DSD over HDMI - Players - Page 2

Not according to Sony. They specify DSD w/ HATS for the connection to these two players.

I highly doubt you will hear a difference between this player and my onkyo 506 when running dsd over hdmi with all other things being =. The big selling point of the XA5400ES is its D/A converters which wouldn't be used if running dsd over hdmi... Save the money and put it towards speakers.

Kal, I have posted before on several Forums . When I asked a Sony retailer to find out if the STR-DA5400ES receiver decodes DSD direct to analogue without internally converting to LPCM, Sony said "The DA5400ES uses the full resolution of the DSD signal (32 bit) from input filtering through to processing the signal; then and only then does the amplifier at the final stage convert to PCM. This way you can ensure the FULL resolution of the DSD signal can be fully utilised. However Tim-Vine Lott of Air Studios says this is misleading and blatantly incorrect. I assume this is because when DSD becomes LPCM ,DSD's 'uniqueness' is diminished. However does it really matter ?

That conflicts directly with Sony's white-paper which describes how, after the upsampling of RedBook PCM to the equivalent of DSD, both PCM and DSD sources are decoded as DSD. (Sounds a bit like what Esoteric offers as an option.)

EDIT: What you are quoting is for the AVR and my statements refer to the processing in the XA5400ES player. Sorry for the confusion.


Does it? Depends on the individual, imho.

. April Hi-Fi News ALL OF A JITTER. Sony claims to maintain the high dynamic range of SACD (at bass/midrange rather than high frequencies) by decimating the I-bit/2.8224MHz DSD to 30-bit/176.4kHz LPCM before D/A conversion. If nothing else, the downsampling is necessary because most numeric DSP (soundfield modes etc) will only operate on LPCM. Paul Miller Editor Hi-Fi News& RR.

But that is in the AVR, not the player, right?

YES. In answer to my question, PM was referring to the AVR STR-DA5400ES .. So it seems Sony is putting out conflicting information in the White paper you mention. Perhaps I should quote the whole of what PM says rather than selective passages.
If you have read our March issue then you'll have discovered by now that Sony's HATS for HDMI regime confers its XA5400ES/DA5400ES combination with exceptionally low jitter with both LPCM or DSD data. The DSD data used by SACD is described in single bits but at an extremely high 2.8224MHz sample rate (64x the 44.1kHz sample rate of Red Book CD). To fully realise the monotonic potential of this bitstream it should be converted using a true 1-bit' 'bit converter'. Unfortunately, to the best of my knowledge, only the very first SACD players (including Sony's SCD-1) employed such a device. Modern 'universal' CD/DVD/SACD players employ similarly universal DAC technologies. These ICs feature DSD compatible inputs but the bitstream is invariably downsampled to a lower rate at a higher bit depth, (conversely the LPCM inputs are upsampled and truncated to a similar 4-5 bits) before D/A conversion. At this point DSD becomes LPCM and its'uniqueness ' is diminished. Sony claims to maintain the high dynamic range of SACD (at bass/midrange rather than high frequencies) by decimating the 1-bit/2.8224MHz DSD to 30 -bit /176.4kHz LPCM before D/A conversion. If nothing else,the downsampling is necessary because most numeric DSP (soundfield modes etc) will only operate on LPCM. Sony's SCD-XA5400ES player is equipped with a 'universal' DSD 1796 DAC from Burr-Brown.

I would say that they are putting out apparently conflicting information if we were reading Sony's documentation rather than PM's restatement. While I have great respect for him, the quoted text does confuse me a bit but I am reading it out of context.

I think Paul Miller is right, It seems to me both the XA5400ES & DA5400ES convert to LPCM .. Perhaps you can question Sony on their White paper but unless you have some sort of connection with them they probably will not answer. In the UK their communication is dreadful. The people who answer the phone on Sony Support just do not understand the products.

Mebbe. I will reread the docs when I get to the office tomorrow.

Greetings Gentlemen,

This is an interesting thread... SACD and associated DSD signal was originally designed as the digital signal "closest" to analog, and was thought of mostly for archival purposes. By removing the "brickwall" decimation filter from the signal chain, Sony/Philips thought they would also remove the part of the signal path that affected the sound the most, and end the "vinyl vs. CD" war. I believe they have- and I also believe they really were onto something... Please note that Philips had an extensive catalog of very high quality tape recordings in 4 channels from the 70's that a lot of people really wanted put on the market, but there was no distribution format for it. They now can be found in the Pentatone SACD collection, in their original 4 channel format, for your listening pleasure...

Things have changed- most music today is recorded digitally, and as mentioned several times in this thread, the 1 bit signal only allows a very limited amount of signal processing- certainly not enough to allow EQing, reverb, mixing and all other tools required to mix down an album. Sony -also a provider of studios gear- worked on that issue, and after careful listening test came up with the DXD fiormat: 384kHz sampling rate, 24 bits depth, as indistinguishable from the DSD and DSD2 signals (DSD is 1 bit/2.82MHz, DSD2 is 1 bit, 5.64MHz). Most "high end" DSD recording studios today employ DXD during the mixing process, and transform that signal into DSD at the last stage of the mastering, using a sigma-delta modulator (traditionally 4th order, although higher order modulators are also in use). Please see the Sony Oxford consoles for more information...

