Upsampling DACS: Take the Pepsi Challenge

HAs anyone used 2 of the following 3 relatively inexpensive upsampling DACs: Perpetual technologies, Bel Canto, MSB Link 3 with upsampling upgrade?? I am trying to sort out the details of the new technologies. The Perp Tech can "interpolate", while the others do not. I am under the impression that the "24 bit" part of this new technology has to do with s/n ratios aroung 140 db, which is great, but a little useless considering the other equipment in the system. The sampling freq is the part that has me all aflutter, because it seems to be getting closer to analog quality "infinite sampling" if you will... What do you think? Has anyone compared these dacs?? Thanks, gang.

Showing 13 responses by greysquirrel

I see that some posts were deleted from this thread as well as the "What the heck is Res Audio" thread. Did someone cry to mommy?
Thanks Onhwy61. I'm not ashamed to admit that I'm a little confused about this subject. The two processes appear VERY similar to me and yet there is quite a stir about upsampling DACs.
Perhaps someone on this thread can clear something up for me? I asked the question before on Audiogon: what are the differences between upsampling and oversampling? Carl, you seem to be one of many fans of the improvement wrought by upsampling. Can you explain the difference or are there differences? My understanding is that upsampling and oversampling are basically the same. By upsampling/oversampling the digital filtering can be more aggressive (outside the audio range), leaving only a gentle analog filter before output. Is this basically correct? Upsampling/Oversampling can't actually create information; the process can only allow more accurate retrieval of what is contained in the 16/44.1 signal. Right? As for oversampling, all delta/sigma (1 bit) type DACS must use oversampling. Are the new 24/96 DACS mostly delta/sigma types or are they ladder DACS? Don't mean to divert attention from the thread topic, but the recent attention to upsampling has me wondering if my understanding of the process is correct. Did mfg's just get better at implementing oversampling techniques to get better sound and needed new marketing jargon to draw interest? Thanks!
Craig, yes dither (white noise) is added to increase resolution. Not sure exactly how that works, but I believe it elevates the recorded data up from the near the least significant bit, reducing quantization errors. Interpolation is basically what is done whenever you increase the sampling rate by some multiple. I don't think that early "oversampling" designs were adding dither before decoding? But if this is the only difference, then my last statement above is probably right: mfg's found better ways to implement oversampling techniques and are marketing them as the "upsampling" process? Either way, it sounds like most listeners have found well done upsampling/oversampling to be a significant improvement.
Megasam, just went out to the local Borders and read the article in "HiFi News and Record Review". Thanks for the notice. Not sure what they were trying to suggest? They only compared the sound of two different units. I would have liked to hear about the differences wrought by sending the Wadia an upsampled data stream via Purcel or Delius (can't keep straight which is the "upsampler"). After reading that article, I'm still left with the same feeling: oversampling and upsampling are VERY similar as far as the processes go. Implementation (how much over/up sampling and amount of dither added) and choice of filter algorithm is likely to be most significant to overall sound. How do you "resample" a digital datastream? You oversample and interpolate the points between original sampling points. You may increase word length also, but this does nothing unless you add dither. I'm not an expert on this, but I do think I'm not far off. I've e-mailed several mfg's with specific questions regarding upsampling/oversampling. Haven't heard back yet, but if it turns out I'm way off base here, I'll report it here and eat my helping of crow. Just want to understand. Greysquirrel
Perhaps I need to ask an even more elementary question to clear my confusion: with an upsampled datastream, what "information" is populated into the additional 16,711,680 word states (16 bit to 24 bit) and what information is packed in between the sampling points (44.1 to 96,192,706,768 kHz)? MSB has returned my e-mail and confirmed that the improvements brought about by upsampling (at least as they implement it) are primarily due to the ability to filter more aggressively outside the audio band, as well as the ability to utilize the greater linearity of the 24 bit DACS they use. Haven't heard back yet from Resolution Audio or Wadia yet, but will post their response when they do.
The following is the actual e-mail (in entirety) sent to me (this morning) by Resolution Audio in response to my question concerning upsampling / oversampling (the question was almost verbatim to my original question above). Indeed, there is no technical difference between upsampling and oversampling. The only difference I can discern is in the marketing. Indeed, digital filters can be very aggressive above the audio band without the adverse effects that analog brick-wall filters have. This is possible because of FIR (finite-impulse response) filters, which have constant group delay (zero phase effect vs. frequency). There is no physical realization of an FIR filter in analog. Using FIR digital filters allows the analog filter to be relaxed significantly, because the first "images" are located at much higher frequencies. In our cd55, we use a passive third order filter which is down only 0.2 dB at 20 kHz, yet the rejection of the images at 700 kHz is about 60 dB. And indeed, the digital filters do not create information that may have existed before the mic feed was converted to digital. Some external "upsamplers" may by their nature apply some other filter/eq, but this is independent of the a/d - d/a process. You are also correct regarding the