Digital vs. Analog compression and limiting - ILLUSTRATED...

[MASSIVE Mastering]

DISCLAIMER: The following is all IME, IMO, YMMV, yada, yada, etc. Take it as you will... Most of you probably know by now my affinity toward schveet analog gear - Some wonder why, others already know why. I had a client in a few weeks ago wondering why I didn't just "strap an L2" on his mixes to bring them up to "face ripping" volume. So, the clock stopped for a while and we experimented, audibly and (more fascinating) visually and the results were worth saving. Originally posted on a different forum with some fresh modifications for my favorite hangout...

There are a couple sayings that I live by that normally... (A) Mix with your ears, not your eyes - (B) Some gear does things with "more style" than other gear.

Here's a prime VISUAL example of (B) which relates to (A) and leads to (C) below...



The top is the untouched mix file - Healthily and happily undistorted, with excellent dynamics and a wonderful "smooth" and "warm" quality to it, sitting around -19dBRMS

The second file was mastered through a fairly schveet analog chain into excellent quality A-to-D converters. It sits in the mid -11dBRMS area and "appears" to still retain a good amount of dynamics (which it does) and a certain "smooth" and "warm" quality (which it also does).

The bottom version is the same file, through the same chain except for the limiting - which was accomplished through my personal favorite digital limiter, the UAD Precision Limiter. It sits a little lower in volume in the low area of -12dBRMS and "appears" to be essentially clipped with less dynamics (which is true) and a certain "cold" and "harsh" quality, mostly in the highs and upper mids (also true).

As good as the UAD limiter is (again, my personal favorite digital limter), the difference in appearance - AND sound - is pretty striking. And this is just one link in the chain - The one that arguably should have the *least* effect on the overall sound.

Again, take it as you will. It doesn't take much to beat the crap out of a mix.

Hence my other saying - (C) When it counts, use the right tools. If you don't have them or don't know how to use them, find someone who does. And of course (D) If it doesn't count, and you don't care, do whatever the hell you want - Just don't be pi$$ed off when nobody else cares either.

One thing I failed to mention is that the analog file peaked at -0.6dBfs while the digital file peaked at -0.1dBfs - So there's another 1/2dB advantage to the analog version - although I'm not going to adjust the volume of a -0.6dBfs file just to gain another couple tenths of a dB... That'd just be silly.

[Tweak]

Nice and illuminating, John. I notice how the UAD precision limiter really "flattens out" the dynamics. I need to keep this in mind on my own mixes. I too often succumb to intense limiting.

[MASSIVE Mastering]

Hey, at least it isn't the "L" series... Goodness, I really don't care for those things... The UAD at least does it with "more style" than most...

I probably should've given a little more history - At "another forum" there was this thread started with an image on it and simply - "How come I can't get my files to look like THIS?"

Look like? LOOK like?!?

Of course, the image was of a skating rink or a 2x4 - It was a commercial release of some sort that was squashed to bits in a not-so-friendly manner, and this poor guy was thinking that his mixes were no good because they weren't completely flat like the "pro" mix was.

This handy little experiment seemed like just the reason to discuss analog vs. digital dynamics (mostly limiting, I suppose) and why some dynamics can actually be good!

Any time I can get one person to not turn it up so darn loud, I feel like I've done something good.

[Tweak]

Well, you made an impression on me. I just turned in my re-mix submission to Dugz with no limiting. Your post was the final straw that broke the camel. Limiters get to be like crutches, at least for me, an easy way to get a uniform signal and avoid the leveling I should be doing at the track level.

I only have the L1,2 and 3. It's interesting that each new version of the Ultramaximer has gotten progressively more severe, allowing us to squash things to a flatter line than the last.

Thanks again!

[MASSIVE Mastering]

Whoops - sorry... No offense to your "L's"...

Even though I suppose, like anything other plug, they're only as dangerous as who's using them. It just shocked me how much "pressure" you could put on the UAD-PL before it fell apart.

But I really do hope this "loudness race" falls apart soon... Scientifically, it can't get much worse - But it really needs to go backwards so we can start listening again.

[Dugz Ink]

John... I really appreciate this post. Thanks for sharing it with us.

When I first joined this forum, everybody thought I was crazy because I "hated" compression... then they got to hear one of my mixes, B Ur Lover, and people started saying "Wait... that DOESN'T have compression?"

The truth is, I don't have the great analog gear that you have, so I didn't even try, because I knew that my software would crush the song to death... turning the bass and kick into a dull mess and ruining the dynamics.

So I did the best that I could, using some light limiting on a few tracks and the final mix... just enough to tame some stray peaks that tend to stick out like a windmill on a hill top.

Maybe we can solicit your services to master the winner of the remix contest. Then we could post that as an example of what good mastering sounds like, too.

D~s

[MASSIVE Mastering]

That could be fun...

But on the "gear" thing, I'm still a fan of making the most of what you have available... A tune I posted yesterday that I *think* sounds pretty decent (?) was done with nothing that would have the word "premium" on it, and totally in the box - NO outboard hardware.

It was a rush-rush recording and there were a lot of corners cut - but for the purpose it served, I think it came out okay.

As usual, the key point is that the band and their instruments sounded nice in the first place...

http://studio-central.com/phpbb/viewtopic.php?t=21883

[Modern Mastering]

Come on John, don't make fun of the "L" series , there are many of us who used to use that when we first started out. I still like the sound of the UAD effects but a nice analog limiter will always do better in almost every case. Nice post!

[Tweak]

I've heard the UAD is having trouble on some G5's in Logic. Can anyone confirm? I've wanted one a long time.

