The Leveller wrote:Since I don't buy into the 'live's audio engine is inferior' argument ...
This depends upon what you conside the audio engine. An average user consider anything that in anyway impacts audio playback and rendering (including wap, automation, clocks etc) - from that point of viw - it is inferior and objectively probably so beyond any doubt.
OTOH, if you are only considering the basic audio path processing without any consideration to how plugins are controlled and how plugins might impact the audio and also exclude consideration of in-the-box plugins (ie default state of a new set with unwarped, unlooped and at project sample rate audio clips) and exclude sample rate conversion then its fine - it appears to do mathematically correct floating point summing and panning - but lets face it - thats *all* it has to do. I would ever say I see nothing wrong with its export dithering either.
I think the problems people experience lie once you venture past this perfect pristince state. Some of them I have to say are logically bullshit and lie within the realm of placebo and faith etc, but I also have to say that over the years there have been a number of occasionaly that have made me question lives audio quality after discounting all logic explanations except for these:
1. Some DAWs *could* do oversampling (x2 or x4) in their native audio path resulting in a higher apparent sample rate through plugins (I am not talking about internal oversampling in native EQs etc). However I know of no DAW that actually does this, so I guess discount it.
2. Lives default bit depth coversion from internal processing to 24/16 bit audio hardware may be flawed in some wierd way, particularly at sample rates where the ninquist frenquency is close to audio range - ie 44.1k, and to a lesser extent 48k causing reflected overtones back down into the audio range - increasing apparent harshness/graininess of sound, potentially reducing clarity etc. I cant think that Live can exist for so long and something like this to still be the case also given that ableton have apparently does quite a bit of internal research and answering of such claims.
3. Lack of clock latency compensation. Yes sure - we know all the obvious impacts upon LFOs and other blatently obvious time syncronised processing thanks to recent threads and testing folks have done. But what about the non-obvious? What about the potential impact upon any plugin that relies upon a perfectly compensated VST time (AU equivalent) structure to do its historical processing correctly?
I dont know these answers, but I cant discount these possibly more subtle effects either.
I think we have the be a litttle more open minded about the possibility of a ghost in the machine somewhere and not allways be so quick to dismiss these claims. I have had several times when i have thought a mix sounded better in something else (including time when I didnt know that the person iwas workign with at the time had recreated the mix in something else), but Ive put it down to not quite standing in the same place, someone else not standing in the same place (and so impacting audio wave refraction, so in turn impacting perceived phasing and perceived frequence balance due to cancellattion and re-enforcement etc). However after recent threads and doing some fo the testing myself and tthinking through some things from the pespective of a develpoer who understand audio processing and the VST SDK at least - there are now alot more things to question and alot more unknowns that cant yet be dismissed.
Maybe something to poke around with next week when I get released from corporate slavery and get to enjoy a month or more as a jobless music bum
