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
Do you use different AD Converters tracking same song? Audio Interfaces
Old 9th January 2013
  #121
Quote:
Originally Posted by emitsweet View Post
seems like a trend , everything appears to be a misunderstanding when I step in, but luckily you have the understanding of thing others miss but we appreciate your knowledge it has made me a better engineer for sure. I recall 3 weeks ago we went back and forth about "signal" and DAWs. Good thing you set me straight on that too. Now I know data structures and registers process signal.

I will add, clocking in all A/D converters I have worked on will add additional latency due to the zero cycle void. This is very very very small delay. No where near 5ms. So it would not impact a phase/time/delay relationship except for very extreme
RT applications. used in conjunction with a RTOS.

So I suppose with higher end A/D such as lavry, additional jitter/error detection filtering, noise filtering cause increases delays is what you are saying?

As long as you don't claim what is coming out of a A/D output is a "signal" we are good here. If you can acknowledge it is in fact data, I can concede that based on your experience, expertise and knowledge of A/D conversion that only Audio A/D are exempt form zero cycle latency. Is that true?

I am happy I learned something here. It is no longer a misunderstanding. It is now data I can store in my neurological unit. I guess I was a little overzealous (falsely) assuming ADCs used in Audio converters are not that much different than other ADCs used in other applications and have similar characteristic specifically with regards to clocking and latency. Since every ADC I have used whether SA, Counter ramp, ladder, Binary weighted and or parallel encoded used in data acquisition, have latency due to gaps in clock cycle respective to analog input signal being encoded into binary data. I guess ADCs used in Audio A/D are exempt from this inherit flaw?

Based on your expertise and vast knowledge of A/D conversion what specific design functional characteristics are used for ADCs commonly used in, and specific to just audio converters? What type of sequential encoding methodology are these specific ADCs based on? Or do they even use encoding in these?
Oh, your wit astounds me....you're able to write all these big words and patronise me, flaunting your 20 years gone experience in a non-music related field, yet you were the one proved wrong after telling me conversion had no latency...you'll have to forgive me if I don't take your design credentials seriously.

I have never claimed to be a converter designer, but I have used, tested and experimented a lot in the course of my AUDIO work - I know the effects of converter latency and why it should be a consideration! you did yet missed a significant factor involved in their design. Doesn't look good does it?

Throw all the words you like at me, try to baffle me with the jargon, but the fact remains that you stated something that was completely incorrect, then patronised those of us who got it right, and now you're trying to score points on the pedantic point that word clock may or may not affect latency at the sub-sample level, whilst simultaneously trying to sweep under the carpet that you weren't aware of the multi-sample delay of your average converter.
Old 9th January 2013
  #122
Just remember fellas, this thread's topic is about a song, after all
Old 10th January 2013
  #123
Gear Addict
 
Mo Facta's Avatar
Yeah, this thread has derailed a little bit.

In any case, Doc, you hit it on the head earlier with your three points. Multi-mic sources across brands is clearly now a no no.

Cheers
Old 10th January 2013
  #124
Quote:
Originally Posted by Doc Mixwell View Post
I am a huge fan of running the IBP hardware. For me, It works better to get these things worked out in the analog realm, while I am tracking. I find you just cannot do the same techniques with the software. Because you cannot record with them. I cannot split an instrument to both an Amplifier and a Recorder with them. But they are fine for playing around inside the box. The hardware works so much better for me, because I would rather take the time on the front end of the recording, than mess with it on the back end. That's just me though, I am sure we can all agree each tool has its own place and "time"..

No doubt... At least we hope so!



Too funny and honestly a little sad at the same time. If every bad mutha$%$$#% mentored someone, we'd have a lot more bad mutha%$%^$## engineers out there, passing it down.


Ahh, in a perfect world this would be the norm.



Right now I am running the Symphony I/O on Internal Clock. I also have a Rosetta 200, which I clock via the Symphony I/O. This is the main ADA I use at my studio. Though we also have a DM2000, an array of MOTU stuff, and a Big Ben clocking all of that. I pretty much for the MOST part use my Apogee gear. I feel the Symphony Clock output, [with a real short WIDE EYE cable] makes a difference to the Rosetta, after listening back and forth, so that's how its rolling.

