Very little technical knowledge other than textbooks. But if I understand taking a 16 bit file to 32 bits is done by adding dither. Just additional bits with no information. Then during processing the added bits are eliminated. But you should get a resolution of 16 bits, which the best DAC or CD player can't produce without this process.

Still don't have a good handle on upsampling  the file to 384 KHz. My completely subject impressions have been that a larger word length seems to add much more than the sample rate. MMMV.

For upsampling, if the start is 48 kHz, just insert 7 zeros for every sample and then use a low pass digital filter with corner frequency at ~20 to 24 kHz or something like that (aka interpolation filter). Sharp cut off. Result: 384 kHz. Filter may benefit from 32 bits of processing, but it cannot add information it does not have.

To my understanding, it's just a different type of processing. Most would be done in the digital domain since then analog will have no info from ~20 to 192 kHz, perhaps making the analog filter requirements easier to meet.

If the recording is done at a higher sampling rate and resolution using some high end ADCs, that's another story, and a DAC that can handle such resolutions might be of some use.

Meh, bs, 16 bit to 32 bit does not do much for SQ though I do prefer DACs to handle stuff in higher bit depths. The sampling rate thing is marketing fluff since pretty much every other DAC does some form of oversampling. Their DAC locking range thing is probably some form of PLL or signal buffer, which every other high end dac also has.

Possible.

Interesting they went for an 8 channel DAC. Thought those were mostly for multi-channel receiver usage. The processor is indeed 32-bit but doubt it's capable of doing more than 18 bits ENOB (THD+N on the part is -110 dB according to this brief)... if that. The implementation might take a few bits off as well. This seems like ESS flagship this days for 2-channel, but dunno for sure.

Anyhow. Again, could be an awesome piece of equipment.

And there could be a room full of engineers tearing their hair out at what the marketing guys have written. I guess that happens.

You're reading too much into the marketing speak, ultra

"requantizes 16 or 24 bit data to 32 bits" = hey, we're using a 32 bit DAC!

I remember when the ESS was hot shit and Twisted Pair kept running out of Buffalo boards to sell to DIYers, now every damn company under the sun is using it and claiming that their implementation is a breakthrough.

Ha! I still remember that, I wanted to grab the Twisted Pear Buffalo a few year ago but it always seemed to be out of stock.

Since its a Sabre, it could be more pins and needles as well.

It's probably gonna end up sounding like a reciever dac too. I think they should of kept the design more simple and stayed with 2 channel and better bur brown dac or something. I don't see why a dac should have that many freakin' inputs anyways.