The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
2buss comparison: Fatso vs. Oxford Dynamics Dynamics Plugins
Old 1st June 2006
  #1
Lives for gear
 
WunderBro Flo's Avatar
2buss comparison: Fatso vs. Oxford Dynamics

Slutz!
I just set up a little comparison I wanted to do for a long time and thought if someone here is interested I can post the resulting files!

I wanted to check out if it pays to go back to analog to do 2buss compression and then back in the DAW for limiting, or if I can get equally good sounding masters with just using plugins...

so here is the setup:
a rock´n´roll song mixed ITB then going through Sony Oxford Dynamics (compression 2:1 attack around 30ms, release around 80ms, GR 3-6dB) and alternatively into my Fatso (BussComp GR 3-6dB, no tranny, warmth set to 3 resulting in warmth-GR of about 2-4 on the meter, hitting the saturation pretty good so that both "0VU" and the "Pinned" were lit up most of the time)

then both mixes were level-matched and went through the exact same mastering eq and limiter setting.

here are the files, encoded as 320kbps MP3 files:

http://www.wildcowboys.com/rr_A.mp3
http://www.wildcowboys.com/rr_B.mp3

Check them out and see if you hear a difference and which one you like better!

My personal findings:
The difference is subtle and similar to my experience with OTB summing. The fatso version sounds a little bit more pleasing to my ear, a little more "like a record" (tm charles dye, haha), a little more "together" and "forward" - just the right thing for rock. I guess on R&B and hiphop I would not go the fatso route, it is not what those styles call for, but for rock it is worth the extra effort.


Some words about the song, it is called "Rockin´and Rollin´In The USA" by the group Tunesmith, a bit freaky and slightly retro, album will be released this year. ..

So tell me what you guys think about the differences or just tell me what you think of the song, production, mix, whatever!
Rock on,
Pat
Old 1st June 2006
  #2
Moderator
 
matt thomas's Avatar
Firslty they were quite similar, I don't think the choice of compressor is going to make or break the song. I don't think either was any better or worse.

B sounded a bit sludgier to me, but I liked it better at the same time, maybe with a slightly differnt compression setting, I think it suited the song. (I mean that in a good way)

I've never used either compresser so I don't know which is which, although I would guess that the "sludginess" is a product of the fatso "warmth"

oh, and nice shameless plug for the band by the way, cunningly disguised as a compressor comparison, lol

narco
Old 1st June 2006
  #3
Lives for gear
 
cramseur's Avatar
I liked the clarity of A better, but B sounded more cohesive...sludgier, "smearier" but at the same time more together...
Old 1st June 2006
  #4
Lives for gear
 
WunderBro Flo's Avatar
good observations guys! pretty much spot on! at least i hear what you are hearing! and i agree, the differences are very subtle. besides i think the sony oxford dynamics are among the best 2buss plugincomps out there, so i would have been dissapointed if there were more dramatic differences. too bad my waves ssl demo has already expired, i would have loved to try the ssl comp on the mix for comparison, as the fatso´s buss comp mode is modeled after the ssl!!!

any comments about the song or mix?
do you think the singer sucks? i am not sure if his tone is original or if it sucks! what do you think?
Old 1st June 2006
  #5
Lives for gear
 

Nice track. My personal opinions:

First of all I hear a big difference. Secondly, I find both to be over compressed. I can't fairly say which is which cuz, 1, I don't know the Oxford, and 2, I don't think the Fatso was used in a way I would use it.

The Fatso doesn't do very well at 6db of GR. 2 or 3 is good. Any more than that and it starts to sound weird. The manual suggests this as well. I know you want to hit the saturation hard, but compressing it like that is not the way to go IMO.

