dac noob is confused

[radiolistener]

Sorry to be crude, but only an idiot would buy a 32-bit DAC.  If the full-scale voltage is 1 V, then the value of a single bit is 1/4294967296 of that, or 2.33 x 10 ^ -10.  That's 23 NANO-volts!  What could you possibly build that could detect such small signals?

1 V is too small, it's better to use higher voltage, because it's hard to deal with nanovolt level noise.
For example 9 V, as used in AK4499 DAC.

Human ear dynamic range is about 140 dB.
Unfortunately existing 24 bit DAC cannot provide so high dynamic range with no distortions and noises.
This is the reason for 32 bit DAC...

Yup, I think the 32-bits refers to a "frame" which contains two 16-bit samples, one left-channel, ore right.

No, it refers to 32 bit word for single channel (mono) and all 32 bits are meaning.
AK4499 can works with up to 4 channels (mono, stereo and quadro) with each channel 32 bit resolution.
But the best 32 bit sound with SNR=140 dB is avaliable for single channel mode (mono) only.
In stereo mode it also uses 32 bit resolution, but SNR=137 dB.

[sahko123]

along with the dynamic range of 140db which for audio is about all nobody will ever need(except for those who are interested in getting the best most realistic audio perfomance(realism means dynamics reality isnet compressed)) and "32" bit performance means that high frequencies can be along with the high samping rate of 768khz can be represented more closer to an analog signal with lower top end distortion which is the biggest problem in lower quality(resolution) dacs

also im not looking for ideal 32bit performance(unless it was possible, but probably overkill) i doesnt bother me that its not actually 32bit it still has tangible performance benefits over other dacs over lower bit width.

also thanks SiliconWizard for the xmos board ill be sure to check it out further