new to both USBPre2 and SMAART7

[nottooloud]

I'm reasonably familiar with the concepts, having taken Syn-Aud-Con classes many moons ago, and having spent some time with the Easera Systune demo. I'm having trouble getting relevant data.

Lenovo R60 laptop running XP, USBPre2,
SMAART 7.3.2.1

Mic in front of a studio monitor, input 1. SMAART's internal pink noise.

If I reference input 2, either using the USBPre's loopback or with an external cable, delay finder gives me .91ms, nice and repeatable, and transfer traces look great. If I reference the generator directly, delay finder gives me something random between 30ms and 200ms, and the transfer traces are junk. Nil correlation. Live IR shows a stable peak, just not placed at 0 degrees.

Switching either or both input and output drivers between Wave and ASIO makes no difference.

Playing other audio from the laptop through the USBPre and referencing input 2 loopback also gives me nonsense data in the delay finder and Live IR window, although the transfer data looks OK. So the only thing I have working is internal pink listening from outside.

What am I doing wrong?

[Adam Black]

Referencing to the signal generator will incur some additional delay. However beyond that it should yield a stable measurement. Am I correct in assuming that you are using the measured delay?

[PaulTucci]

nottooloud,

Have you tried testing the tester? That's a key concept I picked up over there at SAC long ago. It may be appplicable here. Setting up a Transfer Function or an Impulse Response measurement that uses the generator for both the reference and measurements sides of the comparison should yield a difference of none in both the Time and Frequency Domains. See if that works, we'll troubleshoot from there.

Tucci

[nottooloud]

@Adam, yes, using the measured delay.

@Paul, Device Generator Mea Generator Ref Generator gives me 0 delay and ruler flat transfer.

Oy. Maybe's it's some kind of crap Windows audio thing, because today it's differently broken.

Other program playing through USBPre2, measuring 1 ref 2. Delay finder works, correlation is excellent, Live IR window shows a lot of amplitude but the peak position is stable.

Internal pink, measure 1 ref 2, everything works, all windows stable.

Internal pink, measure 1 ref gen, delay finder random, Live IR stable although clearly wrong, with a nice clear peak somewhere non-zero.

So now Ref Gen is the only part that's not working.

It's not the USBPre2, anyway. Just swapped in an E-Mu USB 0404 and all results are identical.

The laptop's a 2GHz Core Duo w/ 2g RAM. CPUs are running 60-80%, depending.

[Adam Black]

Internal pink, measure 1 ref gen, delay finder random, Live IR stable although clearly wrong, with a nice clear peak somewhere non-zero.

The delay when referencing to gen won't be correct. Referencing to the generator will add artificial delay. With ASIO drivers the added delay will be consistent. With WAV drivers however the delay will change each time you start/stop the measurement. This is due to how WAV works and and is unavoidable. Though we intend on making a change what will at least keep it consistent for the session, though not between sessions.

-A

[PaulTucci]

Internal pink, measure 1 ref gen, delay finder random, Live IR stable although clearly wrong, with a nice clear peak somewhere non-zero.

Is there a correlation between the stable-but-non-zero peak and the time window of the IR?

Is the peak at the maximum time of the time window? Larger time window yielding a peak that is further away in time than the reference?

Sample Rate / FFT size = the Time window of the measurement

PT

[nottooloud]

@Paul, no, delay times were all over the place.

@Adam, that's got it. ASIO on both gen and input does indeed give me a repeatable delay time, albeit 33.3mms as opposed to .6ms. If I then stop and restart the transfer, the data's all good. I hadn't been doing the stop and restart with the ref 2 configurations, and that makes all the difference.

Here's an apparently complete fix, though. If I turn on delay tracking, everything works regardless. Seems I can even mix and match drivers. I didn't think that was supposed to work.

Thank you, gentlemen.

Now the only bad news is that I'm going to need more computer. And hey, that's never really bad news, is it?

Dan Richardson

[Adam Black]

Here's an apparently complete fix, though. If I turn on delay tracking, everything works regardless. Seems I can even mix and match drivers. I didn't think that was supposed to work.

Hey Dan,

Mileage may vary when mixing driver types and I wouldn't recommend it. In your case I'd stick with the ASIO drivers.

-A

[nottooloud]

Low overhead playback software on Windows with stable ASIO drivers is a rare beastie.

[nottooloud]

So it finally occurred to me to use the onboard generator's WAV ability. Unfortunately, my laptop is slowish. Playback is fine, but when I start the transfer function, the audio stutters. Classic sound from a DAW trying to run an ASIO driver with too small a buffer and not being able to keep up. Unfortunately, I don't see where in SMAART one sets the ASIO buffer length. Am I missing something?