I too wanted to hit my Fatso hard, so I bought a 1968 to compress the mix, and then the Fatso to hit the saturation hard. Works for me. Kinda lost the use of the Fatso compressor, but, hey, we're all slutz here.
Old 1st June 2006
  #6
Moderator
 
matt thomas's Avatar
Quote:
Originally Posted by WildCowboys
do you think the singer sucks? i am not sure if his tone is original or if it sucks! what do you think?
Oh, he sucks for sure, but then again so do some of my favourite singers

narco
Old 1st June 2006
  #7
Lives for gear
 
WunderBro Flo's Avatar
haha, narco, cool way to look at it, i know what you mean! in rock a singer does not necessarily have to be good to be good!

thanks bump for your comment! yes i know i hit the compressors hard on this mix, i usually use no more than 3dB GR on most mixes, but with this stuff i really like the pumping - i was aiming for it! this is also why i wanted to do the comparison with this material, because it is more obvious this way. (if I did this on a popsong with just 2dB gain reduction, I think it does not matter which comp i use, the differences will be too subtle...) thanks for your feedback! btw. how does the 1968 compare to other 2buss comps you have used, is it special in a way? i would love to hear the 1968 and the portico comp on 2buss....

rock on!
Pat
Old 1st June 2006
  #8
yeah I hear a BIG difference.
Old 1st June 2006
  #9
Lives for gear
 
cramseur's Avatar
So
A= Sony Oxford Plug-in
B =Fatso

?
Old 2nd June 2006
  #10
Gear Guru
 
u b k's Avatar
 

well, they both sound equally overcompressed, WAY too much. why so overboard? it sounds like it was a great mix before the life got sucked out of it.

all the things i don't like are exaggerated in the tone of A, but i don't like the sound of either mix, nor do i like the compression which just saturates all the nastiness.

so if the purpose of this test was to pick the superior tool, my vote is for 'neither', at least the way they were used here. strapping something across the mix after it was mixed misses the point of mix compression, doesn't it? this doesn't sound glued, it sounds hammered.

i wish more people would read 'the purple cow.'


gregoire
del ubik
Old 2nd June 2006
  #11
Moderator
 
matt thomas's Avatar
I see a big difference between mosquitoes and elephants, I find different breeds of mosquitoes to be quite similar.

Anyway..

I actually think the song suited over compression (I'm with you there), but as I said before I would've played with the settings on the fatso a bit more, prehaps even tried it in parallel.


narco
Old 2nd June 2006
  #12
Gear Maniac
 

Subtle difference ?!

Well ... I never used the fasto yet so take it for what it worth, but within fews second of listening, my opinion was that A is Sony plugin while B is Fatso.
Not even close sounding to my ears.

Or maybe I just prefer the sound of the plugin over hardware ! ...

I say B is Fasto heh

Salvator
Old 2nd June 2006
  #13
Lives for gear
 
Ruudman's Avatar
 

Agree, B is Fatso.

ruudman
Old 2nd June 2006
  #14
Lives for gear
 
WunderBro Flo's Avatar
Yup, everybody´s guess was right, A is Oxford and B is Fatso!

So to sum it up, yes, it has been overcompressed A because I like it for this song and B because I wanted to make the difference in sound as audible as possible...guess I achieved that.

funny that some of you say you hear a BIG difference, I hear a small difference and I dare to say that 90% of the average listener would not hear any difference at all. interesting! also interesting that the artist and I both liked the overcompression a lot and we both thought it actually added "life" instead of sucking it out!

so for the final mastering I will use the fatso and hit it not as hard, 2 dB less GR would not hurt I guess!


thanks everyone for your comments!
Rock on!
Pat
Old 2nd June 2006
  #15
Lives for gear
 
Zep Dude's Avatar
 

I don't know that you can draw any generalizations about out of the box bus compression vs in the box based on this test. I don't consider the Fatso a good bus comp (although I love it on individual tracks) and also, the converters can make a big difference when evaluating outboard comps.

