The No.1 Website for Pro Audio
Real life experiences with external summing (ITB)?
Old 25th May 2006
  #31
Lives for gear
 

Jitter shouldn't be a factor in ITB summed mixes. Digital audio files are basically lists of numbers. The summing process is an order mathematical process that combines these files in a strict orderly fashion.

While your DAW is playing in realtime, it's controlled by your system clock. This clock - depending on quality of design/build and your AC power quality in particular, will have some amount of jitter - nothing hardware is ever perfect, so there is going to be some instability in your clock. However - the rendered file is unaffected by this, because it's simply an orderly list of numbers. It's not much different from summing a spreadsheet of figures - it doesn't matter how erratic the timing of the clock is, the numbers will still be summed accurately and in order.

So with a totally ITB mix, jitter is irrelevant. You can burn the audio CD, and jitter won't be a factor - not until it's played in a jittery consumer CD player.

With any OTB process - or even just recording the streaming output of your DAW to another recorder, jitter will be compounded at every turn.

So the price difference between ITB and OTB will always be horrendous because you absolutely need good conversion and stable clocks.

But I really think the 'benefits' of analog summing could be faked digitally to get a similar effect.
Old 25th May 2006
  #32
Lives for gear
 
RKrizman's Avatar
 

Ah, the real Dr. Science. Thank you for the patient and cogent explanation.

-R
Old 25th May 2006
  #33
Lives for gear
 

Actually - just noticed an error in what I said. At the point when you burn the CD, there is considerable room for jitter and error if it's an audio CD.

But if it's a data file (wav), those errors are corrected.
Old 25th May 2006
  #34
Lives for gear
 

I read your interesting points... I view it slightly differently. Someone said jitter is not present when you don't sync digital devices. As far as I know jitter is always present in a conversion process, because the whole idea is that the signal is converted into samples on a certain interval. This is driven by a "crystal reference clock" that is not 100% stable. The instability creates what we call jitter. This jitter, which is present on each track on the whole song as digital numbers (representing samples), part in the input that is feeded into each digital process. We hear this jitter as:

- deteriorated ease of listening
- deteriorated clarity
- deteriorated high frequency response
- worse instrument separation
- less information
- worse timing
- worse soundstage
- deteriorated overall audio performance

among other things as a result of degraded SNR and THD+N. It would be cool if these artifacts would be isolated from the rest of the digital processing, the problem is that it's not. As I already said it's part of each calculation in each digital process. So it's totally dependent on the digital algorithms how much the jitter will damage the signal. Let's say we put a "bad" effect on the mix output bus and have a number of "bad" effects on different input tracks. This means the signal is first converted with jitter, then it enters all the enabled digital processes on the channel strip including the "bad" effect, where each process includes a number of calculation errors, not only due to lack of bit depth precision but also because of the calculation of jitter. Then the signal enters the summing process where the signal on each input channel is distributed to the L and the R channel. This is the critical process, because now the jitter on each track will be the input to the summing process, so not only do we get calculation errors due to lack of bit depth precision, but also due to the calculation of jitter and the weaknesses in the summing algorithm. The result is summed into mix jitter, present on the L and the R channel. Then the signal enters the mix output channel strip and goes through all the digital processing on that channel, including that "bad" mix bus effect (for instance a bad limiter setup for hard limiting). At this point the damage due to calculation errors is of course much worse, since the signal has become worse and worse on the way to the mix bus as a result of all calculations. Now imagine what happens when a not very intelligent bit reduction+dithering is applied on top of the very jittery digital stream...!

By summing externally we can avoid the process when the jitter becomes most prominent (summing and bit reduction+dithering), at the cost of analog signal damage and two additional conversion processes (that adds more jitter, but is not as prominent since the number of processes after it is low). The whole point why I posted this thread was basically to find out what "damage" we prefer by asking what kind of real life experiences you have had with external summing. Someone said they experienced better "depth", which might be true since we know that a jittery signal is perceived with "worse instrument separation" and with "less information", among other things...

EDIT:
I don't think it's likely that external summing sounds better, basically because it requires such a good setup. But it might be that we prefer "analog damage" much more than "digital damage" and the sonic performance gain actually just lies in the "compensation". I still think though that if this is the case we need a really low jittery clock and really good op amps in order to achieve it...

