High resolution digital is dead. The best DAC's killed it.
One was the giving-up of R2R DACs. This decoding scheme, despite it’s (potential) errors in switching, is still the best way to convert D to A. High-end makers have been returning to this since 2011 and the results are very welcome. I won't debate the relative merits of R2R DACs, everything else being equal, but I will say that I've heard a number of non-R2R DACs which show the improvements I'm talking about. I don't think my central thesis, can be explained this way alone. Best, E |
| Post removed |
Ever since the beginning of (consumer) digital audio, there has been an ongoing debate over DAC's, analog vs. digital, etc. The original design intent for CDs has, over the last several years, been almost realized - a quantum leap over the earlier versions of gear designed to retrieve those elusive bits. One thing that I find has been consistently overlooked is the actual data retrieval itself, BEFORE it gets translated via DAC (or even digital cable leading to the DAC). I've encountered many products purporting to be the holy grail of CD restoration/ improvement/ enhancement, but one product has stood out: a CD treatment called Essence of Music. It is a two step process that, in a nutshell, allows the laser to pull out more of the music. This leads to a more accurate reading of the data (improved source quality). I am not an engineer or a scientist, but I do have more than 40 years' experience in audio, and I am stunned by the difference between treated and untreated discs...and so are other audiophiles when they hear it. Over the past few years, I have gone to Rocky Mountain Audio Fests with a huge book of CDs...and I deliberately had two identical discs, purchased from Wal-Mart, Target, or even at the shows themselves. I present an untreated version of the CD, and after 30 - 45 seconds, I ask the exhibitor to exchange the CD, playing the same selection at the same volume. The looks on everyone's faces (including the exhibitors) tells the whole story. Through the years, I have tried other products, and even done the 'green marker' thingie, in an attempt to wrest the last bit of music from those silver discs, but this product is the ONLY one that has consistently delivered the goods, year after year. It is not cheap (what is, in audio these days?), but the difference in sound quality is comparable to upgrading your cables, or amp...you get the idea. I have even purchased 5 or more copies of the same disc, and found that more treatments yielded even better results. The change is apparent even on "low-" or "mid-" fi systems - even playing a CD on a BLU RAY disc player. Try it out - BEFORE upgrading anything else. You may be pleasantly surprised. |
…But that was a digression from the original topic...but not really. CDs that are read properly will not be as adulterated by the error correction algorithm that is present in all CD players, so red book CDs can sound truly awesome. This was the original intention, and between sound retrieval improvements in the source and the processing, we can see how good red book CDs can sound. Keep in mind, that the operating principle has remained unchanged for decades - but we have better ways of reading, and translating, what has always been there. |
Sadly, Mark went from making absolutely killer $500 DACs to making even better $25000 DACs. I guess the market spoke, and being smart, he listened :-( Its also amazing how much better i made the gold full nelson LinkDAC sound by building a custom (dual), custom regulated box. Many, many hours and about $125.00 in parts on ebay. I cannot hear one whit of difference between 96 and 132k upsampling G |
Lots of interesting points made on a subject destined to run and run. Last year i was privileged to present the AES Heyser Lecture in Milano (May 2018) which included amongst other topics some discussion and personal reflections on sampling rate conversion. You can (hopefully) access the lecture and some related files via the link below including examples of DSD-to-LPCM conversion using a novel "spectral domain" approach to sampling rate conversion. Unfortunately sound quality on the video is a bit compromised due to a technical recording issue beyond my control. Anyway, I hope the content may prove of interest. https://www.dropbox.com/sh/nqvvwnlh6n6b98z/AAC6kOb5NT8Gk8pceZSbpl2ha?dl=0 M |
this ought to be great, thanks Malcolm! Downloaded - would not play in browser. I plan to continue my blog @ sonogyresearch.com with some interesting anecdotes from when i liaised to MPEG while working on both contribution quality digital compressed NTSC and HDTV back around 1989-91. The findings (which can be informally confirmed with any JPEG compression utility and your phone) are very interesting in a world where compression is a dirty word....spoiler alert - compressed HD > uncompressed SD. No surprise if one considers the degree of redundancy in any HD representation. G |