That 1968 someone mentioned earlier with a good set of converters will blow away any plugin I've heard on the bus -if you're looking for punch and thickness. For in the box, the waves linear multiband is very nice if you don't want too much color but just a nice overall bus comp.
Old 2nd June 2006
  #16
Gear Nut
 

HI , in my humble Opinion, the difference is huge ... and im sorry to know yet that A) is the Oxford ... cause i like it much better, yes im a Analogue Slut ... B) is defently too much , the Comp is pumping that in the Intro some of the Snares Peaks are gone ... nobody else hear that ?? but A) is almost too much, actualy it is ... u can feel it pump as well on certain Volume settings ... not Club Compatible ...when the Guitars came in, after the Drum-Intro, the Drums disapear which takes away the huge Feel of the Song ... for the right Sound Arrangement i'd like it more if every next Instruments came on Top so that it gets bigger and bigger, the Sound goes better with the Arrangement then, after Refrain back to the Verse it gets a bit smaller and thiner then, ach im sure u guys know what i mean

dont know your settings on both examples , but i would make less GR and less Release and a bit more Attack ...

but i think it rocks so far !
nice

greeetz
Old 2nd June 2006
  #17
Gear Head
 

I'd like to experiment a bit with some of this myself but there's a piece I don't really understand. Maybe someone can give me a pointer or two.

First, generally speaking, it seems to me I get the best results on my DAW when 0 on an analog VU meter equates to approx. -14dBFS on my AD and DAW. If I go hotter than that on my AD (Lucid, btw), the sound is harsh/brittle sounding. So, this setup usually works out ok for me.

Let's say ITB, I have a comp and limiter on the 2Buss, performing a pseudo master. My peak level is .-3db with say an average RMS of approx. -15dB. So, in prep to go OTB, I disable the stuff on the 2Buss. I run this signal out to analog gear, peaks hitting around 0dBVU, eventually through the Lucid back to DAW. If I don't deviate from my normal approach, I'll end with peaks around -14dBFS back at the DAW. Not what I'd want if this was suppose to be the "master".

I haven't cared for the sound too much when I hit the Lucid close to 0dBFS. But, if I don't do this, my OTB "master" is going to be considerably lower than the ITB version.

So, I'm missing an approach to this for my own experimentation. Are my ears deceiving me? Should hitting the Lucid with peaks @ just under 0dBFS sound as good to me as hitting it with peaks @ -14dBFS? Assuming my ears are just not shot, is it just that some converters are better handling this type of gain? Adding gain once the track is back in the DAW doesn't sound like the right thing to do.

I'm not sure I really want to do this in the first place, but I'd like to play with it for some first-hand experience and to decide for myself if there's any benefit. Maybe I have some fundamental issues with my approach. Any guidance or suggestions would be appreciated.

Thanks,
Steve
Old 3rd June 2006
  #18
Gear Guru
 
u b k's Avatar
 

Quote:
Originally Posted by Steve T
Are my ears deceiving me? Should hitting the Lucid with peaks @ just under 0dBFS sound as good to me as hitting it with peaks @ -14dBFS? Assuming my ears are just not shot, is it just that some converters are better handling this type of gain? Adding gain once the track is back in the DAW doesn't sound like the right thing to do.

your ears are not deceiving you. ime, the analog front end of a/d converters tends to sound best with levels averaging -14, peaking at -10.

the way 99.9% of all folks handle the mastering is they use a digital brickwall for the final limiting and gainstaging. this is either done in a daw using a plug, or it's done in hardware like the L3 or finalizer and then transferred *digitally* back into the daw.

if you're doing your limiting in the analog realm with an analog box, record the signal back into your daw at modest levels, then either normalize the file to get the peaks up to the ceiling, or use the aforementioned brickwall plug to get the gain up but with no additional limiting other than the occasional errant peak.


gregoire
del ubik
Old 3rd June 2006
  #19
The Distressor's "daddy"
 
Dave Derr's Avatar
 

USING ALL THE BITS

Quote:
Originally Posted by Steve T
First, generally speaking, it seems to me I get the best results on my DAW when 0 on an analog VU meter equates to approx. -14dBFS on my AD and DAW. If I go hotter than that on my AD (Lucid, btw), the sound is harsh/brittle sounding. So, this setup usually works out ok for me.
Thanks,
Steve
THIS SOUNDS VERRRRY SUSPICIOUS! ADCs should be most linear and accurate when every last dB of headroom is used. By throwing away 14dB on the top, you are losing 2.4 bits! So if you were working at 20bits, then letting it peak 14dB down essentially means you are working at 17.6 bits. Maybe I am misunderstanding but it sure sounds like you are throwing way some good resolution... or maybe the LUCID has some high level problem.