EDIT2:
I found a great CD for this and I think I will leave it at that. It's much easier to draw conclusions when there's a lot more material to process and it's likely that it is much more precise as well...
The CD can be found at: http://www.3daudioinc.com/catalog/pr...5bcc13f34cc7c7
Old 25th May 2006
  #35
Gear Maniac
 
Mark Cattano's Avatar
 

I am certainly no scientist, but FWIW, I was in Manhattan last night at a seminar with a friend of mine who is an editor for Mix Magazine. He introduced me to this cat from New Jersey whom he described as "The Conversion King". He's got this place called Soniclab. This guy is like some genius mad scientist. I grabbed him and spoke to him for about a half an hour. I told him how much I thought my D2B improved the imaging and depth of my mixes. He told me in lay terms that what I was digging was the DISTORTION. You know what? I don't know the science of it- I just know it sounds better.

Mark Cattano
Magneto Studios
www.heedmusic.com
www.myspace.com/markcattano
Old 25th May 2006
  #36
Lives for gear
 
jumpnyc's Avatar
 

I use a D2Bus - Here's why I do it - I use a bunch of outboard when I mix and I would rather do one D/A and then back A/D than multiple conversions on separate tracks. I go out from 16 channels of D/A on my Pro tools 192 i/o and come back into Pro Tools with a Lavry Gold MK3.
Old 25th May 2006
  #37
Lives for gear
 

Quote:
Originally Posted by Mark Cattano
I am certainly no scientist, but FWIW, I was in Manhattan last night at a seminar with a friend of mine who is an editor for Mix Magazine. He introduced me to this cat from New Jersey whom he described as "The Conversion King". He's got this place called Soniclab. This guy is like some genius mad scientist. I grabbed him and spoke to him for about a half an hour. I told him how much I thought my D2B improved the imaging and depth of my mixes. He told me in lay terms that what I was digging was the DISTORTION. You know what? I don't know the science of it- I just know it sounds better.

Mark Cattano
Magneto Studios
www.heedmusic.com
www.myspace.com/markcattano
Interesting! Thanks for sharing this experience with us... I think that another way of describing this phenomenon could be "absence of digital distortion", that might be what that cat from New Jersey was trying to say. (only my theory though...take it for what it is...)
Old 25th May 2006
  #38
Lives for gear
 

Quote:
Originally Posted by jumpnyc
I use a D2Bus - Here's why I do it - I use a bunch of outboard when I mix and I would rather do one D/A and then back A/D than multiple conversions on separate tracks. I go out from 16 channels of D/A on my Pro tools 192 i/o and come back into Pro Tools with a Lavry Gold MK3.
So you record to the DAW and then route the channels out to the analog gear, sum it and converter the stereo signal directly to the target sample rate without dithering? I would really like to hear the difference in doing the whole process like this...
Old 25th May 2006
  #39
Lives for gear
 
Ribbonmicguy's Avatar
Quote:
Originally Posted by jumpnyc
I use a D2Bus - Here's why I do it - I use a bunch of outboard when I mix and I would rather do one D/A and then back A/D than multiple conversions on separate tracks. I go out from 16 channels of D/A on my Pro tools 192 i/o and come back into Pro Tools with a Lavry Gold MK3.
Jump, you slut!

I wanted to have that Lavry Gold for my mixdown A/D as well!

How does it sound? Help you get your sound faster?
Old 25th May 2006
  #40
Lives for gear
 
jumpnyc's Avatar
 

Quote:
Originally Posted by RainbowStorm
So you record to the DAW and then route the channels out to the analog gear, sum it and converter the stereo signal directly to the target sample rate without dithering? I would really like to hear the difference in doing the whole process like this...
Well...I need to let you know that I also spank it in the analog domain when I hit the 2bus - which has a ton of headroom - so it's not a direct comparison. I would have to bring the levels down considerably to do the same thing ITB.

For me and YMMV - this process sounds great. I am also using premium outboard gear - STC8, Atomic Squeezebox, etc.
Old 25th May 2006
  #41
Gear Maniac
 
Mark Cattano's Avatar
 

Hey Jump

Do you go out of the computer, into the D2B, then right back into the computer, or do you use hardware (compressor/pre/eq) after the D2B? Just wondering what your process is.

Mark
Old 25th May 2006
  #42
Lives for gear
 
jumpnyc's Avatar
 

Pro tools D/A -> Patch bay -> Patch in goodies -> 2Bus ->CraneSong STC 8 -> Lavry Gold - > Pro tools or CD Burner.
Old 25th May 2006
  #43
Gear Maniac
 
Mark Cattano's Avatar
 

Thanks Jump. BTW, nice signal chain.


