AK4499EQ - Best DAC ever

The IMHO is there by popular request. Some of the guys seem to think its more important to me to say it than for anyone else. Doesn't bother me to oblige. Besides, its true my views are subject to updating over time as new information becomes available. Not one of those people who is convinced they already know all there is to know, now and forever.

All the foregoing IMHO, of course :)
 
Waiting for sound quality listening reports to start coming in. Eventually someone who already has a very good dac for use as a reference will hopefully post about the audible differences.

Upsampling to DSD can means lots of things, if that is what they did. For example, with HQ Player there are lots of choices of algorithms and they don't all sound equally good.

Regarding A22 looking nice, I'm sure the manufacturers hope people will buy with their eyes, based on looking at pictures and looking at measurement numbers. Usually there is a big burst of sales for awhile for those and similar reasons.

All the foregoing IMHO, of course :)
 
There was also talk of the reason for pcm>dsd is that it allows the signal to bypass certain areas on the chip that introduces noise......is that correct?

DSD can bypass the volume control circuitry.

The thing about noise might have to do with the volume control being a dithered digital volume control, which means it has a fixed noise floor. As the volume is turned down S/N ratio will get worse since the signal is decreasing while the dither noise stays constant. However, passing through that part of the circuitry doesn't mean the signal has to be turned down there. A design could use pots or something after the dac to turn down signal and noise together in the analog domain. Of course, pots bring their own problems.

That all said, running DSD through the digital volume control circuitry at all may not end up as harmless as running PCM through it, hence the bypass option for DSD. If a post DAC chip volume control is going to be there anyway for DSD, then why not use it for PCM too?

In any case, there is nothing that says PCM should be converted to DSD to keep noise down. DSD tends to have plenty of noise issues just because its
DSD.

All IMHO, of course :)
 
Last edited:
Why does very high quality conversion to DSD512 sound good?

The guys around here with experience designing dac chips (using FPGAs, etc.) don't all agree that DSD256 (or higher) should sound better than PCM. Some say the reasons are very complex to understand. Some say they don't know why.

Also, at a very basic level there is no particular reason PCM should not reconstruct a waveform just as well as DSD if not better.

In other words, there is no simple explanation why. It just does :)

IMHO, of course.

EDIT: Probably best not to think in terms of there being a reason for lots of things. If there are reasons, they may not all be equal in contribution to a particular result or outcome.
 
Last edited:
... if the topping were to convert pcm-dsd standalone it wouldn’t be considered as high a quality as what HQPlayer does?

Highly unlikely it would be as good as the best HQ Player can do. Remember, a very powerful PC is required for the best from HQ Player.

Don't think we are near the point where the necessary technology (in the form of one or more FPGAs) will be seen in a $1,000 standalone dac anytime soon.

IMHO :)
 
Last edited:
HQ Player is paid software. It has many processing algorithms to choose from. The algorithms that sound best require a lot of very fast computation.

In other words, one not only needs to buy the HQ Player program, one also needs a very powerful computer if one wishes to run the most computationally demanding algorithms (which could be counted as an additional expense).

IMHO...
 
There was also talk of the reason for pcm>dsd is that it allows the signal to bypass certain areas on the chip that introduces noise......is that correct?

It's a lot more complex than that Bob.

There are a few DACs that have ability to play native DSD by virtue of bypassing internal modulators, vol control, dig filters etc and sending DSD directly to the OP resistive bits (DAC) utilized as an FIR filter.
This has it's own set of problems, namely ISI, which introduces distortion and I'm not clear as to how various DAC's that play native DSD handle this.

If we are talking low data rate DSD such as x64 / x128 then there will be *more noise converting to DSD. x256 starts to move the noise away from audio band and DSD512 moves noise right away from audio band.

Does anyone know if this paper is on AK4499?

A 20 kHz Bandwidth Resistive DAC with 135 dBA Dynamic Range and 125 dB THD - IEEE Conference Publication

Cheers,

TCD