SNR of 24 bit might be problematic physically. However dynamic range of this much is achievable. Zoom F6 and the current Sound Devices MixPre recorders can record in 32 bit float, and they use parallel DACs optimized for different signal levels so you can have a dynamic range closing in on 24 bit results. I don't know this is important enough it will become widely available in the near future, but it is possible. Its usefulness in field recording is you don't have to set recording levels and you never get a clipped recording.
ADCs are pretty much solved as well. The very low end isn't SOTA, but you can get close for not much more, and even low end models are good enough they aren't any kind of impediment for one's needs.
Really other than creature comforts and such everything in audio is solved except transducers.
Yeah I was actually reading something to this effect recently, that higher dynamic range is possible by splicing different parts of the signal range to be handled by separate DACs. Just seems expensive (or just unconventional?). Not sure which, but I think this is really cool that ADC's seem to implement this regularly more than DACs themselves?
As far as ADC's being solved, I was not clear. I mean't solved in the sense of performance at costs seen from DACs for example. Also even if their performance was at such (I know RME's PRO is pretty SOTA), and even if the price was there. There just doesn't seem to be many ADC+DAC bundled devices nor mini-devices like the USB DAC sizes or inline mics as options for headphone cables and such. So solved in my machination would include availability, and various number of various offerings.
Speaking of 24-bit ADC's, it seems LG's V series (at least the V50 and later for sure that I know of) have recording capability of 44.1, 48, 88.2, 96, 192 all at either 16, or 24-bit ranges, and can be saved instantly as M4A, WAV, or FLAC. Really cool to see that.
Also, I don't feel filters are SOTA yet either, aside from FPGA implementations like the sort from Chord using million taps supposedly, where you get a really nice brickwall. ESS is good enough though, but AKM I'm a bit disappointed didn't have a nice 22kHz brickwall from their latest 4499 offering (but it seems they're doing this on purpose or is a technical limit of their design as someone spoke to me about a while ago).
Oh and DSP is definitely not solved. Even my RME DAC doesn't do low pass or high pass filtering as well as I'd expect (it's pretty lazy in terms of the slope dB/octave). Compared to JRiver for example that can muster 48dB/octave, the RME seems to be somewhere around 12db/octave or so(maybe better, but I'm not sure). Not bad, but I want more agressive sloping (though of course this could be included in your notion of creature comforts). Not sure how scaling of processing power that is required here, but surly I would imagine we'd be more far ahead at getting more DSP in more of these audiophile devices.
Also, do you think AVR's and Power Amps should be in a better place by now, or do you think their performance is gated by our best minds on the matter itself? And one final question since I highly value your opinion and have learned a lot from reading your technical exchanges with others. The transducer issue, what is the actual problem? Is it simply a material science ordeal? Inability to find a material to suit the higher performance aspiration. Or is it something else like dumb enclosure designs that eventually ruin nearly all transducer implementations? I'm really curious to know what exactly is the problem. And how far are we from what you would consider transducers to be as "solved"?
Oh and would microphones themselves be considered in your transducer defintion? I've seen some crazy vid Amir posted a little while back about some scientific microphone picking up on the chewing noises of some leaf eating worm and such. Any thoughts?