It seems like a lot of audiophiles get hung up on the "32 bit" sigma delta DAC chips (ESS, AKM, Wolfson, etc). Problem is these chips including the highly touted ESS chips do not provide more analogue bit resolution than the TDA1541 multibit chip from the 80s. At least this is my understanding. Here is a link to a very well written response from a different forum:
http://www.audioasylum.com/audio/digital/messages/15/153466.htmlWell, I''ll just post the god damn thing:
"Posted by Thorsten (M) on July 31, 2010 at 05:38:47
In Reply to: RE: Sabre DAC not 32-bit? posted by sheppard on July 27, 2010 at 13:31:52
Hi,
Well, I think we need to define what we mean with "32 Bit DAC" if we want to see if the claims are true or not.
We can interpret it in many ways:
1) A 32-Bit DAC is one that shows 32-Bits equivalent analogue resolution, that is 183dB dynamic range, measured the traditional analogue way. It goes without saying, not only can we not measure such a DAC should it actually exist, but of course no such thing exists nor is it possible, outside a laboratory system suspended in liquid nitrogen and maybe not even then.
2) A 32-Bit DAC is one that uses 32 individual binary weighted bit switches, so it is theoretically capable of producing 2^32 discrete steps, though it's analogue dynamic range is less than the postulated 183dB (or 192dB as some may say). No such thing exists either (yet), but it is at least theoretically possible to make such a device.
3) A 32-Bit DAC is a DAC that accepts a 32-Bit wide data word and outputs whatever real resolution it is capable of, in other words it is marketing number without any appreciable meaning.
It should be added that recent specifications for computer based audio (Intel/M$ HD Audio) call for systems that are able to handle 32 Bit words, simply because this is how computers like to work, they want 8/16/32 Bit words to work with, not 24 Bit. This is the reason for the 32-Bit DAC's now becoming more common.
This has no meaning other than a DAC should accept a 32 Bit word, for compatibility, not that it that actually does anything meaningful with the whole 32 Bit's. In fact, several "32-Bit" DAC's simply take the 32-Bit Data and dither it down to 24-Bit which is then applied to manufacturers 24-bit DAC Core.
However, as it is very easy for marketing departments to point out that "32-Bits are better than 24-Bits (even though under my definition in 1) there is no such thing as a 24 Bit DAC either) just as 24-Bits are better than 16-Bits" and so on, the 32-Bit part has become a major marketing macguffin for DAC (Chip) manufacturers and their customers.
Okay, on to the Sabre (and for that in principle almost all modern DAC's).
In the year of our lord 2010 all but one DAC targeted at audio use a concept that used to be called "hybrid DAC". This means these DAC's combine several bits worth of multibit core with a delta sigma modulator (aka one bit DAC).
The combination is used to achieve the total resolution by using a process called noise shaping from the "real" resolution of the DAC (that is the number of levels that the DAC can directly represent in the analog domain and the additional resolution attained using noise shaping. It is a little difficult to understand and even for those who understand to explain where this extra resolution comes from, but we do not worry ourselves here about details.
The bottom line is that the DAC will have a number analog levels that can be represented directly by 1-Bit and Multibit conversion. The rest has to be produced using noise shaping. The views on noiseshaping vary, my own is negative, where sonics are concerned, compared to having enough real resolution.
Back to the Sabre.
For the "6-Bit DAC" for the Sabre, this is both correct and incorrect.
The Sabre has 2^6 or 64 so called "unitary weighted" or "thermometer code" bit switches. These are able to represent directly 2^6 or 64 individual levels.
Further, the Sabre uses asynchronous sample rate conversion on ALL Input data and converts into a clock rate of 40MHz. If we assume for ease of calculation a 50KHz data sample rate (close enough to to the 44.1KHz used on CD) we can represent as many as 40MHz/50KHz or 800 individual levels using classic pulse width or pulse density modulation.
For ease of calulation I will round up to 1024 levels, which is equivalent to 10 Bit resolution.
This means that the raw resolution build into the ESS DAC is around 16 Bit for single speed (44.1/48KHz) Data, 15 Bit for double speed (88.2/96KHz) and 14 Bit for quad speed (176.4/192KHz) data.
I have to say that this is appreciably more real, raw resolution than most DAC's in the market offer. The ESS Sabre DAC's can actually represent CD Data in the analogue domain with no or very little noiseshaping, which may explain the fact that many find it superior to many other DAC's.
For reference, a highly regarded (by some anyway) 32-Bit DAC by another manufacturer uses a 32 Level (5 Bit) multibit section and 128 Times oversampling at all datarates (7 Bits), thus meaning the actual core of the DAC is able to provide only 12 Bit real resolution without noiseshaping.
Certain others are even more miserly on real resolution, because real resolution costs real money and why bother if you can fake it in the measured performance by agressive use of noise shaping?
It means the ESS DAC relies appreciably less on noise shaping to represent the full needed resolution than most (or at this time perhaps all?) others using the same principle, though it is less than what is attainable using a true multibit DAC. In fact, it is barely able to match the mid 1980's TDA1541 in terms of real (non-noiseshaped) resolution.
For reference, if we combine analogue resolution (24 Binary weighted bits) and the possibility to run at 8 Times oversampling (3 Bits) the Burr Brown PCM1704 (the last true multibit Audio DAC in production) allows us in effect 27 Bit of analogue levels. Sadly this chip is hampered by a SNR/Dynamic range of much less than 120dB, so much of that possible resolution resides below the noisefloor and is of no use. We woudl have to parallel humungous numbers of PCM1704 DAC's to push the noisefloor low enough to make use of the extra bits.
Of course, non of what is written above has any direct relation or mapping onto perceived sound quality. However, a personal observation is that I seem to like DAC's sonics in about the inverse of the amount of noise shaping used. Equally I also know (of) experienced listeners who have a reasonable track record judjing sound quality whose reaction is the opposite (the more noiseshaped the system the better they like it - SACD/DSD being one extreme example).
So the bottom line is - listen for yourself and select what sounds best to you.
Ciao T"
Some may disagree with this guy (like whether noise shaping is a bad or good thing), but he makes a lot of sense to my feeble mind. I think the only DAC I know that can really approach 24 bit analogue resolution is the MSB Diamond DAC (they claim 27 bit resolution), but this is a true DISCRETE MULTIBIT DAC using resistors used in the aerospace industry instead of silicon. Problem is the Diamond DAC start at $25k.