The No.1 Website for Pro Audio
Alright, so who DOES it?
Old 24th August 2002
  #31
There is only one
 
alphajerk's Avatar
 
🎧 15 years
1'=1ms is my quick rule too although isnt it really 1'=1.13ms? so at 10' thats an extra 1.3ms.

the extra 2.7 came from the example above [system and convertor latency]
Old 24th August 2002
  #32
Capitol Studios Paris
 
jon's Avatar
 
🎧 15 years
Yep. I was mentally back to analog recording mode and forgot about the converter delay. Of course as you know, it has to be added only if they're overdubbing onto a DAW, or monitoring from a DAW -- or both, which is double the delay.
Old 24th August 2002
  #33
Capitol Studios Paris
 
jon's Avatar
 
🎧 15 years
If memory serves, sounds travels about 1100 feet per second.

So, sound will travel one foot in (1 second divided by 1100 feet) = 0.9090909 milliseconds.

10 feet = 9.1 ms...if all the above is correct, that is.
Old 24th August 2002
  #34
Lives for gear
 
C.Lambrechts's Avatar
 
🎧 15 years
Quote:
Originally posted by jon
Chris, did your jazz cats do overdubs? Most of the jazz stuff I've seen was done live.
yup, most of the stuff indeed is live, or just several takes to get the good one. But several tracks would have aditional stuff added afterwards ..... like some second voice sax parts or the pianist would add a keyboard solo afterwards.


hmmmm, come to think of it .... most of the groovy stuff was live .... the dubs were more effects then realy groovy stuff .... so might be hard to tell ......


anyhow .... food for thought and testing I guess.
Old 25th August 2002
  #35
Founder
 
Jules's Avatar
turkish math

Arif Mardin was lining up midi back during his Chakka Kahn 'feelfor you' period. Anal precision midi and 'bubbling' counterpoint sounds.... I learned how from his son Joe who I did some co-production with.... pre DAW we used tape and measured it with rulers it was awsomley tedious.....
Old 25th August 2002
  #36
Capitol Studios Paris
 
jon's Avatar
 
🎧 15 years
Yep...the main strength of DAWs is editing when lots of it is needed. If only it wasn't needed so often...

MIDI slop. What fun.
Old 25th August 2002
  #37
There is only one
 
alphajerk's Avatar
 
🎧 15 years
is my math wrong?

1s=1000ms

i thought it was 1130 ft/s

so that would be 1.13ms delay per foot.
Old 25th August 2002
  #38
There is only one
 
alphajerk's Avatar
 
🎧 15 years
midi slop, i kinda dig jamming up the midi pipe with excessive information to get a "chaotic" reslut of timing.
Old 25th August 2002
  #39
Capitol Studios Paris
 
jon's Avatar
 
🎧 15 years
Heh...he said *chaotic re-slut*.

So, if sound travels 1130 f/s, then in 1 ms, sound travels 1130 f/s / 1000 ms/s = 1.13 feet/ms.

So 1 foot = 0.885 ms (1/1130*1000). 10 feet = 8.85ms.
Old 26th August 2002
  #40
Lives for gear
 
Fibes's Avatar
 
🎧 15 years
At sea level, 72 degrees with low humidity... The 1 foot per millisecond is a good rule of thumb but on any given day the weather can wreak havoc on you...
Old 28th August 2002
  #41
Gear Nut
 
Diginerd's Avatar
 
🎧 15 years
Found this site (It's metric so you nice USA boys are going to have to do the conversion!)

http://www.measure.demon.co.uk/Acous...are/speed.html

Some interesting results for the lazy are:-

All results are at sea level, standard pressure

0c (32F) and zero humidity: 331.45 M/s = 1087.43 fps

18c (64.4F)and 75% humidity : 343.02 M/s = 1125.39 fps

25c (77F)and 75% humidity : 347.56 M/s = 1140.29 fps

30c (86F) and 75% humidity : 350.9 M/s = 1151.25 fps

30c (86F) and 100% humidity (wear shorts!): 351.48 M/s = 1153.15 fps

In short temperature has a much larger impact on speed than humidity..

