward/downward compressor or expander (isn't Neodynium downward only ? - I forget).

Nope, Neodynium/Dynamizer does it all; compression up or down and expansion up or down, for 4 different level bands, or "zones" I think they call it. Go back and look at the graph I provided on page one, that shows upward compression on the lower zone.

I'm really getting into the inflating sound of a little upward compression (or is it expansion?) somewhere around -30db to -20db relatively speaking. Now doing that in a multi-band setting is something I think I'm liking - mostly just to handle the bass and mids dynamics differently, and so I can leave the highs alone.

It's expansion if you're expanding the overall dynamic range of the zone, compression if you're shrinking it. If you're pulling up the stuff between -20 and -30, that's upward compression. Note, however, that anything quieter than -30 remains untouched, so the overall dynamic range of the signal doesn't really change, you're only compressing within that range of levels, within that zone.

As far as the "MBC environment" idea, that's built right into Neodynium/Dynamizer.It has the "key input" section on the left of the display which allows you to apply a series of lo/high/bandpass/notch filters to the signal before sending to the dynamics processor. So it's like being able to send a single fully-configurable EQ/bandpass curve to the dynamics processor(see attached).

G.

 

[NYMorningstar]

Yep, all curves eminate from the lower left corner, as that indicates -inf input (zero signal). And a zero input signal is always going to be a zero output, no matter how much expansion or gain you dial in . So that means x,y of -inf,-inf as a strting point the whole time.

I don't know if you touched on this yet. This thread got too confusing for me when the different plugs were introduced

If the signal started out horizontally flat and rises upward, let's say at -48 instead of -inf. ,then all the signals in the flat area would be eliminated and represent gating.

 

[kylen]

As far as the "MBC environment" idea, that's built right into Neodynium/Dynamizer.It has the "key input" section on the left of the display which allows you to apply a series of lo/high/bandpass/notch filters to the signal before sending to the dynamics processor. So it's like being able to send a single fully-configurable EQ/bandpass curve to the dynamics processor(see attached).
G.

I'll have to check out Neodynium some more, the key input is just going to adjust detector sensitivity if I remember correctly. It's still a single-band so when it does compress or expand the full spectrum gets it as opposed to the multi-band I'm trying out. More ways to get a desired sound - they're all good!

 

[boingoman]

well...

This may prove to be a fruitless venture anyway.


So...my numbers are either incredibly wrong...or the spectral design specs on the MultiComp far exceed anything accomplished by Universal Audio...or something.

I should apply the MultiComp, w/ these settings, to an evenly EQ'd composition...& let you guys hear this thing. It sounds like it rivals anything ever done at Sterling Sound...in a little f_ck_'n software box. Unbelievable.

mark4man

What you are doing by setting two different comps to the same settings is like putting two different cars in third gear at 3500 rpm. It would come as no surprise that one moved faster than the other, would it? And you wouldn't base your opinions of either vehicle based on those results, would you? The same logic applies to your comps.

Figuring out the threshold and ratio of your comp is a good idea. But you shouldn't be surprised that transferring those same settings to a different comp results in a very different sound.

 

[mark4man]

Hello again, all...

SouthSide...

Just to get back to original question for a moment (& at least hopefully solve that side of the thread)...a friend of mine from Recording Institute of Detroit tells me that it's a 1:1.21 ratio, -16.6dB threshold expander. He agreed w/me that ratio is basically the O/I slope, but got his numbers by considering the entire scale of each:

90 -1.2 = 88.8
90 -16.6 = 73.4

then:

88.8/73.4=1.21

(without sounding like a lawyer)...does that comport with your summation, counselor ???


boingoman...

Figuring out the threshold and ratio of your comp is a good idea. But you shouldn't be surprised that transferring those same settings to a different comp results in a very different sound.

I think the disparity is probably more akin to what Glen mentioned early on...that there is both expansion AND compression going on. Which also means that, w/ the Precision Multiband...I'll never be able to do a true A/B anyway (as that plug-in...for each band...offers an 'either/or' selection for compression or expansion...it can't do both.)