I have posted most all of my experiences hearing clocking artifacts on gearslutz, and I would say that it has got to the point, where it feels absolutely nauseating to continue. There is so much nonsense out there these days, and as a Sales Rep for a company that Sells High End Converter, and Clocking Products of varying degrees, I only recommend what I believe to be correct for the applications of the end users.

This means, that if someone asks me, if they should purchase a Master Clock, for only Sonic Reasons, I simply reply that is waste of money and an incorrect use of the tool. Whatever artifacts you can identify are subjective. Meaning that not everyone will hear them, nor understand what they are hearing. But if someone has a legit reason to employ a master clock, you simply have to evaluate what it does to your system.

Any change may or may not be present. There is too much variance to explain one single correct answer, unless you've had some experience clocking these things together and listening to them [which is kind of a waste of time IMO] I believe I am super sensitive when it comes to things of this nature, and I hear what is happening just fine. But it's reeeeeeeaaaaaallllllllyyyyyyy small amount of spectral change/distortion.

Me, personally? I think comparing clocks, and esoteric clocking stuff, is a bit like comparing dithers on a mix. Boring as Sin and a waste of time for me. If it takes me longer than 10 minutes, to figure this stuff, there is a serious problem. I've got other fish to fry up!
Lol, can you imagine adding listening for clocking effects to all the other things that you are listening for?

I remember doing many wire listening tests over the years and some formulations were very audible, others kinda subtle.

The main reason for my clock, is I have a bunch of digital goodies and want to make sure they are all in step.

I was able to hear a tiny tightening of the bottom end and point source accuracy, but yeah, very subtle.

My guess is, the clocking in the Avid Omni is quite good already.

Wouldn't mind a pair of B&W 801 to effect my audio path though.
Old 11th January 2013
  #125
Gear Addict
 
Mo Facta's Avatar
Ok guys, so after a simple experiment the results are conclusive.

The Apogee AD8000 delivers a simultaneous source 43 samples later than the Aurora 8. This equates to about .975 ms at 44.1. The workaround, of course, is to use software monitoring at a low buffer rate and use sample delays on the tracks being recorded on the Lynx. This can also be done on input. Once the material has been tracked, it's simple matter of nudging all "Lynx tracks" 43 samples. This means that whatever buffer rate I select will be offset by a further .975 ms, which is tolerable for 64 or 128 but nothing much more.

After implementing the workaround, the resulting audio was in phase and sounded normal again.

The Lynx and the Apogee are both clocked from the AES16e via the AES/EBU stream.

Cheers
Old 11th January 2013
  #126
Gear Addict
 

Quote:
Originally Posted by psycho_monkey View Post
Oh, your wit astounds me....you're able to write all these big words and patronise me, flaunting your 20 years gone experience in a non-music related field, yet you were the one proved wrong after telling me conversion had no latency...you'll have to forgive me if I don't take your design credentials seriously.

I have never claimed to be a converter designer, but I have used, tested and experimented a lot in the course of my AUDIO work - I know the effects of converter latency and why it should be a consideration! you did yet missed a significant factor involved in their design. Doesn't look good does it?

Throw all the words you like at me, try to baffle me with the jargon, but the fact remains that you stated something that was completely incorrect, then patronised those of us who got it right, and now you're trying to score points on the pedantic point that word clock may or may not affect latency at the sub-sample level, whilst simultaneously trying to sweep under the carpet that you weren't aware of the multi-sample delay of your average converter.

I don't have 20 years experience. I had an intern job 20 years ago for 5 months while I was in school, it was a 1 semester cop-op thing. I ended up changed majors, I never did electronics. When I worked on semiconductors I was actually a Chem Eng student, I did cv deposition, I switched majors and haven't done any of that since.

