1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice


Discussion in 'High-end Audio Forum' started by magiccabbage, May 14, 2015.
263 264 265 266 267 268 269 270 271 272
274 275 276 277 278 279 280 281 282 283
  1. Rob Watts
    PC's are very restricted in what they can do for real time signals. You simply can't replicate the processing that Dave does in a PC - simply because PC processors are sequential serial devices with a very limited number of cores. When you are doing a doing a FIR filter (a tap) you need to read from memory the audio data; read from memory the coefficient data; multiply the numbers together;then read the accumulated data and add that to the previous multiplication; then save the result. Lots of things to do in sequence. With an FPGA you can do all of these things in parallel at once, so a single FIR tap can be accomplished within a single clock cycle (obviously pipelined) - you are not forced to do things in sequence.
    With Dave I have 166 dsp cores running, plus FPGA fabric to do a considerable amount of further processing. You simply can't do that in a PC. To give you another example - converting DSD into DoP. You need a quad core processor to do this manipulation in real time - otherwise you get drop-outs - but in a FPGA I could do this simple operation thousands of times over, and at much faster rates than DSD256.
    What some people do not understand is how capable FPGA's are and how widespread they are used - the backbone to the internet? FPGA's. Search engines? FPGA's. Why? because an FPGA is fantastic at doing fixed real time processing - it takes small die area, and can do complex operations with very low power. Mojo for example has 44 dsp cores, uses sophisticated filtering to 104 MHz, and noise shapes at this rate - but does all this whilst consuming only 0.45 W. There is no way any PC consuming huge amounts of power can do this.
    Intel last year acquired Altera (an FPGA company) for $16.7 billion because they understand that the future of processing is with FPGA's
    A second issue is not what you can do but how you can do it - it is not just about raw power, but how the filter algorithm is designed. I have put many thousands of hours and over twenty years improving and understanding how to make a transparent interpolation filter; and I am still learning things today.
    And a third point is that a DAC is not simply a data processing machine but it has got crucial analogue parts too. If I dropped the WTA requirement, I would still need the same FPGA in order to do the noise shaping and other functions.
    Chord Electronics Stay updated on Chord Electronics at their sponsor page on Head-Fi.
    https://www.facebook.com/chordelectronics https://twitter.com/chordaudio http://www.chordelectronics.co.uk/
  2. Christer
    Very interesting to hear,
    "straight from the horse´s mouth."
     I am all ears.
    And always interested to learn as much as possible.
    I have absolutely zero personal  knowledge or expertize with these things, but every reason to believe you  really are maybe the best expert in the industry regarding these things.
    My only relative  strength as an evaluator of SQ in various products lies only in the facts that unlike many others here, I ONLY use acoustic music for reference and in some cases material where I was actually present at the sessions.
    Plus for my age, still pretty decent hearing.
    After all my moaning  about Dave´s pricing,I still have to confess that DAVE was THE BEST DAC I have ever auditioned via headphones.
    No other  digital product I have auditioned or owned has lifted so many  digital veils as DAVE did.
    But judging from your last paragraph above, could the fuller warmer/darker sometimes more timbrally realistic SQ I often hear from my Benchmark DAC2 via speakers especially, be a result of  better analogue parts than  little Hugo has?
    It sure seems that Hugo has a  very clear and obvious  edge over Benchmark in its data processing parts.
    But that doesn´t always seem to translate into better more realistic final SQ, when connected to amp and speakers.
    I do know that  superior analogue parts seems to be exactly what Benchmark  are claiming for their DACs,plus their inclusion of  galvanic isolation in their similar to Hugo priced product.
    I actually paid exactly  the same price for both.
    According to their latest firmware update/upgrade they seem to be doing it in the form of an Xilinx FPGA.
    Truly portable DAVE SQ in a DAC that could also equally well be connected to a HIFI system, now that would be quite something imho.
  3. JaZZ Contributor
    One can always dream (and I certainly wouldn't say no to the idea) – but that won't be realizable with today's state of technology (from what I understand according to Rob's reasonings). Just think of the Mojo's heat development during playback, and that's with much less taps than DAVE.
  4. Sunya
    The HQPlayer has a CPU offload option that permits the use of the CUDA cores in NVIDIA GPUs for real time parallel processing; the new Pascal GPU in Titan X has 3854 CUDA cores.
  5. Hellvis
    Looks amazing!...But that cost....that cost.
  6. Deftone
    i had to skip canjam london this year because i couldnt get the time off work. listening to dave was at the top of my list, maybe next year [​IMG]
  7. Deftone
    id still love to hear it even if ill actually never be able to afford it. well i could but id have to save up for 7 years.
    get in on the lotto? lol
  8. maxh22

