Spider124 Posted July 18, 2001 Share Posted July 18, 2001 Randy, In the DAC prototype I am currently using, a clock (essentially idential to the Audiocom Superclock, measured performance is about <10ps jitter) is driven DIRECTLY to the bit clock of the PCM1702. There is not a single logic chip or anything in between. That is, plain and simple, the jitter the DAC sees. I'm not really ready to go into excessive detail in how the signal is buffered/isolated/converted in the DAC, but the Computer->DAC communication is completely asynchronous and has no effect whatsoever on the timing of the clock. I'd really have to see the thread to know what exactly I said and under what circumstances. Also, I truly did not believe really good playback was possible from a PC for awhile until I came across this method. It is not an exceptionally new idea, but this is the first implementation I know of, and there are no readily-available commercial designs. (All the firewire stuff uses this sort of asynchronous communication with the actual clock's located on the receiving side, instead of vice-versa. But none of the firewire stuff is out yet.) As I said, I'm looking to convert the design to firewire and upgrade the DACs & analog stages, but ultimately I am losing interest, instead looking towards SACD. Charlie ------------------ "What's that noise?" "It's the carpet, it kinda mutes the speakers." "No, it sounded more like the chandalier falling." Quote Link to comment Share on other sites More sharing options...
Prana-Bindu Posted July 20, 2001 Author Share Posted July 20, 2001 Soundog, you must be woolgathering: I believe you were actually referring to the Perpetual Technologies P-1A and P-3A Digital Correction Engine and DAC (respectively). I am under the distinct impression that the P-1A does not accept an optical signal; I believe the P-3A accepts a toslink cable, but not the P-1A (I wonder why... especially in light of the fact that my cd-player only has an optical out). Nevertheless, I looked into the PT products and the P-1A is looking more and more appealing. All very positive reviews and every problems reported was resolved (ultimately) with PT warranty commitment. As soundog indicated, the P-1A upsamples to 96khz (192 when available) and extends the 16-bit samples to 24-bit samples (not by only using dither -- some is used) by applying a software algorithm which interpolates calculated bits between and within the 24-bit versions of the 16-bit words read off the cd. The calculations that determine the value of a new bit or new 24-bit word are, allegedly, both strictly geometrical AND guided by certain psychoacoustic principles. What intrigues me are the formulae that express these psychoacoustic principles in the interpolation functions. Are they simple? They HAVE to be mathematical (it's a computer using 32-bit floating-point processing). Apparently, these formulae also use the information provied in the five samples prior to and the five samples following the sample being interpolated. The literature did not address whether this "window" of 10 samples overlooks already upsampled samples or whether it sees only the 16-bit samples as found on the CD. Anyway, I wonder how important these psychoacoustic principles are to the effectiveness off the algorithm. Do the formulae they inspire present a statistically relevant part of the total algorithm? Will I be able to hear and distinguish thereby the differences between upsampling with just the geometric aspect of the interpolation at work and upsampling with both the geometric and psychoacoustic functions at work? I'm determined to find out. Spider: CD's have the market in a stranglehold that will not let go for quite some time (maybe) And, besides, do you really want to see all those cd's just collect dust when you start building your SACD or DVD-A or G.I.-Joe-Kung-Fu-Grip-CD collection (whoever wins the fight)? Have you considered upsampling and interpolating new data into your digital signal with your computer? I really wonder how important those psychoacoustic principles are to what the P-1A does.... After all, doesn't Shannon's Theory state that the complete and original analog signal can indeed be derived from its quatisized digital derivative? I doubt that Shannon's theory refers at all to a map of the relevant cortex or to whether we enjoy one kind of expression of a particular frequency more than another. First of all, can a computer upsample a digital signal to 95khz? Or is that definitely a hardware-specific function? Why would it have to be? This is where I need to hit the books.... One more question I was hoping y'all would help with before I go grumble about this on my own: how is the dynamic position within the range of a particular format (CD, DVD-A, SACD, etc.) stored on the CD, DVD-A, or SACD? In other words, how does the DAC play a particular sample at 73db in the range vs. 23db? Is the dynamic extension of a particular sample derived from information within the sample? Thanks for the help, y'all. I'm fixin' to figure this digital nonsense out.... ------------------ May the bridges we burn light our way.... Quote Link to comment Share on other sites More sharing options...
Spider124 Posted July 20, 2001 Share Posted July 20, 2001 Prana, you can upsample/oversample any sample rate to any other sample rate with a computer. The problem is that upsampling w/o application of a filter (oversampling), doesn't really give you any advantage (except perhaps for the sin(x)/x rolloff of a multibit DAC, but that can be compensated for w/o upsampling). However, the problem is that oversampling with a filter sounds pretty awful. This accounts for 50% of the nasty sound of digital, I'm convinced. (The other half is poor analog stages, especially opamps.) For playing back any PCM audio (From redbook to 192khz DVDA), the absolute best-sounding system will always be a non-oversampling multibit DAC. Period. End of Story. If I rebuild this DAC instead of moving to SACD (which is tempting), it is very likely I will keep my digital front-end (obviously) and change the DAC's & Analog stages to those of the Adagio DAC by Thorsten. Charlie ------------------ "What's that noise?" "It's the carpet, it kinda mutes the speakers." "No, it sounded more like the chandalier falling." Quote Link to comment Share on other sites More sharing options...
discorules Posted July 20, 2001 Share Posted July 20, 2001 quote: Originally posted by Spider124: Prana, you can upsample/oversample any sample rate to any other sample rate with a computer. The problem is that upsampling w/o application of a filter (oversampling), doesn't really give you any advantage (except perhaps for the sin(x)/x rolloff of a multibit DAC, but that can be compensated for w/o upsampling). However, the problem is that oversampling with a filter sounds pretty awful. This accounts for 50% of the nasty sound of digital, I'm convinced. (The other half is poor analog stages, especially opamps.) For playing back any PCM audio (From redbook to 192khz DVDA), the absolute best-sounding system will always be a non-oversampling multibit DAC. Period. End of Story. ... Charlie Hold on there. Go to the Philips website for a good technical overview on this topic, in particular multibit vs single bit. Bothe have inherent advantages and disadvantages. Like so many things, its the execution of either methods that counts. Quote Link to comment Share on other sites More sharing options...
sunnysal Posted July 20, 2001 Share Posted July 20, 2001 I am quite satisfied with my two sony cd changers hooked up to my MSB linkdac III...I have tried one or two standalone players (a nakamichi and a sony ES) and prefer the sound of the changers through the outboard dac...so, get comfortable and enjoy all your cds from your couch (I currently have 500 available with the touch of a finger)...regards, tony Quote Link to comment Share on other sites More sharing options...
Spider124 Posted July 20, 2001 Share Posted July 20, 2001 disco, yes, I know that each has advantages and disadvantages. However, I am not talking about specs or performance as measured by Hewlett & Packard, I'm talking about sound. It MAY be possible to build a single-bit DAC that sounds as good as multibit DAC's. There were a couple single-bit DAC chips that sound very very good, but are so old and hard to find that I have only gotten to hear one or two, and never in ideal setups. Here I am making a blantantly wide-reaching statement, but it is absolutely true: ALL new 1-bit DACs sound absolutely awful compared to the good multibits (BB PCM63 and Phillips TDA1541, especially) and the good old 1-bit's. There has been a growing trend to throw analog stages and other junk into the DAC chip, and the result is absolutely dreadfull. IF someone made a 1-bit DAC that handled the newer formats w/o all the nasty analog stages on-chip, I would certainly try it out. Charlie ------------------ "What's that noise?" "It's the carpet, it kinda mutes the speakers." "No, it sounded more like the chandalier falling." Quote Link to comment Share on other sites More sharing options...
