Echo Drivers: Buffer Size / Latency

The only way I can get AA 2.0 demo software to work, is use huge buffer sizes in Audio Hardware Setup. PIII 512 rambus ram. Audio Hardware setup adjusts BUFFER SIZE. Does buffer size affect latency? Concern: latency appearing when recording in Multitrack. I use ECHO Purewave drivers in AA 1.5. Audio Hardware Setup AA 2.0: ASIO Echo WDM. Configuration Checked: Enabe ASIO 2.0 Direct Monitoring. What is ASIO 2.0 Direct Monitoring?

No can shed light on Echo drivers latency / buffers / AA 2.0?

Ypu may get more information from ozpeter later - I haven't got the Mia set up on the AA2.0 machine at all. How big is 'huge'?

Conditions that allow me to run AA 2.0 at 9632:
- up to 19 tracks with buffer size 4096
-  to 50 tracks with buffer size 16384

Quick soln: make a stereo mixdown (like what background mixing used to do in AA 1.5) and record in multitrack against this at the lowest buffer size possible. Latency would be minimal here - right? There's not much more I can adjust besides this in AA 2.0? What happens when I uncheck ASIO 2.0 Direct Monitoring? Thanks.

If I have understood you - in my machine with Echo Mia Midi I have the buffer size at 256 samples - and it seems to work OK with normal usage. My setup is similar to yours Echo ASIO WDM, Purewave etc. . I don't know what is causing the latenct, I have only run with about 2o tracks with a few realtime effects and two effects busses

P4 - 2.5GHz -1.5GB memory - XP SP2

MarkT, is your session @ 96khz, and, if it is, what effects are you running?

Typically I use a buffer size of 512 but I rarely run many tracks, and I stick to 44.1/16 - given the huge extra load that higher sample rates and bit depths impose and the (for many people) subtle differences in sound quality obtained, personally I'd go with significantly better performance from the system and forgo the last ounce of audio quality.

Put it this way - say I was running a commercial facility and in the midst of tracking, the system ran out of steam, the client would immediately be very unhappy.  On the other hand, unless the client had specified 24/96, he'd be unlikely to complain about the sound quality from 16/44.1 and the system limit would be much less likely to be encountered.

Typically I use a buffer size of 512 but I rarely run many tracks, and I stick to 44.1/16 - given the huge extra load that higher sample rates and bit depths impose and the (for many people) subtle differences in sound quality obtained, personally I'd go with significantly better performance from the system and forgo the last ounce of audio quality.

Put it this way - say I was running a commercial facility and in the midst of tracking, the system ran out of steam, the client would immediately be very unhappy.  On the other hand, unless the client had specified 24/96, he'd be unlikely to complain about the sound quality from 16/44.1 and the system limit would be much less likely to be encountered.

I am also skeptical that there is any audible difference between even a 48khz and properly dithered down 44.1khz audio file.  I do believe that there is a reason to mix sessions at higher bitdepths (specifically, to reduce rounding errors during track bouncing and destructive processing,  which may leak into the audible domain at 16bits)

As far as I can see, AA2.0 always works internally at 32 bits now?

Oz, you of all people should know that it always has.  Regardless of the preview or mixdown mode.

What I mean is that there is no longer a premixing bit depth option for playing back in multitrack view.  With complex mixes you used to be able to premix in 16 bit.  

"Pre-Mixing

Determines the bit size used for the background mixing process. Best quality is achieved by leaving this at the default 32-bit setting. However, if you're using multiple sound cards, it may be advantageous and faster to choose 16-bit for pre-mixing, as less data will be transferred across the hard drives. For single output device situations, or faster hard drives, 32-bit is better as it provides optimization at mixdown. "

That would have speeded things up in AA1.5 vs AA2.0 - and maybe other programs secretly provide playback (as opposed to mixdown) in 16 bit to accelerate the whole process?

maybe other programs secretly provide playback (as opposed to mixdown) in 16 bit to accelerate the whole process?

I seriously doubt it.  A simple loopback test would answer this, too.