delta-sigma dacs. These dacs are rated for maximum input rate, currently as high as 192 kHz. These converters all run at the output at much higher rates -- typically 12 MHz or thereabouts. The better ones from Analog Devices use extra filter stages when the input rate is lower. Essentially, the dacs run, say, 256x at 44.1 or 48 kHz, 128x at 88.2 or 96, and 64x at 176.4 or 192. This puts the noise modulator heart of the converter at the same frequency regardless of input. The best multi-bits, including the PCM1704, run upwards of 800 kHz, which allows 16x at 44.1 (and 8x at 96, and 4x at 192 input rates). In sum, your perception of "market jargon to draw interest" is dead-on. In addition to preying on the consumer base which generally does not have engineering degrees (and some manufacturers as well), these products offer the opportunity to sneak in digital eqs which will absolutely sound different. Better? That's a different story. Finally, we have just started talking to a dealer in Indiana. If all goes well, I'll pass along the info in a couple of days. Regards, Jeff Kalt Resolution Audio [email protected]
Carl, it seems that Resolution Audio wasn't frustrated by my question. Why must you?
After reading Carl's recent post in "What the heck is Resolution Audio" I reached the point (that I'm sure all of you have at one point or another) that I have to call a spade a spade so to speak. Carl is full of crap!! At least on the upsampling/oversampling topic he is. He may be even more than that! I noticed he took this topic over to the "What the heck is Resolution Audio" thread and started spouting about things (upsampling) he obviously does not understand. I tried to be gracious on this thread when his hero Jeff Kalt of Resolution audio confirmed to me by e-mail (posted above) that there is NO difference between upsampling and oversampling AND that mfg's were just using the new terminology to draw interest. However, I lost it a little on my recent post on "What the heck is Resolution Audio". Sorry guys (and girls) if you feel I'm polluting the thread with a little anger, but someone has to call him out on this one! Live the good life. Jordan
Carl, a couple of points for your "gray area": 1.) oversampling/upsampling without "interpolation" is meaningless!!! Interpolation is simply selecting the best bit value to assign to an oversampled data point. You could linearly "interpolate" between (real/original) data points or you could fit some base curve between those data points. If you increase the sample rate, you MUST interpolate! 2.) The ONLY difference Kevin Halverson mentioned was semantic (and I'm sure he meant it as such)!! "Upsampling" units started as stand alone oversampling digital filters to improve the performance of lesser processors. There is NO difference!! The term "upsampling" caught on, in part, due to ignorant journalists, and now everybody is using the word to claim some magic improvement. Also, both external oversampling digital filters ("Upsamplers") and internal oversampling digital filters are upstream of the "DAC". A digital processor and the DAC(s) are quite different. A digital processor has 4 main sections: Input receiver (reclocking), oversampling digital filter, DAC (digital-to-analog convertor), and analog output. So, we have once again come full circle: upsampling and oversampling are the SAME THING!! SYNONYMS!! Ask Gmkane for further details. Jordan
Carl, I did let "those who design them speak for themselves" by seeking out an answer on this topic from one of the leaders in digital design and then posting Jeff Kalt's e-mail reply here. Did you search for answers OR did you already know it all?!! I humbly entered this thread looking for answers to my questions regarding upsampling/oversampling. You, however, replied with your usual collection of technical terms (organized in nonsensical ways) AND got hostile to boot! Carl, nobody likes a hostile blowhard. From my limited understanding of how a digital processor works, I suspected that upsampling was the same, or nearly the same, as oversampling. After seriously considering your nonsense, I asked Jeff Kalt for an answer. I chose to ask Resolution Audio for two reasons: 1.) Resolution Audio now markets an "Upsampling" processor. 2.) I could kick myself for it now, but I actually thought that MAYBE you knew what you were talking about and just couldn't express it in a way I understood. Since you spoke so highly of Jeff Kalt's talents as a digital designer, I thought, perhaps, he could express what you could not. As we all know now, you didn't have a clue what you were talking about and upsampling and oversampling are the same thing! Oh, I performed your test. Heard a little golf ball in the top octaves and about a 2 footer in the upper bass. I don't know, I prefer listening to music (Miles Davis sounds good today), but to each his own. Jordan
Craig, I feel the same way. Some of the journalists (it seems) either did not know enough to be speaking on the subject or were intentionally misleading. It's a shame. Carl, most of us will readily admit when we make a mistake; however, some find their personal egos are more important than truth and understanding. How about you?
Very well said, Joe!! I apologize for my contribution to the more distasteful parts of this thread and will try to raise my bar. On the subject of journalists, I wasn't speaking about John Atkinson or "J-10" in my post above. Actually, the specific article that I was referring to was the Wadia/dCs comparison in "Hi-Fi News and Record Review". I do think it was irresponsible of them to compare two different processors and try to attribute the differences to "upsampling vs. oversampling", particularly since they are two words for the same thing AND everything else about the two units was also different. That just shows that they did not understand how to set up an experiment to test for differences (eg. compare same processor and change only the oversampling digital filter, for example). Anyway, point well taken, Joe!! Live the good life! Jordan