I think others here have already said what I should do...(which is also a conclusion I have come to myself)...& that is to just use the f'in MultiComp in my work as a tool appropriate for what it does...because...the thing sounds absolutely gorgeous anyway. And because it sports a 'killer level' type output...if I choose to scale the output back to suit my needs...& if that changes the overall sound as a result...I can tweak various bands to get it back.

Thanks everyone for your help, thoughts & expertise.

mark4man


BTW - Glen...are you south side of Chicago or south side of Asbury Park?

 

[SouthSIDE Glen]

If the signal started out horizontally flat and rises upward, let's say at -48 instead of -inf. ,then all the signals in the flat area would be eliminated and represent gating.

If you're describing the curve on the DP X/Y display, you're absolutely correct. Basically that would be saying that for the range of input values from -inf to wherever that bottom flat line ends (e.g. -48dBFS), the output value remains 0dBFS, which is the same as gating out those low volumes.

I'll have to check out Neodynium some more, the key input is just going to adjust detector sensitivity if I remember correctly. It's still a single-band so when it does compress or expand the full spectrum gets it as opposed to the multi-band I'm trying out.

Take another look at the screenshot. Basically the key input section allows you to set an EQ curve across the whole spectrum, but you can apply up to three simultaneous filters of any type to the spectrum. On could very easily synthesize a 3-band MBC simply by selecting thre bandpass filters (or a lo-pass, a band-pass and a hi-pass, if that's what you're going for) that crossover at typical MBC crossover frequencies.

If you wanted to do just the bass like you said, you don't even need to go that far; a single lo-pass will do it for you.

Or, if you wanted more flexibility beyond just an MBC, you also have shelves and notches to chose from, allowing you to create custom EQ curves equal to or beyond what most MBCs can create.

As far as sensitivity, there is an input gain control on the key input that allows one to adjust for overall gain changes otherwise caused by the EQ curve.

G.

 

[SouthSIDE Glen]

(without sounding like a lawyer)...does that comport with your summation, counselor ???

Looking back at it again, I did calculate right for Slope 2.

Where I made my mistake was assuming Slope 1 would be the inverse. That would be true only if that chart were on a linear scale. The logarithmic scale threw me wrong in two ways: first it made Slope 1 appear much steeper than anything even close to 1.2, and certainly looked like a rise of somehwere in the double digits, in which case 14 would have made sense; second is that it made the two slopes appear of equal length, in which case some simple geometry suggested that the two slopes would be complimentary. Again, true if it were a linear scale, but not on your stretched log scale. My mistake.

Attached is a screen shot of the DP plug from Adobe Audition, which uses a linear scale, and aslo provides a few more numbers than the Steinberg one does. The numbers are not exactly precise, but that's only because i could not select exactly 16.6 as the input value at threshold. The closest I could come was 16.4. But the results are close enough to verify both slope numbers as in agreement whith the above. Note how the apparence of the slopes is entirely different because of the switch to linear scale. That 1.2 slope actually is close to 1:1, where I would have imagined it, and that the two slopes are actually nowhere near the same length. D'oh.

I still contest the idea that there's expansion happening though, even though the ratio is reversed. Maybe we're arguing definitions here only, but to me, a 1:1.2 ratio applied to the bottom of the dynamic range of the signal means that the bottom part of the signal will be increaseed in volume by a ratio of 1.2 dB out for every 1dB in. In other words there will be an increase in the volume of the low end of the amplitude scale. The ratioed movement would be upwards, and therefore the overall dynamic range will be compressed. This would result in an upward compression, even though numbers indicate an expansion. A graphic representation using the revised true numbers is on the right of the attachment.

A true expansion on the lower end of the amplitude scale would mean there would be a ratioed DECREASE in the volume coming out. For every dB going in, there'd be some negative dB value coming out; the quiet would be made quieter. On an XY graph, that would be indicated by a negative slope (higher on the left than the right).

Note, OTOH, that if we're talking the top end of the dynamic range of the signal - the higher volume stuff above the highest threshold setting - then yes, a ratioed upward gain would indeed indicate an expansion.