[MASSIVE Mastering]

I would think that if it's a known issue, it'd wind up in the Mac forum here...

http://www.chrismilne.com/uadforums/

[jsepeta]

I've heard the UAD is having trouble on some G5's in Logic. Can anyone confirm? I've wanted one a long time.

my buddy has a dual 2.5 g5 and still enjoys his UAD1, which used to be in a PC he was using last year. he's a cubase sx/nuendo guy though.

i've wanted to try the UAD myself, but I don't need the DSP just yet.

[mcatalao]

Hi!

There's a very elucidative article by Rip Rowan called Over The Limit, that talks about dynamics, and relates it to the development of the music/recording business through the last 20 years, and the loss of dynamics over the time!!!

http://www.prorec.com/prorec/articles.n ... 2E005DAF1C

[MASSIVE Mastering]

Got a PM from a member here with some questions & comments and thought it was worth revisiting...

One was the chain - Looking back at the sheet, the original file was 44.1kHz 24-bit. EQ'ing was handled in the DAW with UAD's Precision EQ. Loaded out through a Lavry M*DA 824 converter into a Manley Vari-Mu for some light compression. That ran into a Crane Song STC-8M for just a whisker of leveling and a final limiter. Back into a Lavry M*AD 824. The hot peak on the file was -0.6dBfs.

Another question was the visual part of the demonstration - How could a file that looked considerably louder actual be slightly quieter -

The answer isn't anything simple, but it can almost be summed up as "the magic of analog."

Well, that's not too far from the truth... Analog "grabs" things differently than digital - Digital waits for numbers of a particular value to do what it does. It normally knows when to expect those values, so it does it in a very "precise and orderly manner" while analog just "goes with the flow" of things in a very "organic" manner. The analog limiter in this case isn't try to "stop the signal at -0.1dB" It's simply trying to prevent clipping at whatever point I set it at. It's not a "true" brick wall and it's not anticipating the signal - but it's FAST - measured in the low microseconds (millionths of a second). It also doesn't know how hot the signal is going to get, so it takes care of it quickly and sets up for the next spike. It's smooth...

But that's not the only thing that can make a file *look* loud - I had a project in not long ago that the engineer was all worried about - The mixes were fairly quiet, but the file "looked" quite loud. When it got here, the problem was simple - He was using one of those "bass enhancement" plugins, and the file had video whine up around 17.5kHz (must've been a helluva refresh rate, as most VDO whine is around 15.8kHz).

Anyway, on his monitors and in his control room, he couldn't hear either of these anomalies. There were null points in his room and his monitors couldn't handle the lows (around 20-25Hz) and he couldn't hear the video whine over the video whine in the room. These two things together were taking up a good amount of sonic space, but were essentially inaudible on most systems. After filtering both problems out here, the file "looked" normal again (and sounded pretty nice also, although I had him remix instead without the goofy bass enhancer).

Gotta run, but I hope that helps a little. I'll try to hit this up again later, eh?

[tecnolover]

thanks for elaborating on this more. Much appreciated. Still not exactly sure how the 'analog magic' works ie. the 'mechanics' behind it, but it is very enlightening indeed. I think I'll need to research this phenomena further. What confuses me still is how the the analog file can mathmatically be calculated in RMS as louder than the digital one which 'visually' appears to peak higher across the entire wave form. Dynamic range is overall less! Afterall, dbRMS is basically a mathmatical calculation of the average of squares of all the peaks in the file? Is this RMS value weighted in some way possibly that might account for this?


** I started exporting my mixes out through a tube preamp like 6 months ago because it really improved my sound in marvelous ways. Namely, adding lots of warm, musical, 'even' harmonics to the mix. Now I think I'll do half of my mastering analogally (is that a word?! ). Ie. still do EQ ing and stereo widenening in the DAW. It was just so dang convenient to work entirely digital but now I think I'll change that.

Thanks a lot for sharing!

[MASSIVE Mastering]

Ah - But it's the energy between those peaks - The numbers get a little sketchy there...

Take a 1kHz tone and set it to -6dBfs - Depending on your DAW's setup, you'll probably get an RMS reading around -9dBRMS. Take a pink noise file at -6dBfs, and your RMS will probably only be around -18.

What's handling the peaks is important - What's handling what lies between those peaks can make a pretty huge difference also.

[jlgrimes11]

Nice info but the Precision Limiter was designed to be exactly what it said it was. It is designed to be a digital limiter it wasn't designed to sound analog so differences would be highly expected.

It would be nice if you could try the UAD LA2A's limiter and see if you get better results (which is supposed to simulate analog limiting).

[John Force]

This wealth of info needs a BUMP!

[Big Tim]

This wealth of info needs a BUMP!

True, but it confused the crap out of me with the mentions of a remix contest. Then I realised it wasn't 2005 any more...

[JDR]

This wealth of info needs a BUMP!

Thanks for bumping JF, this was EXACTLY the article I was looking for

[Weasel9992]

This wealth of info needs a BUMP!

True, but it confused the crap out of me with the mentions of a remix contest. Then I realised it wasn't 2005 any more...

That's hilarious...I thought *exactly* the same thing at *exactly* the same time in the thread.

Frank

[SupremeWise]

This is definetly some great info!!

Good thing I don't find myself using compression that often, actually it's rare if I do.

[MASSIVE Mastering]

But I really do hope this "loudness race" falls apart soon... Scientifically, it can't get much worse - But it really needs to go backwards so we can start listening again.

Wow... Once again, I was wrong, wrong, wrong. Three years has seen a lot more damage. I wonder what the next three will see...