Randy Bey Posted July 21, 2001 Share Posted July 21, 2001 prana - back on topic, One thing that Tony has that is worth mentioning is an anti-jitter device. To address your changer to DAC quality issue, I would go with an anti-jitter device, which means buying a second digital cable (and the anti-jitter box, of course). I had moderately good results with a Monarchy DIP, but I hear a Jisco may be better, although twice the price. Quote Link to comment Share on other sites More sharing options...
sunnysal Posted July 21, 2001 Share Posted July 21, 2001 that's right randy, I had forgotten to mention the jitter reducer (in this case an MSB director) but I certainly think the jitter reducer and outboard DAC are the two principal contributors to my CD sound...while the transports may be second tier I have obviously opted for "the world at my fingers"...regards, tony Quote Link to comment Share on other sites More sharing options...
soundog Posted July 21, 2001 Share Posted July 21, 2001 Prana- You're right ... it is Perpetual Technologies not Peripheral Technologies ....I've been kinda burned out recently due to fatigue from staying up too late listening to my CDs through these unbeleivable devices ... there is just so much more on my cds then I ever was able to hear. You're right about the P1A .. it does not accept optical (toslink) input like the P3A ...that's why I have the MSB Digital Director ...it takes muliple toslink and coax inputs and outputs them as both toslink or optical. Meanwhile my DVDs are getting completely ignored, I am so fascinated by these new technologies and the amazing sound they produce ... interestlngly, at least to me, is that I am feeding their output through a recent incarnation of an old Dynaco tube pre-amp design into over twenty year old Klipschorns (but updated with Al's great ALK crossover. I get a kick out of technology from my youth meeting and combining with the technologies of today. The Perpetual Technologies web site has some great explainations of how these devices work. I feel like I now have the best of the old LP/tube/horn tehnology combined with modern computer tech. resulting in all the advantages of CDs and LPs without the disadvantages of either!!! Quote Link to comment Share on other sites More sharing options...
Prana-Bindu Posted July 23, 2001 Author Share Posted July 23, 2001 Soundog: Have you tried the P-1A's speaker correction software? Are your speakers supported yet? I'd really like to try putting one of those things together on my computer, especially since I really don't think "psychoacoustic principles" would make that much of a difference in the interpolation algorithm. The units being interpolated are so small, that a geometric approach seems more than adequate. I found an interpolating application put together by some UC Berkley people. It uses fixed-point processing, and the P-1A is touted to use floating-point processing, but I'm not sure what difference that would make, if any. Time to learn UNIX.... Spider seems to be telling us that the biggest problem of such a project is to eliminate digital jitter. So, Spider, how did you manage to get jitter down so low on your PC's output to your DAC? Is your software that separates that data from the clock information into a I2S signal proprietary? You mentioned that you use FLAC compression prior to conversion. Does that mean that you have to have a full file for your application to convert to I2S, or can you pipe the info in and out as the CD transport plays? I'd hate to have to encode all my music to do this.... There can't be THAT much of a benefit to encoding it, though: isn't the highest level of compression available in a lossless format about 2:1? What degree of compression does FLAC offer? Maybe it would just be easier to purchase the P-1A.... The room correction software will be available in a few months, but a salesdude I e-mailed responded that the $700 add-on would give you three room calculations. That seems a bit greedy. Why not have the room software available for unlimited use? $700 is a pretty big premium for a software license.... The speaker correction software is intriguing, but they currently support a limited number of speakers. Thanks for any help y'all can break off!! ------------------ May the bridges we burn light our way.... Quote Link to comment Share on other sites More sharing options...
Spider124 Posted July 24, 2001 Share Posted July 24, 2001 You should just buy whatever DAC you think sounds good As I said, most of them have jitter reduction circuitry, and I'm sure it will sound great. The parts in my long letter on digital audio discussing what works best didn't neccesarily mean you HAD to have one of those methods As for my setup, the data is converted from simple PCM format to I2S IN THE DAC, by a special chip I designed specifically to do that. The PCM data is buffered in the DAC before being converted to I2S, so the timing of the PC's transmissions to the DAC are irrelevant, and have no effect on jitter. The audio does not HAVE to be compressed beforehand, I just chose to do so. You could just as easily stick a CD in there and rip it in realtime. Charlie ------------------ "What's that noise?" "It's the carpet, it kinda mutes the speakers." "No, it sounded more like the chandalier falling." Quote Link to comment Share on other sites More sharing options...