Absolute phase is audible but the majority of listeners are unaware of the phenomenon or have heard it and deemed it insignificant, probably the same ones who can't hear cable differences. Absolute phase is when the leading pressure wave reaches the listener in the same phase as originally produced. Take for example a spoken word like 'push'. The compressive wave launch will reach the listener and drive his eardrums inwards. A reversed phase will create a rarefaction when it reaches the listeners ears and suck them outwards. Think of a kick drum. Incorrect phase then means the poor listener is using more mental correction effort to try and add realism on top of all the other reproduction issues where the brain has to try make allowances for background noise, distortion and little or no room treatment etc. Absolute phase is often subtle and occasionally astonishing. I had my CD collection marked to indicate phase. Lots blank (meaning can't tell a difference) some normal phase and some reverse phase. My Wadia DAC has a phase button which I leave to suit the majority and only occasionally change. @ geoffkait The effect is not system dependant, it is very real and generally impresses when first noticed. What is system dependant is the extent to which it is heard. The battery test you mentioned works to show the direction of woofer travel. The tweeter will not respond via the XO because its series cap will do its job and block DC which is merely academic because the speaker designer, one has to assume, knew what he was doing and designed for proper phase tracking around the XO frequency. It is not necessary to concern yourselves whether the tweeter is in phase or not. Changing both speaker terminals will invert phase to the speaker XO and the tweeter will follow suit. 2,3 or 4- way speakers same story. @ erik_squires Richard Vandersteen's big play is for achieving Time coherence which is accomplished by physical driver alignment and in the XO and is one of the best values around. Its the end result. Tweeters or midrange drivers may be connected electrically out of phase to place them acoustically inphase. Polarity is chosen for best performance at XO Think you mentioned full-range drivers, which operate without any XO as such and have their own special sound but introduce other issues. You mentioned 'non R2R DACs' you were impressed with, care to share. |
eric_squires asked: The modern DAC, the best of them, no longer do this. The Redbook playback is so good high resolution is almost not needed. Anyone else notice this?Not really. I totally agree that as we solve problems in DACs and in the timing of the signal they are fed, and as recording engineers operate with the systems margins, and as studio formats allow for better filtering, and... (key here - lots of "ands"), we are finding that red book reality is approaching red book theory. But are today’s DACs, on their own, that much better? Well, Maybe. I will say that when fed a low-jitter signal, my 20-odd year old Theta DS-Pro (Mike Moffit’s original baby) sounds very, very good. Is it state of the art? No. Is it damn good and better than most modern DACs (which admittedly often cost less?) yes. |
| Post removed |
The Theta Casanova from memory used the same dac chip/s as the Casablanca which had hybrid PCM-1792A converters in it, few bits of R2R and Delta Sigma mix, it was the time of the transition period from expensive to produce R2R converters such as PCM1704 to much cheaper to make Delta Sigma. Shanling had two identical (earlier 4x transformer model) T100 units, but one used 4 x mono PCM1704 R2R dacs, the other used 2 x stereo PCM1792 hybrid dacs. The we did a/b’s with them and the one with 4 x R2R PCM1704 dac’s was clearly the better sounding unit with redbook. Cheers George |
FYI, i had occasion to speak to the guys at Theta about the Casanova recently. I was not terrible familiar. According to them the DSPro is significantly better. Casanova is a cool concept withe the all-digital backplane. I considered something like that back in the 1980s. Never took it anywhere for lots of practical and life reasons... mostly i had another company and another job that each demanded my time. So our experiences may not in fact conflict. |
FYI, I had occasion to speak to the guys at Theta about the Casanova recently. According to them the DSPro is significantly better.That's because all the 8 "Pro's" (except for the Pro Prime which was DS ) were all R2R and for PCM Redbook it's better, just like what we all expereienced with the two Shanling models in my last post. Cheers George |
No, is not that simple.The cassanova card (at least the one i was holding) is also R2R. And all the DSPro's from Gen 1 were R2R. In fact, in 1994 i don't think there were sigma-delta chips yet. One of the big changes to the V and 8 was re-clocking (jitter, here we go again) I was going to avoid the technical details, but it had to do mostly with the DSP power to execute the digital pre-filtering. G |
