Recommendation for DAC needed.

Benchmark DAC3 beats pretty much everything at its price and most all above, but it’s already practically totally transparent. An SMSL SU-8 is very close to performance (backed by measurements) and costs a lot less.

The Bryston doesn’t offer anything over the Benchmark. Similar/worse performance for more money; it’s excellent, just overpriced. I do dislike the asthetics of Benchmark though.

@rvpiano 
 
Not sure about the Qutest, but even the Chord Hugo 2 doesn’t touch the Benchmark or even SMSL SU-8. The Auralic Vega and Altair also don’t perform better than the Benchmark (similar to SMSL SU-8).

@celander  
 
The Topping D50 is indeed excellent for the money. The ~$1700 Benchmark DAC 3B is still better (the ESS hump may be audible, not sure on audibility thresholds), so if he has $3K to burn, then <$2K for the DAC3 B is still saving >$1K.

@celander 
 

I presumed running it at 0.00 dBFS would remedy that concern.

 
That graph is plotting dBFS along the horizontal axis, 0dBFS means music that is at the loudest it can go before clipping, undithered CD has a noise floor of -96 dBFS; for modern music, 105dB is peak, and since the average treated room noise is around 30dB (my living room is ~45dB as it’s open, so things like the refrigerator’s compressor and whatnot contribute), so that’s a range of 75dB, so everything from 0dBFS to -75dBFS needs to be looked at. Now, at -35dBFS the IMD of the D50 is -65dB, so 5dB, so audible but not in-room. There’s still stuff like IMD+N though. 
 
Also, while the 2Vrms output of the D50 will allow most any amp to reach its spec’d wattage output, the much more powerful output of the DAC3 will allow you to do heavy DSP (which involves lowering headroom) while still keeping >2Vrms. 
 
But yes, the D50 and similar put to shame so many esoteric DACs that can cost $5K of more. Now, something DACs do perform better than the Benchmark, it’s just we are well past audibility, I would only pay a tad more if the performance was similar but the asthetics were better.

@rvpiano

The digital out of the CCA is bit perfect if using a non-Google casting app If you cast from say Roon or Android Hi-Fi Cast, it will have the same performance as even $5,000 streamers. It’s only if you Cast from say YouTube or your Chrome Browser where casting stinks. 
 
Keep in mind that digital signals are identical except for jitter, and the jitter present if you cast from say Roon, would be pretty much limited to what DAC you use, as the CCA is not the limiting factor, again, unless casting from say the Chrome browser.

@nthuerk 
 
So every DAC except Schiit’s MultiBit sounds like ass, sure. Schiit only recently got AP gear to actually test their performance while designing, and notice how they made the Modi 3 and not another MultiBit DAC. MultiBit is snake oil. 
 
@rbstehno 
 
PS Audio is good, but similar/worse than Benchmark, absolutely not better if you did a blind listening test.

@stevehardy1 
 
Paul McGowan still recommends preamps, so...

@shadorne 
 
Digital volume control is only an issue with 16Bit DACs, 24Bit should be not audibly difference than analog volume controls, and for sure 32Bit.

An analog volume control reduces both signal and noise together. A digital volume control only reduces the signal level

You put that out as a statement, not technically in relation to PS Audio, but you did talk about them in the prior sentence, so I can see that.

There is nothing on the market with an analog noise floor of -192 dBfs.

Correct, as the noise floor is reduced due to the performance of the analog components in the DAC. The 32Bit chips on their own are -192dB, as that’s what 32Bit is. The ESS 9038 and AKM 4499 is -140dBFS, and was ~-130dB a few years ago, so slowly progressing forward.

So, even ignoring how they can lower the noise floor with attenuation, if you did the same -40dB attenuation, that would be -100dBFS, still a greter dynamic range than what we have in-room (around 80dB).

@shadorne

A 32Bit DAC will have a noise floor of ~ -192dBFS. Meaning even if you turn the volume down from full scale to -40dBFS (much quieter than what we listen at), then the noise floor has risen to -152dBFS, still better than 24Bit.

Like I said, it’s a non-issue with 32Bit.

Its even more of a non-issue with say an ESS 9038, which has access to its internal pathway and decreases the noise floor just like an analog volume control (down until it hits its own noise floor of course). Good DACs have been able to do this for >5yrs.

@jeenam  
 
Sucks to see that, but it is bit-perfpect if used with the programs/apps I mentioned, measurements. 

@jeenam  
 
Without measurements, you can’t take his statements as truth. I think I did find that Spotify too is not bit-perfect through casting, which is a shame. I have no clue why Google cant get this right, but I know as a fact that using Roon or an app like Android HiFi Cast will be bit-perfect.