I think the discussion of any DAW's summing architecture applies most directly during the mixdown stage as opposed to the recording or playback stage.
This issue was kicked around,
(and still is), on the Cakewalk forum for ages and ages -- then Cakewalk shipped Sonar v3,
(and I know Dachay still won't believe me, but.....), there was a quantifiable difference in the sound produced by the same mix when bounced out of Sonar v2.2 compared to v3.xx. I had a thread kicking on around here at the time where I posted clips of two mixes -- one out of Sonar v2.2, the other from v3.0
(both still sucked, mind you) -- but they did sound different and wouldn't null to zero when re-imported them into a new Sonar track and flipped the phase on one of 'em.
Therefore, something HAD TO BE different in the audio engine.
(Actually, to give credit where it is due, Dachay actually pointed out that while there may be a perceptible 'difference' in the sound, it didn't necessarily equate to 'better'....)
Anyway => Sonar v3 shipped with a shiny new audio engine -- as touted by the marketing folks -- but plenty of folks on the Cakewalk forum were still complaining that "my mix sounds clearer/different/betterer/more-sooper-dooper-unrealer etc. in Cubase". Ron Kuper, the lead dev. genius for Sonar got on and posted saying that the explanation was to be found not in the summing algorithms of DAWs, but in the way 'pan law' was applied between the different applications.
When Sonar v4 shipped one of the touted features was 'configurable pan law' -- that is, you could specify different approaches for the amount by which any track's level was adjusted according to where it was panned in the stereo field. The bakers at Cakewalk maintained this was the explanation for any perceived differences between different DAW engines. Ron K. went to extreme lengths in numerous threads -- particularly with some dude called 'shea' -- to address these issues scientifically and to demonstrate these assumptions with fact against the S4 engine.
It was never explicitly stated,
(that I saw), but Ron K. did seem to imply on a couple of occassions that there were some underlying structural changes to the audio engine of Sonar v4 compared to v3 which may have aided this cause... Ultimately, I don't know...
Anyway, the consensus of those threads seemed to be that if you do a mix in any DAW, then do another mix in Sonar using exactly the same pan law and FX set, re-import them both into new tracks in any DAW, flip the phase on one and play them back -- THEY WILL NULL TO ZERO. This means they are EXACTLY THE SAME.
Game over - end of the
"DAW Wars", right there.
...
If you want to waste more time, then there is the 'Awesome DAWSum' experiment on another reputable BBS -- here you get a CD full of the same mix with the same settings all exported from a bunch of different DAWs,
(you will have to shell out $$ for the CD - no download). Again, any meaningful analysis of their forum will quickly prove that once again, like arseholes, every person has an opinion and most of them stink...
No-one agrees on anything and personal preference and individual bias rules all...
The ultimate bottom line is this -
Don't waste time worrying about your sequencer software. Choose one. Record some damn music. Arrange it. Edit it. Learn from the tool in front of you and the community resources from which you can draw knowledge. Learn... Grow.
Repeat...
Ultimately, your art --
be it music or engineering -- is more important than the tool you choose to use.
...and above all - have fun!
Q.