So maybe session notes should contain the temperature of the room where the material was recorded!!! yuktyy
Old 28th August 2002
  #42
There is only one
 
alphajerk's Avatar
 
🎧 15 years
Quote:
Originally posted by Diginerd
So maybe session notes should contain the temperature of the room where the material was recorded!!! yuktyy
i actually think they should include both temp and humidity... especially for analog equipment. gear takes on its "mood" from [im assuming] the enviroment... i can tell rainy days gear plays better to my ears than sunny days. im not sure if its coincidence, but it always sounds better when its raining.
Old 29th August 2002
  #43
Gear Nut
 
Diginerd's Avatar
 
🎧 15 years
Quote:
Originally posted by alphajerk


im not sure if its coincidence, but it always sounds better when its raining.
Hmm, any really good studios in Vancouver? Seems to rain there all the time...? That and it's cheap, which is why the X-Files were made there..
Old 29th August 2002
  #44
Lives for gear
 
loudist's Avatar
 
🎧 15 years
It seems to me...

That if you record a click or take an already recorded track on your daw with transients and send that out to the headphones, super close mike one ear of the headphones, record it back into the daw, then you could measure and compute your latency visually. Yes? No? Too easy?

Discuss.
Old 29th August 2002
  #45
Founder
 
Jules's Avatar
Yarzzzzzz!

I think Loudist has a point.

Good idea..

Old 29th August 2002
  #46
Gear Addict
 
🎧 15 years
Okay I just measured in my Mix Plus. From the ADAT bridge analog outs to the line input on my spider using a click track.

I got 86 samples delay at 44.1k.....
Old 30th August 2002
  #47
Founder
 
Jules's Avatar
Care to do a click to headphone to mic to PT round trip measurement as well?


Old 30th August 2002
  #48
Lives for gear
 
loudist's Avatar
 
🎧 15 years
It would be interesting to see what the 'headphone round trip' latency is for all the PT systems owned by the members of this forum, for comparative studies.

Jules... how about leading this off
Old 1st September 2002
  #49
Capitol Studios Paris
 
jon's Avatar
 
🎧 15 years
You guys have e-cue's numbers, kevin's and mine. Anybody else look into this?

Maybe it's a boring issue, but definitely not a NON-issue.

Jon
Old 1st September 2002
  #50
Founder
 
Jules's Avatar
hmmm I have several variables that could make for slightly different 'round trips' depending on the configuration I am using. (Prism Dream ADA converters or Cranesong ones, with often a Digidesign 888 or 888/24 in use for Foldback duties)

My set up is not so straightforward - I do a lot of I/O changing around on my rig between tracking and overdubbing.

Tracking = 16 Prism in's - 888 or 888/24 for foldback
Overdubbing = 8 Prism ins + 2 x ch Cranesong Hedd - Prism for foldback
Mixing (not in question here but anyway) = 16 Prism outs

Either way I will see if we can get round to it early this week.

Old 1st September 2002
  #51
Lives for gear
 
loudist's Avatar
 
🎧 15 years
Right Jules, the different configurations latency are what could be most interesting.

Jon, the headphone round trip scenario (palyback of a hi hat from the DAW to the headphones miked and recorded, then compared time wise) seems (to me) to be the real working assessment of the latency.

It might be interesting to see what the 'round trip' latency is from the different users here who have the same systems. As well as the ones who have expansion cards and different computers... ect.

This seems a worthwhile project.
Old 1st September 2002
  #52
There is only one
 
alphajerk's Avatar
 
🎧 15 years
why even bother with the micing the headphone? how about just plugging from the out to the in and measuring. that will give system delay. the mic will give added nature delay.
Old 1st September 2002
  #53
Capitol Studios Paris
 
jon's Avatar
 
🎧 15 years
The analog signal path time will be in the pico or nanosecond range i.e. nothing you can hear. The D/A and A/D processes are the latency pigs.

My numbers were for PT with AD8000s, out and back in.

I have the number of samples of delay between the repro and sync heads on a Studer A820 too, if anyone's interested. Always handy to have if someone forgets to put the Studer in sync when making slaves or transfers to PT.

Old 1st September 2002
  #54
Founder
 
Jules's Avatar
"why even bother with the micing the headphone? how about just plugging from the out to the in and measuring. that will give system delay. the mic will give added nature delay."

It was a 'vibe" thing to FULLY replicate the path from DAW to 'musicians ear'. Perhaps pathetic technically but meant in good spirit.

