|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
88.104.184.35
In Reply to: Re: DSD1790 in Denon 2900 posted by audiophilect on April 20, 2007 at 10:31:18:
Hi
Can you state which filter is supposed to be OFF..is it the roll off 100khz or 50 khz or another one set from the menu.?
I would assume that IF these DSD filters (simple caps) are switched in then the signal must already have been converted to analogue by the Cirrus.
Looking at the circuit diagram (service manual) the pcm and dsd are passed as either to the same pin on the DSD 1796 chip.
Follow Ups:
The filter I am referring to is the bass management filter on Denon 2900.If you look at the block diagram for 1796, there is not a separate path for DSD stream. It goes through the same 8X oversampling filter etc. Therefore, I suspect that DSD is being converted to a PCM-like stream.
I don't think the block diagram is correct. There would be no need for special analog filters in "DSD mode" if the signal were converted to PCM....there would be no additional out of band noise to filter.....so, I don't think the 8 times oversampling filter is used in DSD mode. When you see a DSD signal converted to PCM there is no "DSD out of band noise" on it(observed with Ocsilloscope). I have seen the output of all the Burr-Brown DSD DACs and everyone of them have the "DSD out of band noise" on them...even the $4.00 8 channel jobbies. The Oppo players that play SACDs convert the DSD signal to PCM at 24/88 and the output waveforms looks exactly like a 24/88 output..no additional noise.
First of all, the original poster said "PCM1790", which is not even listed on the Burr-Brown (TI) website. However, Burr-Brown *did* make a PCM1790 as a custom part for some large Japanese manufacturers. As far as I know, the PCM1790 does *not* accept "DSD" data. So if the player actually uses a PCM1790, I believe that it *must* convert "DSD" to PCM. (I could be wrong about this, as I am going on memory here. I think I recall seeing a Pioneer service manual for a DVD player that used the PCM1790 and it included an abbreviated data sheet for the PCM1790.)But assuming that the later part number of "PCM1796" is correct, this does indeed accept "DSD" data without converting it to PCM. The block diagram in the data sheet is, in fact, incorrect (as are the data sheets for all of the other Burr-Brown DAC chips that accept both PCM and "DSD" data.)
The only place I have seen a correct block diagram is in a pre-print that the Burr-Brown engineers presented at an AES convention when the PCM1738 was introduced, entitled "A 117 dB Dynamic Range Current Mode Multi Bit Audio DAC for PCM and DSD Audio Playback". In this paper, the block diagram clearly shows the "DSD" data being directly fed to the "current-segment DAC" block. (The official data sheet does *not* show this.)
DSD1790 is made for Denon, therefore is not listed on BB's website. Its block diagram in the 2900 service manual is nearly idential to those of 1738, 1792, and 1796. Maybe these diagrams are indeed not accurate in describing how DSD signals are handled.BTW, what are the primary parameters characterzing a DSD DAC (or DSD part of a DAC)? In the case of PCM DACs the stopband attenuation of the digital filter is a good indicator of the performance (and its price). What are the good performance parameters for DSD DACs?
The ultimate DSD playback deck is the one that sounds the best...period. Trying to figure out which DSD capable DAC chip would sound the best by the specs is really silly. There are some that say that you should change DSD to 24/192 and decode for best sound (Alex Peychev, among others). There are those that feel all digital filters used for PCM suck...but if you must use one....then I suppose having more stop band attenuation might be better for lowering noise so you can use less output filtering. Some like the Analog Devices DACs, some like the AKM (Alex again)....some don't like the AKM (Steve Nugent of Empirical). Some like the Burr-Brown....some like the NPC chip set....some make their own DACs (DCS and Meitner, among others).....It all comes down to execution.....a $3.00 8 channel DSD capable DAC chip in a totally tweaked out modded player will sound much better than the "best speced" mono DAC chip in a stock player. I am not against measurements....it's just that the mesasurements most are looking at have very little to do with the resulting sound. Changing one decoupling/filter cap on a DAC chip can make a big difference....let alone the whole output stage, clocks, power supplies, damping, shielding, etc. Once you have played with these things then you know just how complex the whole thing is......in the end, you must trust your ears.
I remember talking to Burr-Brown about this chip and they would not tell me much but I got the impression that it was the first implementation of the upgrade from the 1738....sort of a quasi 1794/6/ etc. So, I think it does DSD separately, has current out and slightly better specs than 1738.....could be wrong...it was a while ago.
But according to Denon, DSD data is not converted to PCM unless Bass Management is set to "ON". When the DVD2900 came out, Denon did spend a fair amount of time promoting this feature.One day I may modify mine with an SDI output (I already have the Pixel Magic DIY kit). Of course, I'll need a video processor to take advantage of it.
:-)If anyone's interested, here's a link to their product archives for the DVD2900 (several .pdf downloads as well):
the BB data sheets seem a bit "coy" about exactly how DSD data is handled in the DAC.As we know, the 1738 (and also the 179x) is a hybrid DAC consisting of multi-level for the upper 5 bits and 5th order delta sigma for the low 18 bits.
Either DSD is being piped directly to the delta sigma modulator, or (possibly) it is being partially decimated to fit into the hybrid architecture.
I suspect it could potentially be the latter (there was an early Sony white paper for the XA9000ES i think that hinted at this).
That could explain the block diagram discrepancy. Otherwise it seems silly that they haven't corrected the block diagram over several generations of DACs.
| ||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: