Personally I avoid bit-perfect audio devices like the plague. All audio transducers and listening environments (speakers, headphones, living rooms, etc.) suffer from many forms of distortion, be they frequency response aberrations, phase misalignment (e.g. multi-driver headphones / speakers with crossovers), nonlinear distortions at the extremes of driver excursion, unwanted excess reverb and exaggerated modal frequencies in the case of listening rooms, the list goes on and on. Or how about for headphones, just the fact that a lot of music meant to be played through speakers spread 60 degrees in front of you are being played through a pair of drivers clamped on your head instead?
Many, many of these aberrations can be corrected to large extent by preprocessing the digital audio to be sent out in real time.
In my years of listening and testing it has been my experience that if the audio to be played back is still totally unprocessed (or as you would say, bit-perfect) by the time it is to be converted by the DAC to be amplified and played back by the speakers / headphones, then 90% of the fight for high fidelity has already been lost.
Why does practically no-one ever understand this?
Heck, even UAPP has added an advanced parametric EQ in recent versions. Would the number of audiophiles who make good use of it need to be counted on the fingers of more than one hand anytime in the future?
While all the issues you refer to are real, no one, anywhere, is addressing all of them in any current shipping product - let alone on the portable side of things. Companies like Linn, Devialet and Meridian do a creditable job with a lot of those issues. Linn do it, for example, by measuring every speaker they make at the factory and then having the necessary correction profiles downloadable by their Exakt engine. They also handle some basic room-mode correction. Devialet do something similar, but just with speaker correction. Meridian works on in-room measurements (or did the last time I owned their higher-end stuff).
But none of this is done at the transport ... it's either in dedicated boxes (e.g. the DEQX stuff) or it's in integrated streamer/transport/DAC/amp or DAC/amp/speaker units.
I do not want my transport mangling my audio before it's even gotten to a device that can do something useful with it. If you can build a transport that can address all those things, properly, great ... but it doesn't exist yet and the options to do it are all large, expensive and decidedly non-portable.
From a simple logic perspective, I want to preserve as much of the original data, un-molested, until it gets to a device that CAN do something usefully corrective/adaptive with it. And that's not ANY portable device today that's smaller than a laptop.
And with Android, we're not just talking about NOT doing those GOOD things that you describe, we're talking about a system that, as standard (and, as it happens even in custom solutions) tends towards a lowest-common-denominator, optimized-for-power-use-not-quality, general purpose, half-assed ASRC process that exists purely to simplify the device designers life and reduce the parts count and certainly NOT to render quality audio.
The vast majority of these devices don't even bother to run the clocks necessary not to have audible artifacts in output even when sample rate conversion isn't required. The standard conversions in the AOSP do non-integer interpolations with are inherently lossy, because they have to shift everything to match a sub-optimal clock frequency.
It's lazy, cheap, nasty and doesn't sound good.
We're talking about portable devices here, many of which don't have enough grunt to run their UIs fluidly let alone the massive overhead that comes with the type of processing required to even partially address SOME of the points you're raising. Even the EQ solutions on some of these units leaves a lot to be desired.
So, until that can all be addressed PROPERLY, I'll keep my output bit-perfect thanks - and I'll defer any other processing to devices that CAN do a decent job of it.