Regardless, I originally stated there was no noticeable latency between
AUDIO converters I have used. I also stated the only latency I knew of with ADC was while encoding and due to clock propagation delay of gates. You claimed "Latency and clocking are not related" with A/D converters, and that's incorrect.

page 19

http://www.ti.com/lit/an/sbaa147b/sbaa147b.pdf


"Figure 14B shows a timing diagram"
Old 12th January 2013
  #127
Quote:
Originally Posted by emitsweet View Post
I don't have 20 years experience. I had an intern job 20 years ago for 5 months while I was in school, it was a 1 semester cop-op thing. I ended up changed majors, I never did electronics. When I worked on semiconductors I was actually a Chem Eng student, I did cv deposition, I switched majors and haven't done any of that since.
So one could possibly say that "I designed AD converters" as a means of support for your opinion was stretching the truth somewhat? I designed loudspeakers whilst at university as part of my course...I wouldn't call myself a loudspeaker designer, I've really only retained the most basic knowledge there. Certainly haven't done anything with it in 10 years.

Quote:
Originally Posted by emitsweet View Post
Regardless, I originally stated there was no noticeable latency between
AUDIO converters I have used.
That might have been what you MEANT, but you missed out the word "noticeable". You've been arguing "no latency" for a couple of pages.

There certainly IS noticeable latency if you parallel the direct source with the round-trip through the converter - you'll hear it as phasing. You've been stating that converters have NO latency from the start - it's only since it's been pointed out that you're wrong have you started adding in the "noticeable". And as I've said from the start, in most cases I agree. But not all - which is why we need to be aware of it, because it COULD cause problems in some circumstances.

Quote:
Originally Posted by emitsweet View Post
I also stated the only latency I knew of with ADC was while encoding and due to clock propagation delay of gates. You claimed "Latency and clocking are not related" with A/D converters, and that's incorrect.

page 19

http://www.ti.com/lit/an/sbaa147b/sbaa147b.pdf


"Figure 14B shows a timing diagram"
This diagram, as far as I understand it, actually supports what I'VE been saying. I've never said that the delay isn't induced during the clocking phase - I've said that changing the clock source from internal to external won't affect the latency, which is what you've stated a few times now. I suppose it MIGHT vary it at the subsample level (I suppose it has to to put 2 converters in sync) but that really is super insignificant even compared to the converter latency.

I don't know if this is a language barrier thing, but if that's what you've been saying all along, then it hasn't been clear.

As an aside, you might like to read the following paragraph on page 19 - I'll quote it here for you to save you hassle:

"For this reason, we measure latency for a delta-sigma A/D
converter by starting at the beginning of a sample period, and measuring to the time that data can be retrieved. It may also be practical to include in the latency time the time needed to retrieve the data, since delta-sigma A/D converters nearly always have serial interfaces. For audio converters, this additional latency can be very significant, even up to several tens of sample periods. For low-speed industrial converters with sinc filters, it sometimes amounts to only a few modulator cycles. For delta-sigma A/D converters, filters with constant group delay are almost always used, so there is no difference between group delay and latency. The latency-time of a delta-sigma converter is often called Settling time."

Surely point proved now - from the horses mouth? Even your "low speed industrial converters" have latency.....
Old 12th January 2013
  #128
Gear Addict
 
Mo Facta's Avatar
Well, for the record, when I monitored a simultaneous source (a mic signal split via auxiliary sends) through both the Apogee and the Lynx (at the same time), the phase issue was massively noticeable. It certainly wasn't negligible. We're talking about almost a millisecond between the two converters. Some will be less, some will be more I imagine, but there's nothing unnoticeable about it.

Cheers
Old 13th January 2013
  #129
right, its way noticeable for me. I guess ADA with similar latency won't be as significant right away but wait until you start processing the tracks

I certainly see how you can nudge the tracks, but you gotta monitor through with delayed tracks?!

To me messing with phase is all just different comb filtering, in some way entering a myriad of different outcomes. I work real hard to make the bleed sound good with my mics

But man all this is giving me a headache! I just want to get a good drum sound. seriously!!! that's a lot more work then I want.
Old 13th January 2013
  #130
Gear Addict
 
Mo Facta's Avatar
Well the point of delaying the Lynx while tracking is to time align the two converters, thus eliminating potential phase problems so I can get on with worrying about mic placement. It's not that difficult to do and will only add .975 ms of additional throughput latency to the system, overall. That's why I figure if I use 64 or 128 samples per buffer total system latency will still be well under 5ms, total.