So in essence, the "downsampling" process from SACD to LPCM can be done transparently, at the condition both bit depth and sampling rate are high enough.

Now, regarding the SCD-XA5400ES SACD player- it does play the DSD stream natively. Both PCM and DSD signal paths are visible on the PCB- I have put both on the scope, and noticed the PCM channel to be quiet (GND) during SACD playback, and vice versa during CD playback. Please note that the Sony is using a PCM1795 DAC from Burr Brown that supports both signals on separate inputs. That DAC is the classic "current steering" topology, very dear to the Burr Brown design team; although there are better spec'ed DACs on the BB offering, I find their latest to have improved definition, and a more controlled slam (The 1792/1794 have a bit of an overemphasized slam to my taste). Those DACs are "multibit DACs" in the sense that they will turn the PCM OR the DSD signal into a 5 bits signal using a digital Low Pass Filter- this is done in order to minimize the impact of jitter, very detrimental to traditional 1 bits DACs (both MASH and straight 1 bit DACs are completely obsolete today because of their high sensitivity to timing errors). All modern signa-delta modulators today employ this technique for jitter reduction; however, the BB DACs employ a current steering technique in the DAC itself, still sensitive to jitter compared to switched-capacitors designs. Additionally, they display a high level of out-of-band noise that needs to be filtered out (3rd or 4th order LPF is required to end-up with a flat noise floor). All those draw-backs do certainly not reduce their sound quality- but Burr Brown DACs are sensitive to jitter, and need stiff analog low-pass filtering. Period. The SCD-XA5400ES designers were well aware of this fact, and designed a very clean clock section- it can be seen close to the DAC in the upper right section of the PCB: locally regulated supply, double buffering local oscillator. They also built a 3rd order, multiple feedback LPF in the output section using "orange drop" polypropylene capacitors in the output stage (not Wima- but what is the difference, really ). Simple and very effective- and likely one of the many details that make this player a winner without spending millions.

Now, when this player is connected to the 5400ES receiver and you are playing music, what really happens? Well, the receiver proposes many digital signal processing functions that cannot be performed on the DSD signal itself, so although the connection from the player to the receiver operates in the "native" format (DSD for SACD both stereo and multichannel, 16 bits / 44.1kHz for CD), it will turn this signal into a workable payload- and all points this to be 4*Fs... As they are using 32 bits DSP, the resolution is maintained above 24 bits at all times (32 bits are never maintained because of rounding errors). Not as good as DXD, but close! So close that it will probably make no difference to the end user. They could have used 8* instead of 4*, but this at the cost of doubling the MIPS requirements for the same processes, which is likely not worth it (double the number of DSPs, etc). The signal is then upsampled to 384kHz at the end of the signal chain to feed their digital input, class D amplifier stages, or fed directly to the DACs at 4*Fs for the analog outputs.

All in all, I find those compromises very acceptable. The DSD format is really a great storage and transport medium- but is never, I repeat never directly transformed to analog without being turned into a multibit signal- even when DACs supporting native DSD inputs are used, because all of those DACs without exception are multibit DACs- this is also true for Meitner or DCs offering (the RING DAC is a 5 bits, multibit DAC). The whole point is to keep the PCM resolution "high and deep"...

Cheers!

Please note that I mistakenly wrote that the SCD-XA5400ES was sporting the PCM1795... It in fact sports the 1796. Those two DACs are very similar, with the exception that the 1795 supports 32 bits input vs. 24 bits for the 1796. The difference is really non-existent as both resolutions largely exceed the resolution of the DAC itself- yet another TI marketing move - except for the fact that with 32 bits coefficients, the 1795 filters match the Matlab (floating point) filter design better, and display less rounding error. As one can see from the datasheet though, measurable performance is identical, and all the math I know would indicate that it really doesn't matter a bit. It would have been nice to see more filter options for the 95 as it is a very recent release, but alas, the filters are the same exact ones used in the 96... Lazy

Cyril Martin seems to know his stuff, I wonder who he works for ?

I'm a lazy epicurean... I drink Champagne for a living

Thanks. That's what I was trying to convey.

I am noticing a lot of people with a strong desire to pair the STR-DA5400ES with the SCD-XA- although it may not make a world of a difference from the technical stand-point. Yes, HATS makes sense because it removes clock dependencies between digital sources and playback, but the STR-DA5400ES is perfectly able to synchronize or re-synchronized (I haven't had time to dig) audio signals. I would suspect the results to be dangerously close with any player supporting SACD streaming over HDMI- has anybody tried comparing the SCD-XA5400ES to any other HDMI player, avec ou sans HATS?

None of the Esoteric do DSD over HDMi?, I was under the impression they did

DJOel

Other than for the Sony DA5400ES (jury seems to be out on this one too) - who manufactures a product that has both SACD or DVD-A and uses PQLS or HATS for HDMI DSD or other transmission?

