Time shift .. what am I missing

[JLepore]

ok ... still doing my eval on SAW here so I decided to hook up and record a practice session. Setup is as follows:

SAW->SAC linked. All my inputs are coming in through SAC.
ch 1-7 drums, ch 8 CD player. Output 1 -> my ears
Get everything dialed in, sounds good enough. Do the link, get the first 8 tracks armed and record playing through a song.

Everything in my ears sounded fine (running 2x64 buffers) - no real latency.

Go back and play back the tracks (through SAC as well) and there is a SIGNIFICANT time difference between the drum tracks and the CD track. Every beat is a significant flam - and no way I am off by THAT much time consistantly!

What am I missing here? We're not talking a few milliseconds here - we're talking it sounds like I'm flamming - 100ms + easily. All thr drum tracks sound fine time wise in relation to each other (lots of bleed, so it would be noticible with the snare bleeding into everything)

[Bob L]

What about playing the tracks together in SAW... listening there or routing just the complete SAW mix to output 1 and then feed that back to a blank SAC chan with the source assigned to SS Out 1... is that in time... in other words did the tracks record in time properly along with the CD track?

If so... which I imagine they did... now let's look at how you changed the rouing to play them back as individual tracks in SAC along with the recorded CD track... what is different...

How is the CD assigned in SAC... thru an spdif input perhaps... or straight analog in... if digital... could there be a sync issue on that input that does not properly receive its clock source correctly when looping back from SAW?

Bob L

[JLepore]

The CD is analog in ... coming through an ADA8000 channel, as are the drums.

I initially tried the playback by assigning the SAW Out 1 to a dead channel, and I didn't notice it, but the track playback level was pretty low. I then changed all the inputs in SAC to the SAW 1-8 playbacks so that I could tweak my EQs, etc. I spent some time doing this listening only to the drum tracks.

Once I unmuted the CD track, I noticed it immediately. I will go back an listen to it directly through SAW in the morning (got everything shutdown at the moment).

I did notice there were no dropped buffers during the entire session (although at about 10% load, I wouldn't really expect any).

[JLepore]

Ok .. went back and reloaded the session(s). The shift is the same if I'm playing it back in SAW or through SAC. If I insert a 23.61ms delay in the CD track, everything lines back up (to the human error point anyway).

Checked by buffer settings on both SAC and SAW were set to 2x64. So now I'm wondering if it's the latency of the system doing it .. the CD has to make it all the way through the system to get back to my ears for me to play the note that then has to go into the system to get recorded .... still don't think this adds up to 23+ ms though. This would certainly be a problem for anyone playing to a click/tracks that were coming through the system for monitoring then.

I can try monitoring directly from the CD without going through the console first (which puts me right back where I started anyway) and see if that makes the difference.

[Bob L]

There certainly is not this kind of latency issue... there are thousands of recording sessions and overdub sessions that would be rendered useless every day if this was the way it works... so... something else is slipping by us that is not making sense.

I tried to explore different possibilities of how last night but did not come up with anything that could explain why all the drum tracks are in sync and just the CD track is way out... unless... the system was slipping buffers... it could be that the system is not holding stable at the 2x64 settings... try 2 x 128... even with that latency, you should easily still be able to perform within reasonable time... then see if the results are the same... perhaps the CD track slipped one or two buffers at the start and was then out by that much after that.

Bob L

[JLepore]

ok .. will try it out again tonight when I get back.

I'm assuming that it has nothing to do with using the demo version of SAWLite to do this (kinda the point of the demo I would guess!)

[Bob L]

Demo should be fine... I have not heard any feedback of this problem from others... so... try a new recording and see... it could have just been a glitch brought on by the machine not handling the low settings...

Also... I am assuming you are using ASIO drivers... not MME.

Bob L

[Bob L]

Just thought of something... is there a span across multiple soundcards between the drum inpit devices and the CD input... if they are on two different soundcards... could be a clock issue where the cards are not properly sync locked.

Bob L

[JLepore]

Just tried it quick with the same results, but have to run out for a meeting.

Same machine I have been using for my SAC work:
i5-650 / 4G @1600Mhz / SSD / Motu 424 / 2048 MKII interface / 1 profire 2626 (in disconnected mode - just acting as a preamp/output) / 2 ADA8000 - all connected to the same 2048 for clocking.

In this case, I'm actually just using the ADA's for everything (keep the cables on the same side of the rack). They are clocked ADAT from the 2048. ASIO drivers

I can try moving the physical inputs around so that there is a mix of drums and CD on each ADA8000 - right now drums happen to be on the first 8000, and the CD on the second.

[Bob L]

Here's an idea... try switching two drum inputs with the CD inputs... record a new test and then see what happens... when you playback now... is the CD in sync with the rest of the drums and are the two drum inputs from the second unit now out of sync?

Bob L