It's a crappy situation, I know, but there will be times where I'll need more than eight simultaneous channels in a multi-mic setup and will need to use the Apogee.

Until I can afford an Aurora 16, I'll have to make do.

Cheers
Old 7th September 2013
  #131
Lives for gear
 

what have i done.

I should be more precise...

What i ment was for example: tracking layers/overdubs of el gtrs with UA 2192 BUT tracking vocals with Orpheus AD. Two totally different worlds/converters.

I find all converters has it own sonic signature which has it weakness and strength... a matter of taste.

again, will it sit or glue in the mix using several AD tracking (OVERDUB) and NOT at the same time?

Old 7th September 2013
  #132
Quote:
Originally Posted by Mo Facta View Post
Ok guys, so after a simple experiment the results are conclusive.

The Apogee AD8000 delivers a simultaneous source 43 samples later than the Aurora 8. This equates to about .975 ms at 44.1. The workaround, of course, is to use software monitoring at a low buffer rate and use sample delays on the tracks being recorded on the Lynx. This can also be done on input. Once the material has been tracked, it's simple matter of nudging all "Lynx tracks" 43 samples. This means that whatever buffer rate I select will be offset by a further .975 ms, which is tolerable for 64 or 128 but nothing much more.

After implementing the workaround, the resulting audio was in phase and sounded normal again.

The Lynx and the Apogee are both clocked from the AES16e via the AES/EBU stream.

Cheers
I must have dipped into my stash when I posted the above "in step" because I could add a PCM42 @ 900ms in front of a signal and it would be waaaay out of "step". I meant jitter free (or as jitter free as I can get)

Mo,
How did you check your latency, what was your input test signal?

Was thinking of using a click made from a single truncated square wave cycle.

Would this work?

Thanks,

John


BTW, I did some cleaning up my listening space acoustically and forgot to turn the clock on the other day...

This is monitoring through an Avid Omni clocked with an Isochrone OCX (PT 10 HD Native)

Thought the mix sounded a bit flabby and the image was not so great.

Saw the clock was off, so I saved and restarted with the clock on and the difference with and without external clocking is actually not so subtle at all.

The bottom end is way tighter (almost sounds like when you swap amps and get massively better damping) and the image is more precise, deeper, taller and the mix sits more obviously in front of the monitors.

It kind of freaked me out a little!

Any one else hear a lot of jitter reduction on the Avids with external clocking?

Is my power dirtier than I think it is?
Old 8th September 2013
  #133
Lives for gear
 

Quote:
Originally Posted by Simma Lugnt View Post
what have i done.

I should be more precise...

What i ment was for example: tracking layers/overdubs of el gtrs with UA 2192 BUT tracking vocals with Orpheus AD. Two totally different worlds/converters.

I find all converters has it own sonic signature which has it weakness and strength... a matter of taste.

again, will it sit or glue in the mix using several AD tracking (OVERDUB) and NOT at the same time?

Old 24th April 2015
  #134
Lives for gear
ha what an ending. So -- is there a document somewhere with the latency differences between converters? I'm particularly interested in the differences between lynx aurora and apogee 16x. I need to insert delay compensation plugins between these 2 for parallel processing before summing. OR pony up more $$ grands to unify converter brands. Plugin compensation seems a lot cheaper
Old 25th April 2015
  #135
Quote:
Originally Posted by goldi View Post
ha what an ending. So -- is there a document somewhere with the latency differences between converters? I'm particularly interested in the differences between lynx aurora and apogee 16x. I need to insert delay compensation plugins between these 2 for parallel processing before summing. OR pony up more $$ grands to unify converter brands. Plugin compensation seems a lot cheaper
Measure it yourself...send a pulse out and in again on each converter, and measure the difference.

Really though, it's going to be a few samples difference. Provided you're not putting things in parallel between converters, you'll be fine.
Old 25th April 2015
  #136
Lives for gear
Quote:
Originally Posted by psycho_monkey View Post
Measure it yourself...send a pulse out and in again on each converter, and measure the difference.