soundog Posted July 24, 2001 Share Posted July 24, 2001 Prana- The speaker correction software is not yet available for Klipschorns and I may have to use the kit they send to measure the output of unlisted, old or modified speakers. That kit as well as the room correction option is not yet available. Quote Link to comment Share on other sites More sharing options...
maxg Posted July 25, 2001 Share Posted July 25, 2001 Well I think I understood the question, but I certainly did not understand most fo the responses. This whole digital conversion to analog thing is so complex (jitter, bit-rates, lasers, caches, buffers) I have given up entirely trying to get to grips with it and decided to take a radical step and let my earsw decide. I was fortunate enough recently to hear a top of the line Accuphase SACD player and separate DAC (model 100 and 101). This is probably close to the king of the line in digital players (certainly judging by the $28,000 asking price). For the demo I heard it playing an XRCD (Famous sound of the three blind mice) which I provided as I do not have any SACD experience to compare it to. Yes - surprisingly (lol) it did sound better than my Marantz CD6000 (£250) but not, in my opinion $27,750 better. The setup was not all that dissimilar to my own in that it had an Accuphase amp (E307) and silver speaker cables going to a pair of Heresy's. Then, just for interest, we compared it to a ClearAudio record deck with its top of the range arm and cartidge through the same amp with the same interconnects. To be honest the recording was not the same as the TBM one we had used for the CD test but it was a nice Jazz recording of similar musical dimensions. To no-one's surprise the Clearaudio was "clearly" better. What was a surprise was how much. Every element of the music we could identify was massively better, the highs, mids, bass, staging, tonality, balance everything. If I scored (arbitrarily) the Clearaudio at 100% then I would estimate the Accuphase at something around 88-90% with the Marrantz at 85-87%. With the price of the Clearaudio setup at a mere $12,000 the Accuphase makes it look like a bargain. I wonder if my wife will see it like that? On the other hand I would have to say that I am amazed at how small the sound difference was with the Marantz, and, if I am going to be honest that represents the best bargain of all in terms of music/$ ratio. But my God the sound of that Clearaudio is tempting... Quote Link to comment Share on other sites More sharing options...
Prana-Bindu Posted July 25, 2001 Author Share Posted July 25, 2001 Maxg: That is truly disheartening. This digital nonsense is a big enough hassle.... Anyway, I was always under the impression that vinyl was on a quick and certain path towards obsolescence. What will vinyl be able to offer that a recording made at 24-bit/96(or 192)khz and played back with proper jitter correction will offer? Everything I've read indicates that that level of digital resolution is not distinguishable by the human ear from the original analog waveform. Or is it...? ------------------ May the bridges we burn light our way.... Quote Link to comment Share on other sites More sharing options...
discorules Posted July 25, 2001 Share Posted July 25, 2001 quote: Originally posted by Prana-Bindu: Maxg: That is truly disheartening. This digital nonsense is a big enough hassle.... Anyway, I was always under the impression that vinyl was on a quick and certain path towards obsolescence. What will vinyl be able to offer that a recording made at 24-bit/96(or 192)khz and played back with proper jitter correction will offer? Everything I've read indicates that that level of digital resolution is not distinguishable by the human ear from the original analog waveform. Or is it...? In fairness to the new format(s), Maxg did not listen to a actual SACD disc on the SACD player. He used a XRCD which do sound good but may not be comparable to SACD. Having said that, I can appreciate what Maxg was saying about analog playback. It still sounds more musically complete than cds. Just the other night I listened to the vinyl of Copacabana by Barry Manillow. I also have a Micro-Acoustics 2002 cartridge on a Kenwood direct drive turntable. The music had a real quality that I still have heard matched by digital. Quote Link to comment Share on other sites More sharing options...
soundog Posted July 25, 2001 Share Posted July 25, 2001 The cost of my Perpetual Technologies gear that makes my CDs and Sony changers sound so great is $1099. for the P1A and $799. for the P3A. Nowhere near $24,000 or even a mere $12,000. And all I do is point and click! Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.