gutb said:

  It sucks because many people can't afford it.
 
DSD is intrinsically superior to PCM for sound quality. That is simply a fact.

Click to expand...

 
Umm, that is not even close to a fact.

DSD much better matches the dynamic resolution of real audio. Indeed, it is much higher in terms of total resolution as compared to PCM. Early DSD (64x) on SA-CDs aren't THAT much better because much of its bandwidth is "wasted" on noise shaping. DSDx128 and above addresses that shortcoming, and much better than PCM.

gutb said:

DSD much better matches the dynamic resolution of real audio. Indeed, it is much higher in terms of total resolution as compared to PCM. Early DSD (64x) on SA-CDs aren't THAT much better because much of its bandwidth is "wasted" on noise shaping. DSDx128 and above addresses that shortcoming, and much better than PCM.

Click to expand...

You just made that up, right?

https://en.m.wikipedia.org/wiki/Nyquist–Shannon_sampling_theorem

Edit: DSD only exists because Sony wanted a format they could copy-protect.

riverlethe said:

You just made that up, right?

https://en.m.wikipedia.org/wiki/Nyquist–Shannon_sampling_theorem

Edit: DSD only exists because Sony wanted a format they could copy-protect.

Click to expand...

Exactly, DSD sucks and Sony sucks. It was a great scam though. I fell for it once and lost money.

gutb said:

 
DSD is intrinsically superior to PCM for sound quality. That is simply a fact.

Click to expand...

 
No.

ttan98 said:

 

coli said:

  See http://www.akm.com/akm/en/product/datasheet1/?partno=AK4497EQ

Click to expand...

This chipset looks interesting, it's spec already appears elsewhere on the net. However not that many here are are interested in DIY kits and also there are neither a commercial product that I am aware of or a DIY kits based on the chipset available as yet. Maybe later this year, in the meantime I am enjoying my dual AK4490.
 
Cheers.

Click to expand...

I hope this thread's not too old... but ironically I Googled: "CS4398 AK4490 ES9018 PCM1792A" and this thread was one of the first entries that popped up
 
So my question was similar to the OPs, only I'm trying to determine if the aforementioned DAC chips have a "base signature". I understand that the O/P stage plays a huge role in the final [sonic] product and I'm sure other components, such as re-clocking (when & where necessary) and so on do as well.
 
If it's not possible then perhaps we could assign a "signature" to specific DAP / DAC manufactures. To be more specific, I am looking at portable DAPs. Companies like A&K, Onkyo, Cowon, The Bit, Questyle, Calyx and FiiO come to mind. Several companies use different DAC chips in their lineups (A&K and The Bit come to mind) so that may also not be possible.
 
Finally, I understand that the best thing is to attend a Can-Jam or similar and try out each personally, with one's favorite IEMs / headphones, but this hasn't been an option for me so far... *sighs*
 
I'll take whatever you can muster in terms of facts and/or personal experience

gutb said:

  It doesn't do native DSD, no. It must convert DSD streams into a multibit format first.
 
I'll also point out that the ONLY way for your typical commercial D/A chip to handle native DSD is to completely bypass ALL of its internal filters and volume control modules, as those manipulations are completely impossible within the 1-bit realm. Filtering can only happen in the analog realm, after the D/A stage.
 
SO many lies and misinformation on this topic.

Click to expand...

Actually, the ESS chips do native DSD, but the interface does not decode a native DSD or a DoP stream directly, it needs an external chip to feed it (such as the XMOS chips), whereas it is my understanding that the AK does not.
 
The ES9018 in fact does native DSD. Everything, both normal PCM up to DXD, and DSD are converted internally to a 6 bit format (SDM) that has the same frequency of the inout or a multiple thereof, at least 2.8Mhz. Then this is converted by single bit DAC array that converts directly this 6 bit format. So there is no loss with DSD: you just convert each 1 to the value "111111" (or some other internal value, I do not know what they choose). This permits in theory also some volume control directly on the DSD stream.
 
 Roberto

I've listened to BB, CL, Wolfson, ESS and AKM and this is my personal findings:-

1. Burr-Brown PCM1792A- thick mid, good bass and slightly rolled off highs. Somewhat full bodied sound.
2. Cirrus Logic CS4398- lively sound; bass is less, mid is less thick and a bit neutral while highs is extended. No so full bodied sound.
3. Wolfson WM8740 - more bass, thick mid, a bit warmer sounding, highs a bit rolled off. Somewhat full bodied sound.
4. Sabre ESS9018- clean sounding, a bit to the clinical side, light bass, almost neutral mid and high is very extended, some 'sweetness' - not so full bodied sound
5. Verita AKM4490- good bass, thick mid and extended high, some sweetness. Overall full bodied sound - I would say, this one sound more musical then the rest.

coli said:

  DSD sucks

Click to expand...