I do know that some waveforms including square waves can clip the digital filters on ADC's and DACs without showing up as clipping. I also wonder if there could be an input buffer problem or shorted node in the wiring that starts to sound nasty as it is pushed harder with hot levels.

Boy it would be fun to know what you're hearing. It scares me when we are working in a theoretically derived medium (digital audio), and what we hear doesn't agree with some of the basic theory. As we all know, there have been false promises and just plain "math errors" in digital audio before...
Old 3rd June 2006
  #20
Gear Head
 

Thanks for the responses, guys.

Gregorie: Thanks for validating that pushing my converter hard might have some unpleasant artifacts. Several times, I've just closed my eyes to dial in what sounds "right" on the input to this converter, and I'm always within that range. In my case, I'd have to do the limiting back in the DAW. I did think of that but was frankly a little skeptical how adding that much gain back at that stage would sound. Looks like I need to play around with it some.

Dave: Thanks for the questioning response. I posted because I wasn't sure if what I was hearing was "normal". I really need to perform some more tests.

I should amend something I stated earlier, though. I was mainly commenting on levels when running a mix out of my DAW back eventually through a converter. So maybe some of this has to do with the density of the source, if that makes any sense. With individual tracks, the sweet spot can vary (for me) somewhat, but still -14 is a good target for me from a holistic view.

I do appreciate your comments. Thx.

Steve
Old 4th June 2006
  #21
TML
Lives for gear
 

Use both. Both work. The fatso sounds better in most ap's......not even close even without the att and rel controls......use the 384 all the time....fatso doesn't feel the same my 2 cts
tim
Old 4th June 2006
  #22
Lives for gear
 
Jazzpunk's Avatar
 

It seems to me that this track was originally intended to stay ITB with the OTB comparison thrown in as an after thought. I suspect the end results would've been quite different had you been planning (ie anticipating the colors the fatso would impart during the actual mix) to use the fatso on the 2bus from the beginning.

How about a real comparison with two versions of the same track, each mixed from scratch with their respective compressors on the 2bus from the get go?

(Or am I misunderstanding and this was an 'Oxford plug vs. Fatso' for mastering comparison?)
Old 4th June 2006
  #23
Lives for gear
 
Sui_City's Avatar
 

Quote:
Originally Posted by Dave Derr
THIS SOUNDS VERRRRY SUSPICIOUS! ADCs should be most linear and accurate when every last dB of headroom is used. By throwing away 14dB on the top, you are losing 2.4 bits! So if you were working at 20bits, then letting it peak 14dB down essentially means you are working at 17.6 bits. Maybe I am misunderstanding but it sure sounds like you are throwing way some good resolution... or maybe the LUCID has some high level problem.

I do know that some waveforms including square waves can clip the digital filters on ADC's and DACs without showing up as clipping. I also wonder if there could be an input buffer problem or shorted node in the wiring that starts to sound nasty as it is pushed harder with hot levels.

Boy it would be fun to know what you're hearing. It scares me when we are working in a theoretically derived medium (digital audio), and what we hear doesn't agree with some of the basic theory. As we all know, there have been false promises and just plain "math errors" in digital audio before...
Sorry Dave,

But am i missing something here?

Is calibration of a DA not done using gain pots on the analog side of the conversion? If so, how would that lead to a loss of 2.4bits?

I understand your concern about a converter that responds differently at different levels. That would make me run a mile. But the calibrating of -14/-16/-18dbfs to 0vu is a pretty standard thing, is it not? Allowing 14/16/18db (etc) for peaks.