I had better add in the chain Mic / pre / Distressor / Fatso!

I wonder how many picco seconds are in a sample?
Old 2nd September 2002
  #55
Lives for gear
 
loudist's Avatar
 
🎧 15 years
Jon, doesn't it matter how one builds a cue mix in PT in terms of harware overhead?
Old 2nd September 2002
  #56
Capitol Studios Paris
 
jon's Avatar
 
🎧 15 years
Loudist,

Not sure I understand your question. I make the cue mixes on an analog board...no latency there. The distance from the mic is usually well under a foot, so no problem there, either. If distant micing, then 1 foot equals roughly the sample rate divided by 1000, in DAW samples.

What matters is the D/A plus A/D times.

If you are listening to analog tape and recording to a digital machine, you have to nudge the overdub earlier by the amount of A/D latency.

If you are listening to a digital machine and recording to one as well, you have to nudge the overdub earlier by the amount of machine 1's A/D + machine 2's D/A.

The most standard situation would be listening to your DAW and overdubbing onto it. In this case, you measure the D/A+D/A time and nudge the overdub earlier by that amount.

Thing is, no one seems to ever do it.
Old 2nd September 2002
  #57
Founder
 
Jules's Avatar
While this is under investigation... a DAW tip

one DAW timing shift that can matter a lot is one with Auto Tuned vocals.

The more the tuner works the higher the delay factor.

Thus an tuned vocal performance can end up with slightly different timing shifts within it's self. A bummer..

One easy way around this is to use the Voc Align plug in to align the tuned vocals back to where they landed originally.

It beats the 'guesstimate' global nudge (mine is 60 samples forward) method and the zoom / slice / nudge method for speed.

Old 2nd September 2002
  #58
Lives for gear
 
loudist's Avatar
 
🎧 15 years
Jon, If you are using the analogue mixer to build your cue mix, are you 'stemming' out from your 16 or 24 D/A's to the big frame board?

I understand that the latency is in the A/D and D/A conversion, but I am wondering since you stem out, if that has an impact on the latency, compared to using the 2 bus in PT or a set of aux sends in PT... like some others might have to do because they don't have an analogue mixer.

In theory. it shouldn't, but we have all experienced some pretty illogical things concerning digital audio.

Why I was suggesting the headphone round trip is that if one is using aux sends for cue or even the 2 bus that its load and overhead is in the loop and therefore its possible effect on latency. It shouldn't have an effect, but then again...
Old 2nd September 2002
  #59
Capitol Studios Paris
 
jon's Avatar
 
🎧 15 years
Loudist,

Yes, I generally send all the channels out individually. But summing them first in PT shouldn't make any difference worth bothering about.

If I recall, PT aux sends add a sample or two...but could be wrong there.

In comparison, the converter round trip is in the range of 122-128 samples on my system.

Jules,

Cool tip!
Old 2nd September 2002
  #60
Lives for gear
 
loudist's Avatar
 
🎧 15 years
Jon,

It seems that we are at odds.
I am curious from a scientific methodology standpoint (eliminate or confirm theory or supposition by tests).
While all your numbers and calculations are logically sound, they aren't proven. Not that I suspect that they are wrong, just not comfirmed, thats all.

I have heard varying degrees of latency on differing PT systems, while not having the opportunity to sus out the culprit(s) as sessions were in progress.
Having an direct patch out/in round trip of the converters is definitly valuable data, a baseline to be compared with a mix out with plugs 'mic round trip', compared with the same mix without plugs 'mic round trip'. If there is no significent difference in latency then logic is in concert with the hardware. If there is a significent difference in latency, then the software/hardware doesn't follow in a logical manner.
Neither has been comfirmed.

And just to throw in another variable... sampling rate.
Does this change the latency in a predictable way?
Is 96K twice or half the latency of 48k?
Or is it not a multiple/division of 2?
πŸ“ Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 56 views: 13314
Avatar for burns46824
burns46824 6th June 2019
replies: 189 views: 22369
Avatar for MarsBot
MarsBot 29th August 2017
replies: 944 views: 111712
Avatar for mistermorrison
mistermorrison 3 weeks ago
replies: 1195 views: 158700
Avatar for origin35
origin35 3 weeks ago
Post Reply

Welcome to the Gearspace 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…

Forum Jump
Forum Jump