BTW - Glen...are you south side of Chicago or south side of Asbury Park?

Oh, please! My name is Glen, not Johnny, I don't depend upon Bruce Springsteen to give me all my good ammunition, and I say "pop" instead of "soda"

I do like havin' a party and a rarely want to go home, I'll admit that, but that's as close as I come to being from Joisy . I'm born and bred on the southside of Chicago, home of Al Capone, Buddy Guy, Mayor Daley, Muddy Waters, Chess Records, The Checkerboard Lounge, and the only real baseball team in Chicago, the White Sox (though they are sucking right now. But that's OK, the Bears (also currently residing on the southside) training camp (south side suburbs) started today... )

G.

 

[kylen]

...Take another look at the screenshot. Basically the key input section allows you to set an EQ curve across the whole spectrum, but you can apply up to three simultaneous filters of any type to the spectrum. On could very easily synthesize a 3-band MBC simply by selecting thre bandpass filters (or a lo-pass, a band-pass and a hi-pass, if that's what you're going for) that crossover at typical MBC crossover frequencies.

If you wanted to do just the bass like you said, you don't even need to go that far; a single lo-pass will do it for you.

Or, if you wanted more flexibility beyond just an MBC, you also have shelves and notches to chose from, allowing you to create custom EQ curves equal to or beyond what most MBCs can create.

As far as sensitivity, there is an input gain control on the key input that allows one to adjust for overall gain changes otherwise caused by the EQ curve.

G.

Woops - minor communication problem. Neodynium is a single-band dynamics processor - yes? The key input allows the detector to be more or less sensitive (hyped) to a particular frequency(s) depending on the filters that are set. Neodynium can be more sensitive to bass for example, if the bass filter is hyped. However, once the single-band dynamics processor kicks in the entire input is processed , right? That's what I meant - That's different from the multi-band comps I'm messing with in tracktion. There's 3 different 'key inputs' and bands and detectors. If the bass dynamics processor kicks in the the MBC none of the other bands need kick in, only the bass band is affected. That's what I'm working with right now in T3, a totally different sound and feel - you knew that.

 

[NYMorningstar]

I'm born and bred on the southside of Chicago, home of Al Capone, Buddy Guy, Mayor Daley, Muddy Waters, Chess Records, The Checkerboard Lounge, and the only real baseball team in Chicago, the White Sox (though they are sucking right now. But that's OK, the Bears (also currently residing on the southside) training camp (south side suburbs) started today... )

G.

What about LeRoy Brown

 

[NYMorningstar]

and the only real baseball team in Chicago, the White Sox (though they are sucking right now.
G.

You should let them know you struck out chessrock for some inspiration

 

[chessrock]

Na ... according to another thread, I'm still on strike two.

We still gonna' go out for that beer, Glenny?

I still need you to explain all this stuff to me. I mean, apparently this issue was so clear cut and dry. And all you had to do was plug in those values that he gave you in his first post, and all this hoo-ha should have been solved in like one post. I need to understand how we're still on this subject after two pages.

We already know that the only possible value that his original values would have given us was the threshold (assuming that's where his coordinates were derived from) ... and that we couldn't help him with the ratio until we knew the slope of his curve (which he didn't provide us with until later). Now you mean to tell me that we don't even know if what he's doing is expansion or compression?

Jeez Glen ... if I didn't know any better, I'd think you were throwing me some wild fast balls.

.


- Keith the Jackass

 

[SouthSIDE Glen]

Keith the Jackass

The only part of any post you've made in this thread that makes any sense or has any truth to it whatsoever. I responded to your invitation in the other thread. Come on down, boy. I'll be happy to explain it to your face.

And, you incredibley dense imbecile, that was strike two on the shoving your idiotic words in my mouth. You're already struck out of this thread.

NYM, that was a good thought about the Sox inspiration. but striking Mr. Ricker out is not exactly a real accomplishment. When it comes to audio tech, Keith couldn't hit a Chicago-style 16 inch softball with a 4 by 8.

BTW, Leroy Brown was a wimp

G.