And all the DSPro’s from Gen 1 were R2RSorry, but the Pro prime was only one of the Pro series to be Delta Sigma it used the DS chip SAA7350GP. The Pro Prime II was not DS it had 2 x PCM67P-K R2R chips In fact, in 1994 i don’t think there were sigma-delta chips yetSorry again, but whatever you want to call it (Delta Sigma, Bitstream, 1 Bit, Mash ect ect ) In 1984, at the Eindhoven AES convention, that Robert Adams of dbx presented a paper on their 1-bit adaptive delta-modulation A/D- and D/A-conversion system. I had "one" of the first DS based chips in 1990 the Technics SL-PS70 No, is not that simple.The cassanova card (at least the one i was holding) is also R2R.Sorry again the manual and other sources states it uses for D/A 24bit Delta Sigma dacs and that they are 94khz compatible. Just as the Casablanca does also. And the A/D section conversion is also Delta Sigma but just 20bit. I had occasion to speak to the guys at Theta about the Casanova recently. According to them the DSPro is significantly better.This is also correct, as the DS Pro’s except for the Pro Prime, are all R2R and not DS. Cheers George |
George: 02-15-2019 3:23pm read what i wrote, DSPro, not Prime, not Basic, DSPro. regardless, you seem to agree that the DSPros were R2R, so what's your point? Maybe I'm missing it...And all the DSPro’s from Gen 1 were R2RSorry, but the Pro prime was only one of the Pro series to be Delta Sigma it used the DS chip SAA7350GP. The Pro Prime II was not DS it had 2 x PCM67P-K R2R chips And the card i was holding had a PCM17XX (i forget) R2R chip on it. Or a BB series that I thought to be R2R (e.g. 17XX, PCMXX) but that was some time back. Now, as to when PDM/sigma-delta was in wide use (not in labs -- bear in mind that the Theta DSPro G2 is from 1993-4), i recall in 1987-1988 i was looking at the BB PCM54 predecessor to the PCM55 and 64 and it -- and the whole series to come -- was *very new*. BB was in its R2R days. Phillips was selling lots of what 1741s or something like that (ancient history!). But R2R was not special - it was simply how most things were done then. I don't know what gen Cassanova i was holding or what variant - it was a card brought to me bare, with a project in mind. We scuttled that idea after speaking with Theta. But regardless, my point was that the Theta DSPro was a significantly better performer to the Casanova i was holding, and according to the engineer i spoke with a Theta, the big differences was horsepower to perform the DSPs. Maybe it was more - i didn't have the time or interest to play twenty questions with someone helping me.... So maybe it was PDM. |
so what’s your point? Take a pill sunshine. I said That the DS PRO PRIME was part of the "DS Pro" series and it was the only one of 9 in that series that was Delta Sigma, the rest all being R2R. I don’t know what gen Cassanova i was holding And that the Casanova was also Delta Sigma, it states it in the manual specs and online, here is the page https://ibb.co/zf4T5Xp You must have been holding on to something else too tight. |
Interesting thread. I found my destination digital with the Aries Cerat Kassandra DAC. No up sampling, no filter, and using 16 x AD1865 NK chips. Huge power supplies, tube out (10v) I/V transformers, and huge output transformers. It weights 67 kilos but soundly beats all others I had at home inc: Hilo interesante. Encontré mi destino digital con el DAC de Aries Cerat Kassandra. Sin muestreo, sin filtro y usando 16 chips AD1865 NK. Enormes fuentes de alimentación, transformadores I / V de salida de tubo (10 v) y enormes transformadores de salida. Pesa 67 kilos, pero supera a todos los demás que tenía en casa: 1. Ch Precision C1 2. Chord DAVE 3. Lampizator Golden Gate 4. Naim CD555 5. Audio Note DAC 5 Signature Check out: https://ariesceratespana.com |
I couldn’t agree more with the OP. The desire for high-res was to reduce the timing artifacts from older DACs. I have the latest Beresford Caiman SEG DAC which is optimised to more accurately convert an SPDIF clock signal. The stability on Redbook is incredible. Standard Redbook choral music (Thalis Scholars) exhibits only natural harmonics and none of the harmonic timing artifacts that I’d previously assumed were on the recording. High-res was a temporary workaround. |
| Post removed |
@gosta All local files and streams including Tidal, etc. are first converted to DSD 512 on my server PC before heading to my Windows 10 streamer and finally through the DAC 8 DSD chipless converter. As long as you have server horsepower, you can even use DSP while doing this, which is fantastic to tame a bass mode or to use as a balance control. The PCM side of the DAC is ok, but once you hear DSD 512 there's no going back. There's more texture, better tone, soundstage expands in every direction, imaging is more precise AND has more density (that is the hardest trick to pull off IMO). It really is something. |
