Re: Layman with a mission
Re: Layman with a mission
- Subject: Re: Layman with a mission
- From: John Pitcairn <email@hidden>
- Date: Tue, 12 Oct 2004 19:23:29 +1300
From: Philippe Wicker <email@hidden>
*If* the problem is really the latency added at
some points in the chain, then that technical aspect - where is the
latency added and how can it be minimized - deserves to be discussed
here in order for involved developers to be aware of that potential
issue and do the corrections if any.
Agreed. I was about to post here, but Dennis appears to have
pre-empted me and suffered the user-backlash I was expecting. I can
say that latency using the RME HDSP Multiface is considerably higher
in Logic 6/CoreAudio than it is in Logic 6/ASIO (OS 9), which makes
something of a mockery of the claims that OS X is a lower-latency
solution.
I am hoping to engage in some sensible discussion about where this
additional latency occurs, without any partisan finger-pointing, and
without simply being referred to low-level Emagic/Apple customer
support who will simply not understand the subtleties. If it is all
down to Logic processing, we'll know where to go next.
Data, for those who have not simply dismissed this out of hand:
Using an audio loopback test to determine monitoring latency in OS X
vs OS 9 with RME Multiface, on the same machine, at a 64-sample
buffer in all cases. Direct-monitoring latency through the unit's
hardware is also provided as a reality-check:
OS 9 / ASIO / Logic 6.x:
Hardware monitoring - 62 samples
Logic software monitoring - 192 sampes
OS X / CoreAudio / Logic 6.x:
Hardware monitoring - 62 samples
Logic software monitoring - 322 samples
So it seems we have an additional 130 samples (3ms at 44.1kHz, very
noticeable in some circumstances) for software monitoring via
CoreAudio.
Questions:
Is this all the fault of the app (Logic)? In which case, why so much
additional processing delay running Logic under OS X? Stefan?
Is there any additional non-driver buffering (not by the app) in
CoreAudio compared to ASIO? Doug/Jeff?
Is this the fault of the driver (RME are bcc'ed on this)?
Bottom line is ... why?
--
John Pitcairn
--------------------------------------------------------------------------
Revolver Design | Opus Locus | Auckland, New Zealand
--------------------------------------------------------------------------
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden