Audition 2 and ASIO - problems

could you post a link to the Ambience reverb VST?

http://magnus.smartelectronix.com/

So - I'd be interested to hear whether others having any difficulty with multitrack replay of intensive effects loads can get them to play through those methods.

I loaded my 12 tracks of cd audio, ASIO buffer size set @ 128 samples, and these are my results:

Audition: I can run 7 tracks of Ambience's Amoebe Hall preset without problem.  Any more and the audio starts glitching, no matter if I disable the time display and turn off the output meters.

Krystal: Oddly enough, I can only run 5 tracks without glitches, any more and it starts having problems.

Sonar: I can run 13 Ambience's (1 on every track, then another on track 1) perfectly.  Any more, starts glitching.

Hopefully others will do the test as well...

Interesting...

Try "Bitviewer" in the master channel of both AA2.0 and Sonar to ensure both are replaying in 32 bit.  (see http://www.kvraudio.com/get/1654.html if you don't already have it).

Try stopwatching a mixdown in Sonar and AA2.0 with the identical number of instances of Ambience in both.

Thanks so much for that tool!  I was worried that Sonar was playing back in 24bit.  Luckily I was wrong.

BitViewer said 32bits were passing through it when I threw the VST on the master track in both Audition 2.0 and Sonar 4.

The other way to verify that the "playing field is level" would be to mix down a session in Sonar, insert that into the Sonar session, reverse the phase of it, and see if when you then play back in Sonar you get silence.  In other words you'd be proving that Sonar's playback is the same as its mixdown.  Take it from me, that's the case in Audition.

I'm by no means discounting that Sonar's mix engine may be superior to Audition's but before saying that I'd want to be as sure as possible that Sonar is not cutting corners in some way.

Just tried the "Ambience" test with Tracktion version 1 (the free one needless to say) and performance seems broadly in line with Audition's figures.

The other way to verify that the "playing field is level" would be to mix down a session in Sonar, insert that into the Sonar session, reverse the phase of it, and see if when you then play back in Sonar you get silence.  In other words you'd be proving that Sonar's playback is the same as its mixdown.  Take it from me, that's the case in Audition.

I'm by no means discounting that Sonar's mix engine may be superior to Audition's but before saying that I'd want to be as sure as possible that Sonar is not cutting corners in some way.

That is the "problem" with Sonar 4 - it doesn't support 32bit mixdowns (Sonar 5 actually supports 64bit IIRC).  

If I put the mixdown on a track in multisession and solo it, bitviewer states that it is 24bit.  

I couldn't quite figure out how to get the silence in Audition's multitrack view.  I mixed down the session, inverted the phase, stuck the mixdown on its own track, but I could still hear everything (bitviewer states the mixdown is 32bit here).

Just tried the "Ambience" test with Tracktion version 1 (the free one needless to say) and performance seems broadly in line with Audition's figures.

I have a free version of Tracktion lying around here somewhere.  I'll install it and test in there (could you give me some help with the basic interface, tracktion looks very alien to me, and I was barely able to use it when I tried a long time ago)

Somewhere back in this thread it was stated that with the new ASIO support and the removal of the backgroud mixing stuff, AA reads directly from the HDD and spits out the audio using the ASIO drives to create magic. Why isn't my HDD clicking away like crazy when playing back audio? Also, if I load a session from a mapped network drive I see no degradation in playback performance. That doesn't make sense to me.

One comment regarding this conversation...

I only have Sonar version 3...but based on that...

I use Audition exclusively for my mixing of wave files because I do NOT consider the effects in Sonar to sound anyplace near as good as the ones in AA.  The best real time reverb in Sonar...to my ears at least...sounds less good than Audition's "Studio Reverb".

I have similar opinions of the dynamics processing and EQ available.

Perhaps it's just me...and perhaps they've improved in later versions...but this could also be a reason why you can run more effects in Sonar.

Bob

I am very skeptical about that,

but we can subjectively test it by recording a track in audition, then apply studio reverb, then loading the dry track in sonar and tweaking with the verb in attempt to get the same sound, or vice versa.

Rummaging around the net (OK, wasting time) I notice that Sonar doesn't actually support VST effects natively but via some kind of wrapper - and that this is a source of complaint from users, and I came across a company statement indicating that to encompass VST support within their native code would require a considerable re-write.

I can't help but wonder therefore whether what seems to be the ability of Sonar to support many more instances of a VST than Audition is due to some kind of corner-cutting via this "wrapper" that they seem keen not to part with.  Though intuitively, you'd think that something done, as it were, second-hand with a wrapper would be a cause of inefficiency rather than the reverse.   But then again, Asio4All seems to be classed as a "wrapper" of a different type, and works well.

Interesting stuff.  But at the end of the day I guess one would have to know a considerable amount about the Sonar code and about the Audition code for a real assessment of the differences to be made.  And that ain't going to happen!

I believe that the latest version 5 of Sonar does now support VST plugins natively. Whether it still works as well I don't know.

Rummaging around the net (OK, wasting time) I notice that Sonar doesn't actually support VST effects natively but via some kind of wrapper - and that this is a source of complaint from users, and I came across a company statement indicating that to encompass VST support within their native code would require a considerable re-write.

I am interested in this. Can you post a link?