Or am i just fukked?
Old 4th June 2006
  #24
The Distressor's "daddy"
 
Dave Derr's Avatar
 

FS Digital

Trimming the input to obtain a Full Scale digital conversion is the right thing to do, but it sounded to me like Steve T was actually hitting the ADC at a less than full scale signal, meaning he wasnt using all the bits. If it was just a matter of adjusting the operating levels so the converter worked at full scale, just below clipping, then my comment was irrelevant, and I apologize for the confusion.
Old 6th June 2006
  #25
Gear Guru
 
u b k's Avatar
 

there are two separate issue here: bit resolution, and analog gain staging.

digital resolution aside, my experience is that the analog stages of converters don't do as well when hit with the kind of levels that approach 0dbfs. since my place is calibrated so that 0vu = -14fs, running a compressed 2mix into the converters near the top of their meters requires ridiculous headroom from the converters and the output stage before them, and things always sound less choked to me when i keep things more conservative.

and my understanding is that signals that average -10dbfs have more than enough bits to capture the dynamics of the signal their getting. am i off?


gregoire
del ubik
Old 7th June 2006
  #26
The Distressor's "daddy"
 
Dave Derr's Avatar
 

Full Scale Digital Signals

Quote:
Originally Posted by u b i k
there are two separate issue here: bit resolution, and analog gain staging.
and my understanding is that signals that average -10dbfs have more than enough bits to capture the dynamics of the signal their getting. am i off?
If you're talking about capturing every bit of dynamic range without clipping, one would want to have the very loudest point hit 0dBfs. Its hard to say where the average should be since a drum kit for instance, would probably average a bit lower than -10dBfs if you let the loudest strokes hit 0dBfs. A full contemporary rock Mix would probably average -3 - 5dBfs with the snares and kicks hitting 0dbfs. I'm only guessing here, but average vs peak is hard to pin down. The important thing to get maximum resolution, is that at some point you get verrrrrry close or right up to 0dBfs.

However, You are right in that modern 20 - 24bit converters are so quiet and have such a large dynamic range that you could probably let -10dBfs be a peak level and still not hear any difference when its all said and done. But you are losing almost 2 bits of resolution.

These days I think people get lost in technological details, and if they get the big things right like the SONG, the arrangement, performance, EQ, and compression close... losing a bit here or there or using an old 16 bit converter at 44.1KHz will hardly mean a thing when its over.
Old 7th June 2006
  #27
Lives for gear
 

I've been reading some stuff by Paul Frindle, and he believes the "resolution" theory of digital audio is completely false. (Interesting to do a google on his name and read his ideas).

If you get hung up on this idea of "resolution" and "using/losing bits", you end up recording too hot, and getting a less accurate waveform.

I find it interesting that 24 bit audio has a theoretical dynamic range of 144dB. But the very best converters are lucky to reach 120dB. My Lucid AD9624 has a dynamic range of 114dB.

144dB - 114dB = 30dB. That suggest to me that if you are trying to digitize a waveform and achieve peaks right up to 0dBFS - you are exceeding the analog limits of the dynamic range of that converter (by 30dB!). The signal must therefore be damaged, so in the quest for "more resolution", you have unintentionally distorted the waveform.

I don't mean digital clipping - exceeding 0dBFS. I mean well before you reach 0dBFS, you are saturating the analog section of the converter, and probably getting a very harsh sound.

I used to worry about "empty makeup gain" - but i'm less concerned with that now. Any loss due to using slightly less bit depth is probably going to be waay less than any damage caused by saturating the analog stage of a converter.

As an example of what I mean, consider the reverse (hitting a D/A converter). Create a sinewave that peaks at 0dBFS, and then listen to it as you play with the master fader. With cheap converters, you can hear the harmonics being added to the sinewave the closer you get to 0dBFS. With my Benchmark, it's much less obvious.

So I believe exactly the same thing happens with A/D converters - I just haven't set up a test to prove it.
Old 8th June 2006
  #28
The Distressor's "daddy"
 
Dave Derr's Avatar
 

Quote:
Originally Posted by Kiwiburger
I find it interesting that 24 bit audio has a theoretical dynamic range of 144dB. But the very best converters are lucky to reach 120dB. My Lucid AD9624 has a dynamic range of 114dB.