    I was going to write this but you beat me too it.

    @Rob Watts wouldn't the the new Titan X with it's 3856 cores be more capable of reproducing the original signal than LX75 fpga with it's 166 cores?

    I know this is like comparing apples to oranges since FPGA and ASIC are completely different but on paper the Titan X should Be the more capable processor.
  9. ecwl
    I think one of the issues is that even with HQPlayer and a multicore super desktop, you can at best upsample/filter to 32-bit/768kHz and then send it to a DAC with the current USB protocols. Whereas the Chord DAVE I believe first upsamples to 768kHz first and then goes to 12.2MHz next. And then the 12.2MHz signal gets noise-shaped and converted to 5-bit/104MHz. So until someone develops a protocol to transmit 5-bit/104MHz to the Pulse Array DAC off HQPlayer, you're still not approaching the Chord DAVE's FPGA capability.
    Similarly, I've seen people say that they can use HQPlayer do upconversion and then noise-shaping in DSD. Well, as far as I know, you can still only send 1-bit/DSD512 = 22.5MHz through the USB protocol. So 1-bit/22.5MHz is still a far cry to 5-bit/104MHz. Not to mention the difference in DAC design between Pulse Array and pure DSD/PWM.
    So as much as I see the potential advantages of using PC's and HQPlayer for more computationally intensive upsampling/digital filtering, I'm not sure if it's comparable yet. I'm guessing there's probably no theoretical reason why the USB protocol can't handle higher data rates for specific DACs. I'm surprised HQPlayer has not moved in this direction yet with a hardware partner...
    Jawed likes this.
  10. Mython Contributor
    LOL - yeah, all very nice, and no doubt immensely powerful with oodles of potential, but rather taking Rob's remarks out of context, don't you think? (both of you) [​IMG]
    Not meaning to put word's in Rob's mouth, but I consider it quite apparent that he was speaking of current, mass-market, released-&-widely-available, home computers; not speaking of bleeding-edge, barely-out-of-prototype, and not widely-available, novel high-end-server GPU chips.
    It wouldn't surprise me if there are also uber-powerful bleeding-edge, barely-out-of-prototype, and not widely-available, high-end FPGA chips in existence, too!
    miketlse likes this.
  11. miketlse
    Oh well in that case, let us buy one of the pascal GPUs for the quoted $10,500, then add the double Xeon motherboard, then the noisy cooling fans and power supply, add a case etc.
    Then spend a few months writing the parallel processing program, and then test everything.
    @Sunya I think you will have persuaded a lot of people that the desktop sized DAVE, which runs silently without any user debugging or testing, is remarkably good value for money. 
  12. Sunya
    Lol, the Pascal based Titan X is $1200; the lower spec GTX 1080 with only 2560 CUDA cores is $700, all readily available. And nothing stops you from using the CUDA cores from the previous Maxwell generation. My point is PCs are perfectly capable of parallel processing if the software is designed for it. The developer of the HQPlayer estimated he would need the Stratix V FPGA from Altera (he's more familiar with Altera than Xilinx) to run his algorithms. The Spartan 6 FPGAs is the cheap low end range from Xilinx BTW.
  13. miketlse
    You linked to an article containing the details for the new Pascal GPUs that are only currently available in very limited quantities, with an estimated price of $10,500.
    How silly of me not to realise that you were referring to a completely different piece of hardware.
  14. Sunya
  15. Sunya
263 264 265 266 267 268 269 270 271 272
274 275 276 277 278 279 280 281 282 283

Share This Page