I am aware of NONE.

Who said that?

I don't see any in the first page!


DJoel

Latest Pioneer receiver LX82 (in UK) has PQLS & new model Onkyo's from 707 have low jitter DACS. However it appears so far that over HDMI DSD direct or LPCM there is virtually no difference in sound quality.

Disbeliever, when you state "However it appears so far that over HDMI DSD direct or LPCM there is virtually no difference in sound quality"- you are comparing using the STR-DA5400ES, correct?

Quick update for those in the market for a new player.

I recently purchased a Pioneer DV-58AV and an Integra DPS-6.9.
The 58AV is intended for more general use and entertainment to stay with my family room system (a Kuro PDP-111FD Professionally calibrated by UMR) and my SC-07 (also professionall calibrated by UMR) with Polk RT-1000i mains and voice matched Polk center and surrounds in a traditional Multi-channel audio configuration for 4.0/5.0/5.1 listening. The Integra is for my dedicated theater.

The Pioneer was broken in as a demo unit and the Integra is brand new (<30 hours on it) so far here are some general thoughts re SACD via HDMI and DSD direct and DVD-A via HDMI (PCM direct):

1) HDMI control works wonderfully on both (Except the Pioneer HDMI control will turn on the Integra when turning off another HDMI control Pioneer Player in the same chain);
2) The menus and displays (especially with SACD - using track and other listings) are FAR FAR superior on the Integra DPS-6.9;
3) The Integra is a bit faster and produces finer menu output (again well superior to the Pioneer's dated and old school menus);
4) The remote is better on the Pioneer and more comfortable to hold. The layout of the Integra remote is confusing and awakward - it's a bit too thin and navigation keys are not all equally sized.
4) For DVD-A the Pioneer is audiably superior:
The Pioneer produces a more live and fuller mids and highs from DVD-A (very accurate in my opinion), while the Integra by comparison is muddier, with tamed highs and a less defined midrange;
5) For SACD - the Integra (Despite its substantially lower specs for the same internal DACS (not used in this configuration) is audiably superior. They are close but the Integra - even not broken in yet - seems richer and more dynamic particularly in the dynamic range offered even with SACD.

With both comming across the same cables via HDMI to the SC-07 (with its Wolfson DAC configuration with some form of Jitter reduction over HDMI with its sampling rate converter that does NOT do PQLS for either unit above) the results are pretty impressive.

Overall (so far):
So far if you are a DVD-Audio fan and have some SACD you like - the Pioneer is the one to consider at this sub $500 price point.
If, however, you love your SACD and have only very few DVD-A discs - the Integra is the one to get.

I have many of both so my jury is out. Off to play some more SACDs and compare before a final decision.

Test samples (my tastes are far ranging) so far include

1) DVD-A: How the West Was Won - Led Zeplin 48/24 recording - Stairway to Heaven (live concert recording from the 70's in LA). The Pionner is dramatically better with this - the guitars and tube amps are viscerally portrayed through it as is the tennor of the crowd and the sung voice. The Integra by contrast seems muted and very laid back, loosing much of the midrange dynamics especially in the guitar and strung distortion chosen for same. Not having heard this on the Pionneer first I would have been very dissapointed with the recording quality of the DVD-A;

2) SACD (Hybrid): Pentatone RQR SACD Classical Sampler - Beethoven Piano concerto No. 4 in G Op. 58
* I can see why Kal likes them so... they have a recording quality that is truly astonishing

3) SACD: Let's Dance - David Bowie - not such an impressive recording but the reproduction difference is still present in minor forms - particularly with the reproduction of the band instruments in favor of the Integra being more open overall and more natural. The Pioneer seems less foward and warmer with a less dynamic reproduction.

Cyril, YES with the Sony combination,I have also used ONKY0 706 which outputs DSD to analogue without conversion to LPCM in Pure Direct Mode . However I prefer the sound from the Sony DA5400ES (Despite being Class D) that does convert to LPCM.(must be the lower jitter) Sony states it gives full resolution from 32 bits but in fact as you have posted it does not and also confirmed by Tim-Vine Lott Air Studios (Pioneer Consultant) who told me that Sony's statement in reply to my question is blatantly incorrect.

I own a Classe SSP-800 pre-pro and was looking to buy a new multichannel SACD player as my current SACD player (Lexicon RT-10) is now somewhat outdated. As the Sony ES 5400 sends DSD only through HDMI, and the Classe does not decode a DSD data stream directly (as I understand it), are the two units incompatible?

The Sony player will output PCM from SACD over HDMI. I did one test with an early SSP-800 and they failed to communicate but the system was at a local shop and I cannot be certain of the actual setup.

Subsequently, I was successful in getting the XA5400ES to communicate (SACD - PCM) via HDMI to an ARCAM AV888 and the Meridian HD621/861.

Since there are no switches or menus to select PCM output, it would seem that the EDID handshaking is at the core of the issue. I would be surprised if the current SSP-800s would not work with it but, as of this moment, I have no such evidence.

Out of curiosity when you say Universal do you mean it plays dvd-a as well.

Thanks

Djoel