144dB - 114dB = 30dB. That suggest to me that if you are trying to digitize a waveform and achieve peaks right up to 0dBFS - you are exceeding the analog limits of the dynamic range of that converter (by 30dB!). The signal must therefore be damaged, so in the quest for "more resolution", you have unintentionally distorted the waveform.
Actually, its not that you are clipping the top end, its that the last LSBs (least significant bits) are virtually trash. You don't lose resolution on the top end of the dynamic range, you lose it on the low end of the dynamic range.

All good converters should perform best when all the bits are used. This is almost by definition, and why the distortion specs are always measured with a full scale sine wave. It's possible that there are math errors in the filtering, or bad converters that have Full scale problems, but usually designers do all their testing... both audible and equipment testing, at or near full scale signals. It would be a bad design that had the analog front end clip before the ADC clips, especially if metering didnt show it! A sign of a really good converter is when you can put in a really low level signal, say -40 or -60dBfs, and still have respectable distortion. But everyone who designs and uses ADC's kind of assumes the high level or full scale signals MUST BE EXTREMELY CLEAN.

The big advantage of modern high resolution converters IE, above 16 bit, is that you hopefully have so much "quality" and resolution, that if you leave a little headroom, you aren't degrading the signal very much.

By the way, there are really very few converters who put out "perfect" resolution, that is, where every bit they put out has a full 6dB of dynamic range. Usually, a 16 bit converter will work around a perfect 15 bits, a 20 bit converter often only works at 18 bits, 19 at best, and 24 bit converters.... well, the last 5 bits are often in name only! I think it was George Massenburg who called them "Marketing bits". I have complained before saying that at least one spec must meet the theoretical spec insinuated in the converters resolution. Hell, I can make you a 100 bit converter, that truly will put out words 100bits in length...with at least 17 of them useful! Any takers?
Old 9th June 2006
  #29
Lives for gear
 
norman_nomad's Avatar
Quote:
Originally Posted by Dave Derr
Actually, its not that you are clipping the top end, its that the last LSBs (least significant bits) are virtually trash. You don't lose resolution on the top end of the dynamic range, you lose it on the low end of the dynamic range.

All good converters should perform best when all the bits are used. This is almost by definition, and why the distortion specs are always measured with a full scale sine wave. It's possible that there are math errors in the filtering, or bad converters that have Full scale problems, but usually designers do all their testing... both audible and equipment testing, at or near full scale signals. It would be a bad design that had the analog front end clip before the ADC clips, especially if metering didnt show it! A sign of a really good converter is when you can put in a really low level signal, say -40 or -60dBfs, and still have respectable distortion. But everyone who designs and uses ADC's kind of assumes the high level or full scale signals MUST BE EXTREMELY CLEAN.

The big advantage of modern high resolution converters IE, above 16 bit, is that you hopefully have so much "quality" and resolution, that if you leave a little headroom, you aren't degrading the signal very much.
This is my understanding as well, but when talking in terms of a "little headroom" what is your best practice? I try to keep my peak signals at around -14dbfs to -6dbfs tops... would I really gain much of anything by pushing this forward... or upward a little?

I've been under the impression that I could even let my peaks fall lower if I wanted to, because even at lower levels, the noise floor and quantization distortion of a well implemented 24 converter is unlikely to creep into the audible range...
Old 9th June 2006
  #30
Lives for gear
 

Quote:
But you are losing almost 2 bits of resolution.
Are you really "losing" two bits of resolution? To actually "lose" them wouldn't you have to be recording a signal with a wide enough dynamic range that you're approaching the noise floor of the converters? If you're recording a signal with 40 or even 60 dB of dynamic range...which is a lot...then peaking at -10 dBFS or -20 dBFS or lower is really no big deal in a 24-bit system.

Quote:
Hell, I can make you a 100 bit converter, that truly will put out words 100bits in length...with at least 17 of them useful! Any takers?
I'm sure there would be people who would read about it and would just have to have it so they could get the "most bits" possible.

-Duardo
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
XHipHop / Music Computers
20
Ghost Logic / Gearslutz Secondhand Gear Classifieds
0
benelli / Gearslutz Secondhand Gear Classifieds
0
Dutchmuzik / Gearslutz Secondhand Gear Classifieds
0

Forum Jump
Forum Jump