Mark Cattano
Magneto Studios
Old 25th May 2006
  #44
Lives for gear
 
RKrizman's Avatar
 

Quote:
Originally Posted by RainbowStorm
This means the signal is first converted with jitter, then it enters all the enabled digital processes on the channel strip including the "bad" effect, where each process includes a number of calculation errors, not only due to lack of bit depth precision but also because of the calculation of jitter. Then the signal enters the summing process where the signal on each input channel is distributed to the L and the R channel. This is the critical process, because now the jitter on each track will be the input to the summing process, so not only do we get calculation errors due to lack of bit depth precision, but also due to the calculation of jitter and the weaknesses in the summing algorithm. The result is summed into mix jitter, present on the L and the R channel. Then the signal enters the mix output channel strip and goes through all the digital processing on that channel, including that "bad" mix bus effect (for instance a bad limiter setup for hard limiting). At this point the damage due to calculation errors is of course much worse, since the signal has become worse and worse on the way to the mix bus as a result of all calculations. Now imagine what happens when a not very intelligent bit reduction+dithering is applied on top of the very jittery digital stream...!
I think you're torturing your point a little, here. When you record a signal to digital, the jitter of the A/D's clock will affect the integrity of the signal. But once it's in the DAW, that effect has already been encoded on the track. At that point, if you're mixing ITB, the jitter of your clock no longer matters--it will not affect the outcome. All the numerical calculations still occur in the same order, whether your clock is lurching or not. As long as you stay totally in the digital domain, you can do two BTD's, one with a jittery clock and one with a less jittery clock, and the resulting files will be identical.

Now, if you choose to sum those tracks analog, then the jitter of the D/A clock will matter and will affect the sound.

It's one thing to prefer one method over another--it's another thing to know what accounts for the differences.

-R
Old 26th May 2006
  #45
Lives for gear
 

Quote:
Originally Posted by jumpnyc
Well...I need to let you know that I also spank it in the analog domain when I hit the 2bus - which has a ton of headroom - so it's not a direct comparison. I would have to bring the levels down considerably to do the same thing ITB.

For me and YMMV - this process sounds great. I am also using premium outboard gear - STC8, Atomic Squeezebox, etc.
Cool... Ok, so that "spanking" process will of course compensate a lot in this way besides the external summing. I'm just curious, what does the waveform look like once the stereo signal has returned to the DAW? Is it spiky or "smooth" in terms of transient peaks?
Old 26th May 2006
  #46
Gear Maniac
 
tunasafedolphin's Avatar
 

Quote:
Originally Posted by RainbowStorm
Cool... Ok, so that "spanking" process will of course compensate a lot in this way besides the external summing. I'm just curious, what does the waveform look like once the stereo signal has returned to the DAW? Is it spiky or "smooth" in terms of transient peaks?
And how does it matter what it looks like?
Old 26th May 2006
  #47
Lives for gear
 
amost's Avatar
 

Quote:
Originally Posted by RKrizman
I think you're torturing your point a little, here. When you record a signal to digital, the jitter of the A/D's clock will affect the integrity of the signal. But once it's in the DAW, that effect has already been encoded on the track. At that point, if you're mixing ITB, the jitter of your clock no longer matters--it will not affect the outcome. All the numerical calculations still occur in the same order, whether your clock is lurching or not. As long as you stay totally in the digital domain, you can do two BTD's, one with a jittery clock and one with a less jittery clock, and the resulting files will be identical.

Now, if you choose to sum those tracks analog, then the jitter of the D/A clock will matter and will affect the sound.

It's one thing to prefer one method over another--it's another thing to know what accounts for the differences.