DSD is one bit so cannot be edit. Therefore, in editing process, either it was converted to multibit-DSD (DSD 8bit) or DXD (PCM 384kHz 24 bit int or 32bit float).
And conversion between DSD<->PCM is not lossless. Do that a lot and you have ton of noise. 
 
Though, very few studio recording/editing in multibit-DSD and no conversion between DSD and PCM. Yep, very few doing good job while most of them doing poor job in recording and editing. That's another reason DSD is suck. 

So why bother with DSD when 384KHz and 768KHz DAC existed? If studio recording or editing in DXD?

Well some people are prefer DSD because of their downside

If you converting PCM to DSD (I test it with foobar in PC/Audirvana in MAC with ES9018), it sound like you EQ the song. More dB on low and mid frequency while less on high frequency. And yes without proper LPF, ton of high frequency background noise.
Some of my friends found this preferable...

Maybe It depend on individual tastes...

I think in future, we are going to have hybrid format between DSD and PCM. Something like 3072KHz 32bit float/24bit int. And we can either calling it Multi-bit DSD or Hi-res PCM.

The whole idea behind DSD superiority is the avoidance of decimation in the front end (ADC) and harsh digital filtering end point of DAC. The response impulse is almost analog in nature and have much lower noise floor up to 15kHz than conventional 24bit PCM. Also note, noise of DSD gradually rises up instead get chopped off abruptly like in PCM. Noise gradually rise up at 22kHz and above for DSD64, while DSD128 will start to rise above 44kHz and so on for higher sampling rates. More info below:

http://www.nashaudiophile.com/analog-pcm-and-dsd/

I've been listening and collecting SACDs for almost a decade, my impression of SACD and now DSD downloads are always sound more 'analog' than PCM. DSD gives a excellent micro-dynamic and soundstage, everything seemed to snapped into its space. It is less 'fatigue' compared PCM. Converting PCM to DSD in some way help to reduce the 'hard' sounding of PCM but this requires a moderate processing power.

Also native DSD recording sounds the best, they have the least editing than PCM. Do note, some DSD downloads are actually from PCM master so do check the originality. In the end of day, you have to listen for yourself to make the judge. There's no right or wrong!

Almost everything contained in this thread is either a gross oversimplification or flat out wrong, or both, (especially likely to be both if it comes from GUTB). Just sayin.

senorx12562 said:

Almost everything contained in this thread is either a gross oversimplification or flat out wrong, or both, (especially likely to be both if it comes from GUTB). Just sayin.

Click to expand...

Thanks for the heads up... I guess... :rolleyes:
I don't suppose you could contribute SOMETHING (such as a hyperlink) to point us in the right direction... ?

I don't think I need a link to support the proposition that trying to compare the sound of dac chips without any reference to their implementation is a fools errand, which is where the op started with this thread. No chip has a "sound" on its own. The only sound any chip makes without a shedload of other variables that have to be specified before any meaningful comparison could be made between them would be if you dropped them on the floor. I don't have a hyperlink for that, sorry. And I think GUTB's posts speak for themselves. No hyperlink needed to prove he doesn't know what he is talking about.

senorx12562 said:

I don't think I need a link to support the proposition that trying to compare the sound of dac chips without any reference to their implementation is a fools errand, which is where the op started with this thread. No chip has a "sound" on its own. The only sound any chip makes without a shedload of other variables that have to be specified before any meaningful comparison could be made between them would be if you dropped them on the floor. I don't have a hyperlink for that, sorry. And I think GUTB's posts speak for themselves. No hyperlink needed to prove he doesn't know what he is talking about.

Click to expand...

 
I don't agree 'No chip has a "sound" of its own. Take for instance, different type of op-amps give different type of sound signatures. If you do DIY audio, you will know why... The same goes to DACs. No chip sound the same because they are made to be different from each others. I do agreed, the design around using the particular chip do sound different but the main signature inherent on the chip itself determines the overall sound signature.
 
All modern Multi-Delta-Sigma DACs, for example CS4398 has two separate path for PCM and DSD. For PCM, it goes to the usual 8x over-sampling and sharp cut-off digital filter, while DSD bypass all these, also known as 'Direct DSD' and go directly for the conversion to analogue. Check out the datasheet for CS4398 below:
 
https://www.cirrus.com/en/pubs/proDatasheet/CS4398_F2.pdf
 
The shorter path for DSD means there's different between to PCM and DSD when playback.

FWIW...
 

traveller said:

  Interesting "interview" about a collab between Teac & AKM (from 2015).
 
What struck me most was the 5 "sound-tuning" options available with the AK4490 chip...
 

 
Makes me wonder how A&K configured them* for the 300/320/380.
 
*Teac also went for a DAC-per-Channel cfg as is the case with the 320/380s...

Click to expand...

 
 
 
Also what is assumed to be the general path of a typical DAC (but taken from A&K's FB site):