1. Register NOW and become part of this fantastic knowledge base forum! This message will go away once you have registered.

Delay Compensation for AD8000

Discussion in 'Pro Audio Equipment' started by Steve Hudson, Jan 15, 2002.

  1. Steve Hudson

    Steve Hudson Active Member

    I noticed in a recent thread that our erstwhile co-moderator Julien said he'd caluclated the delay compensation factor for his Apogee AD8000SE's at 19 samples. Just last week I asked Apogee what they recommend, since there's nothing in the manual; their tech support says 67 samples from "analog input to digital output" and 114 samples "from analog input to analog output."

    Am I interpreting this information correctly by assuming I nudge or Time Adjust any track that has an analog insert (like an outboard compressor or reverb) by 114 samples earlier? In what case would the 67 samples apply? I am not using any digital I/O to/from my outboard racks.

    Thanks in advance for any clarifications.
     
  2. anonymous

    anonymous Guests

    Best thing is to record the results and then zoom in and measure the total of the audio's 'round trip' out and back in to the DAW

    BTW this applies for ALL multichannel i/o converter interfaces (motu 2408 etc)

    send to an out patch into , say a compressor then back into the interface and record it...

    measure how much you have to 'nudge' it back to be perfectly in time with the origional...

    knowing the amount that is needed to 'line it up' allows you to advance the source track BEFORE sending out so you can monitor in record input , while tweaking the outboard EQ / compressor / whatever.....

    :)

    My calculations may have been wrong!

    :)
     
  3. Steve Hudson

    Steve Hudson Active Member

    Julian -

    Thanks for the suggestion. I'm anxious to see how close the delay is to what Apogee says it should be, and will report back with results.
     
  4. mwagener

    mwagener Active Member

    While we're on the issue of A/D - D/A delay: how much does frequency come into play. Is it the same delay for, say 100 Hz as for 5,000 Hz? I recently put a 1K tone through the digital inputs of my board and the same tone through the analog inputs. I looked at the output of the board on a 2 beam oscilloscope to match the waves exactly. There was a difference of 47 samples (analog in late compared to dig). When I switched to a 100 Hz tone the difference was 67.

    Any ideas?
     
  5. anonymous

    anonymous Guests

    Hello Michael! (I owe you a Twist CD, on it's way soon)

    It was suggested I use a 'click' to line stuff up.

    I DONT LIKE YOUR TONE!!!!!

    :D

    :w:
     
  6. mwagener

    mwagener Active Member

    Originally posted by Julian Standen:
    I DONT LIKE YOUR TONE!!!!!

    :D

    :D :D :D :eek:
     
  7. Steve Hudson

    Steve Hudson Active Member

    Julian/Michael, et al -

    I conducted the following test to determine the delay induced by running a signal thru my AD8000SE:

    1) Record a click track into PT.

    2) Bus audio from the click track thru the AD8000 to a second track and record.

    3) Viewing both tracks at sample level, determine the number of samples the second track is delayed from the first track. Result was 119 samples.

    4) For grins, I repeated Steps 2 & 3 after inserting an 1176 to see if there was any additional delay induced by adding effects to the chain (as I suspected, there wasn't).

    5) I ran each track thru a separate channel on my board then inverted the phase on one; I expected the signal to go away but was reduced from the original signal by about 30db. Given the typical dynamic range of stuff we work on (nonclassical), should this be acceptable?
     
  8. anonymous

    anonymous Guests

    Ooops! Perhaps it WAS 119 rather than 19 Oh dear! I'm sorry! Shamfully misleading I do apologise.

    :)
     
  9. mwagener

    mwagener Active Member

    Originally posted by Hud Hudson:

    2) Bus audio from the click track thru the AD8000 to a second track and record.


    How did you get the click track out of PT into the anlog in of the AD8000? If you went out analog then you would introduce a delay through the PT D/A converter wich you would have to deduct from the 119 samples in order to find out the *actual* AD8000 delay. Just to add to the confusion :p I send a 1 k tone (sorry Jules) from an external analog source through the AD8000 analog in, then digital out (AES) into the Sony DMX R-100 and multed the same tone into the Sony direct via it's onboard A/D converters: the AD8000 was 14 samples faster than the Sony. The Sony had a delay of 42 samples and the Apogee a dealy of 28 samples from A to D. 119 samples (about 2.5 ms at 48kHz) seems a lot to me. Are you running the test at 44.1 or 48 kHz?


    5) I ran each track thru a separate channel on my board then inverted the phase on one; I expected the signal to go away but was reduced from the original signal by about 30db. Given the typical dynamic range of stuff we work on (nonclassical), should this be acceptable?


    Was the -30dB signal full range, or was just the low or high end audible? If it was not full range, then you are probably hearing the difference between the two converters or the difference between analog and digital.
     
  10. Steve Hudson

    Steve Hudson Active Member

    Michael -

    I am recording in PT at 44.1. After recording the original click track ("Track 1"), I used an internal bus in PT to record a second track ("Track 2"). The internal delay within PT was 10 samples. Then I patched directly from an AD8000 analog output (via patch bay) to an AD8000 input feeding into another track ("Track 3") in PT. The delay induced by this analog chain was 119 samples. Since in this second instance (Track 1 to Track 3) I did not do any busing in PT, would PT still have induced the 10 samples delay I experienced when duplicating Track 1 via internal (digital) bus?

    On the issue of the phase cancellation test, the audible click artifact remaining after I switched phase on one channel seems to be full range.

    I'd like to be able to do the frequency dependent tests you conducted, Michel. Can you suggest an inexpensive tone generator?

    Thanks again for all the help. This is a great forum.
     
  11. I'd like to be able to do the frequency dependent tests you conducted, Michel. Can you suggest an inexpensive tone generator?

    You can use the digirack signal generator plug for this.

    Sincerely,
     
  12. Steve Hudson

    Steve Hudson Active Member

    Thanks, Casey! So many plugs, so little time...
     
  13. Ang1970

    Ang1970 Well-Known Member

    Originally posted by Hud Hudson:
    Since in this second instance (Track 1 to Track 3) I did not do any busing in PT, would PT still have induced the 10 samples delay I experienced when duplicating Track 1 via internal (digital) bus?

    This would seem to be a logical assumption, yes. But it's kind of a moot point, since you can't get around either the ~10 sample latency induced by the PT busing or the ~109 sample latency induced by the AD8k. 4+2 or half a dozen, take your pick.
     

Share This Page