-R
Okay well you know what I just got 002 and had a Lynx card so I've been experimenting. I thought I was hearing differences listening to the same file one clocked with the 002, the other the Lynx and I might have been but when I BTD....the audio statistics were exactly the same and when I phase reversed one....silence. That's the first time that null test has ever been completely silent for me.
Old 26th May 2006
  #48
Lives for gear
 

Quote:
Originally Posted by tunasafedolphin
And how does it matter what it looks like?
It is only a sign of how much potential noise artifacts there is within the waveform... A very smooth waveform will of course be much easier to work with further down the line, than when you have jittery transients as a result of bad errors in the digital processing. Some of these distortion transients can of course be due to distortion within the mic as well, which is why it is important to record with a mic that can handle a lot in the capture moment without distorting.
Old 26th May 2006
  #49
Lives for gear
 

Quote:
Originally Posted by amost
Okay well you know what I just got 002 and had a Lynx card so I've been experimenting. I thought I was hearing differences listening to the same file one clocked with the 002, the other the Lynx and I might have been but when I BTD....the audio statistics were exactly the same and when I phase reversed one....silence. That's the first time that null test has ever been completely silent for me.
I'm talking about recording directly to the DAW software, so this is an AD conversion process. You can't isolate clock stability in any simple way as far as I know, that requires deep level jitter measurements and needs to be done at a very stable room temperature...
Old 26th May 2006
  #50
Lives for gear
 
RKrizman's Avatar
 

Quote:
Originally Posted by amost
Okay well you know what I just got 002 and had a Lynx card so I've been experimenting. I thought I was hearing differences listening to the same file one clocked with the 002, the other the Lynx and I might have been but when I BTD....the audio statistics were exactly the same and when I phase reversed one....silence. That's the first time that null test has ever been completely silent for me.
You may hear a difference listening to the same file with 2 different clocks, because in the act of listening you are performin a D/A process. The file itself is the same, but it sounds different because you're using different clocks to create analog for your monitoring.

OTOH, if you take a multitrack session and do a BTD using once clock, and then do a BTD using another clock, those 2 files will be identical (except for 2 factors--the dither, being random, will be different on each pass, as will any reverb or cycling time-based effects--just because they are different each pass, regardless of clocking.) So when you listen back to those 2 files using the same clock for monitoring, you shouldn't hear a difference (except for the aforementioned caveats).

-R
Old 26th May 2006
  #51
Lives for gear
 
amost's Avatar
 

Quote:
You may hear a difference listening to the same file with 2 different clocks, because in the act of listening you are performin a D/A process. The file itself is the same, but it sounds different because you're using different clocks to create analog for your monitoring.
Yeah, made sense to me as I was doing it.
Quote:
OTOH, if you take a multitrack session and do a BTD using once clock, and then do a BTD using another clock, those 2 files will be identical (except for 2 factors--the dither, being random, will be different on each pass, as will any reverb or cycling time-based effects--just because they are different each pass, regardless of clocking.) So when you listen back to those 2 files using the same clock for monitoring, you shouldn't hear a difference (except for the aforementioned caveats).
-R
Good info. I'd wondered why the null test had never really been dead on for me before and it's probably because the mixes had time based plugs on there I guess.
Old 27th May 2006
  #52
Lives for gear
 

Been using a dangerous 2 bussinto an Oram sonicomp for about 18 months love it no way Im going back to just ITB especially if Im doing indie bands.
Old 27th May 2006
  #53
Gear Maniac
 
Mark Cattano's Avatar
 

Thought I'd chime in one more time here. I've had the D2B for three days now, and all I can say is WOW. As per the previous post, there's no way I'd ever go back to mixing ITB, now that I can hear my mixes summed analog. It really does sound like listening through a high end analog console. Simply stunning sound, especially when I think about what it sounded like before I got the D2B.

Mark Cattano
Magneto Studios
www.heedmusic.com
www.myspace.com/markcattano
Old 27th May 2006
  #54
Lives for gear
 

Quote:
Originally Posted by Mark Cattano
Thought I'd chime in one more time here. I've had the D2B for three days now, and all I can say is WOW. As per the previous post, there's no way I'd ever go back to mixing ITB, now that I can hear my mixes summed analog. It really does sound like listening through a high end analog console. Simply stunning sound, especially when I think about what it sounded like before I got the D2B.

Mark Cattano
Magneto Studios
www.heedmusic.com
www.myspace.com/markcattano
Very interesting! So at least one of us finds analog summing much better sounding...! I'm sure the difference is not "that" huge, but the small and right improvements can have a big WOW effect...! Do you think you could post A and B clips? That would be awesome...!
Old 27th May 2006
  #55
Lives for gear
 
tomdarude's Avatar
gimmie

2 Lynx Aurora16
1 Lavry GoldmkIII A/D

and a 32ch. Tonelux with faders, busses & EQ and a bunch of outboard



....and there´ll be no way I go back to ITB heh



ps: ...and I promise to never complain about jitter!
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
Forum Jump
Forum Jump