What do you make out of this? (VDAT vs ASIO(Tracktion 2) rec

voidar · Post by **voidar** » Sun Jul 10, 2005 7:00 am

I did a little test today and I am not sure what to make out of this.

I had Tracktion 2 MTC-slaved to VCR-128 and simultaniously recorded the input of my guitar through my A16U to both Tracktion 2 and VDAT.

The A16U is slaved to my Pulsar II via wordclock.

I then did some manual editing in WaveLab to have the tracks line up, and this is what I came up with.

http://www.home.no/voidar/mp3/vdatvst2.mp3

I have the VDAT recording on the left and the T2 (ASIO) recording on the right.

Both tracks line up perfectly for the first second, but then the right tracks seems to drift periodically every couple of seconds?

I recorded this in 24 bit 44.1KHz and used ASIO2-32 drivers for T2. I even had the VCR-128 clk hooked up to the ASIO2 dest clk input.

Theories?

Is this a user error by me or simply an error in T2, or perhaps this is representative for most ASIO hosts?
Is this why people prefere VDAT for recording audio?

[ This Message was edited by: voidar on 2005-07-10 09:14 ]

voidar · Post by **voidar** » Sun Jul 10, 2005 2:24 pm

I must say VDAT is great. I just recorded a track using nothing but SFP. Kind of forces you to perform better, though overdubs are quite easy to do once you get the hang of it

.

Old fashioned editing.

[ This Message was edited by: voidar on 2005-07-10 15:24 ]

Music Manic · Post by **Music Manic** » Sun Jul 10, 2005 2:34 pm

Yep I said it long ago.Vdat is special.As much as Symbiote said about O's and 1's,there's something going on with Sequencers algos.

I feel Vdat is Raw and doesn't mess with operation of OS much other than streaming it to hard disk.

I think these sequencers also have code to check on timing of plugins,MIDI,latency etc,and this must interfere with performance surley.

emzee · Post by **emzee** » Sun Jul 10, 2005 4:40 pm

Interesting........ thanks for sharing that one.

symbiote · Post by **symbiote** » Mon Jul 11, 2005 6:56 am

Hey, I never said sequencers and protocols aren't ever broken =P. I wouldn't be too surprised if ASIO was a bit broken, being designed by Steinberg and all, or Tracktion having broken recording, being low-priced and fairly new.

I fully expect VDAT to be as good or better than anything native. I was just arguing that, under normal conditions, the recordings should be the same.

voidar · Post by **voidar** » Mon Jul 11, 2005 7:51 am

The levels and quality of sound seems to be equal until the jitter, as they seem to superficially phase out before that.

I encourage people to try this with other ASIO sequencers aswell.

I think marcus and someone else had something to say about jitter in native sequencers too?

[ This Message was edited by: voidar on 2005-07-11 08:52 ]

wolf · Post by **wolf** » Mon Jul 11, 2005 9:52 am

When syncing a sequencer to another clock, it has to introduce a kind of "prevention" for occasional stuttering audio playback due to incorrect sync signals (like MTC is by nature). Also it can't "predict" the tempo and therefor events in contrary to beeing synced internally.
There are several strategies for a sequencer to circumstance these handicaps. One is to align the audio stream to the clock via stretching the audio signal. Regardless if this is done via e.g. a flexible samplerate or granular synthesis, it leads to sound degration as well as timing issues.

wolf · Post by **wolf** » Mon Jul 11, 2005 10:03 am

errh, forgot the conclusion

If you load both recorded files into vdat (of course manually aligned in a wave editor), they should play back without phase/timing issues.

voidar · Post by **voidar** » Tue Jul 12, 2005 6:25 am

On 2005-07-11 11:03, wolf wrote:
errh, forgot the conclusion
If you load both recorded files into vdat (of course manually aligned in a wave editor), they should play back without phase/timing issues.

No, the errors are recorded. I manually aligned the tracks in wavelab for playback.

wolf · Post by **wolf** » Tue Jul 12, 2005 6:27 am

yes, but do they still have phasing issues *after* having them aligned ?

voidar · Post by **voidar** » Wed Jul 13, 2005 8:51 am

On 2005-07-12 07:27, wolf wrote:
yes, but do they still have phasing issues *after* having them aligned ?

Aligning them is really the only way to hear whether they were or not.
Listen to the file, you will hear it. At first you will be hearing a "mono" signal as they are perfectly aligned, manualy by me, but after a couple of seconds the signal recorded by Tracktion via ASIO starts to drift, chronicaly.

I only recorded using VDAT and Traction 2. All playback was done in WaveLab.

When recording I had T2 slaved to my VCR-128 via MTC, so what you initially said could be true.
The reason why I had to post-align the tracks is because T2 adds some samples pre and post to the recorded file.

I could try this again without slaving T2, just manually press record on both. It would not be much more work to align the tracks anyway.

[ This Message was edited by: voidar on 2005-07-13 09:58 ]

alfonso · Post by **alfonso** » Wed Jul 13, 2005 12:57 pm

The use of MIDI protocols to slave an audio app is considered forbidden if you want to have a minimum of reliability.

You can use the ASIO2 protocol to sync ASIO apps with VDAT, instead, the way to do it is explained in the manual.

The midi protocol is not sample accurate, so a very great number of timing errors can occur.

I've never tried it, but accuracy should be perfect with ASIO2 sync.

at0m · Post by **at0m** » Wed Jul 13, 2005 1:31 pm

Using ASIO2 Sync to slave SX to VDAT/VRC-S, it often takes 4-8 bars for SX to catch up and sit perfectly tight. Make sure you have set identical frame rates and displays, and you'll see they're a perfect match.

voidar · Post by **voidar** » Wed Jul 13, 2005 2:49 pm

For the record. I slaved using MTC which is considered to be a powerfull syncronization tool, and not MIDI Clock which is a whole different thing.

Framerate was 24 fps for both.

I was running ULLI on 4ms.

[ This Message was edited by: voidar on 2005-07-13 15:56 ]

alfonso · Post by **alfonso** » Wed Jul 13, 2005 3:16 pm

On 2005-07-13 15:49, voidar wrote:
For the record. I slaved using MTC which is considered to be a powerfull syncronization tool, and not MIDI Clock which is a whole different thing.

Framerate was 24 fps for both.

I was running ULLI on 4ms.

[ This Message was edited by: voidar on 2005-07-13 15:56 ]

MTC is not sample accurate. It's perfect for the midi sequencer, but absolutely not enough to run 2 audio apps together. You need an audiorate sync signal, MTC is a 7 bit a-sync format, totally separated from the digital clock of the cards on wich the audio is calculated. It's impossible not to have errors. What happens to your setup is absolutely normal.

valis · Post by **valis** » Wed Jul 13, 2005 5:43 pm

Alfonso is correct, you've chosen a lower resolution sync so MTC is the weak link...

Faxi Nadu · Post by **Faxi Nadu** » Wed Jul 13, 2005 5:51 pm

:0
i read over and realised my reply was not to the point

[ This Message was edited by: Faxi Nadu on 2005-07-13 18:53 ]