Really though, it's going to be a few samples difference. Provided you're not putting things in parallel between converters, you'll be fine.
you mean by recording a ping thru each? i did a quick ear test with a snare, seems my apogee is around 1.7 ms faster than the aurora. I guess it might even change per setup, but i thought there'd be some document around here with stated processing latencies of various converter brands.
Old 17th August 2015
  #137
Lives for gear
 

Quote:
Originally Posted by goldi View Post
you mean by recording a ping thru each? i did a quick ear test with a snare, seems my apogee is around 1.7 ms faster than the aurora. I guess it might even change per setup, but i thought there'd be some document around here with stated processing latencies of various converter brands.
Interesting...

It seams that we need a new thread for: Apogee vs Aurora on drums which seams almost like a preamp game API vs Neve on snare and kick...
Old 30th December 2017
  #138
TNM
Lives for gear
i've been looking into this topic with interest, as all my synths go into different stuff.. 8 apollo ins, 8 focusrite adat, 8 audient adat, and an apollo twin for a further 2 inputs.

I was thinking about measuring the subtle differences so everything is time aligned perfectly by nudging after recording.. then it hit me..

every single synth i use has different converter latency too.. So even if i had, say, 2 apollo 16's for 32 identical ins, there'd still be timing variations from the synths themselves, not to mention sways in midi timing.

I think it's only important when say you were recording a multi mic'd drum kit and were doing it from two totally different converters.. then there might be phasing issue.. so you would want to use the one converter for that example..
I think we can drive ourselves insane with this stuff..

it's funny, 17 years ago, when i had a pro tools TDM Mix cube rig and 64 ins coming from literally 5 different sources LOL, i never knew about this stuff and never had an issue. All sounded great. Now the paranoia started, from reading up on topics like this.

Yes, i suppose one could even measure the delays from the stuff feeding into the converters (in the case that they are not real instruments and have their own d/a converter going into the interfaces A/D to be recorded), on top of the differences between the a/d converters themselves, and get everything sample accurate.. But i'm not gonna!
Old 31st December 2017
  #139
Lives for gear
 
JayTee4303's Avatar
With something like 12 interfaces and 12 PCs here, room to room routing runs into these kinds of issues on a frequent basis.

A couple of things I do to minimize problems...

One instrument follows one chain. No matter how many tracks we grab at once from one source, we don't run them thru different converters. Think "bleed" here too.

Money signals get money converters, by facility design. I usually use published dynamic range as my decision making authority. Pay attention to spec details, weighting, etc.

ADAT hops cost one sample per hop. Drums into an Octapre, thence via fiber to the Live Room host, one extra sample. LR host thru the backbone (Z-System's Optipatch Plus) to the Control Room PC, another extra sample delay.

Measuring this sort of thing, as opposed to guessing, helps in the pre-production decision making process.

The ADAT hops aren't enough of an issue to affect where we tap signals for foldback..artist monitoring. Therefore, engineer's convenience rules the day. I like to turn them around AT the recording PC interface. That's where I am, easy to re-route once you have a signal or group of signals recorded, and artist monitors now employ the recorded, as opposed to live, signal.

Don't rub the artist's face in unnecessary discussion about necessary tech. Nobody's EVER mentioned two extra samples lag on drums, but they might if I publicised the info during a session.

I HAVE gotten called on 5-7 ms RT lag on softsynths. Maybe they actually felt it, maybe they were showing off, but THEY brought it up.

Nudge as needed in post. Movie style clapper helps, but I just clap my hands during countin, after "rolling audio", and train them to give me a four count or more between there and when they start playing.

Lastly, like I said earlier...we don't choose converters artistically during sessions, but the facility IS designed to have higher dynamic range converters where they need to be on a permanent basis.
Topic:
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
mosrite / Remote Possibilities in Acoustic Music and Location Recording
101
SoniqKwality / So Much Gear, So Little Time
1
willi1203 / Remote Possibilities in Acoustic Music and Location Recording
6
mos / So Much Gear, So Little Time
1
FROGGY / Remote Possibilities in Acoustic Music and Location Recording
18

Forum Jump
Forum Jump