Jump to content

Sagittarius

  • Posts

    20
  • Joined

  • Last visited

  • Country

    country-ZZ

Retained

  • Member Title
    Newbie

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Thanks Miska and Rovo. Unfortunately, it seems that it is likely a Windows issue .
  2. I would like to know if HQPlayer decides that it will not try to play a file in some modulator/filter combinations based on some measure of available system resources and how this decision is made? For example, I have been trying to explore which filters I can play with my system. When I covert 44.1K PCM to 256x DSD with Sinc-L/ASDM&EC-super 512 +fs, payback is ok. However, when I increase the DSD limit to 512x with the same filter/modulator combination, HQPlayer does not even try. Unlike some filter/modulator combinations where there is a noticeable delay before HQPlayer starts playing a file and I hear the computer fans revving up, in this case the blue bar which shows across Hqplayer window when its setting up a modulator/filter flashes for a fraction of a second and then nothing happens. I have CUDA enabled and 16GB GPUs (both 4060TI and an old Tesla P100 that I managed to get to work). Memory consumption is around 6GB for the 256x case whether I am enabling CUDA on the 4060TI or the Tesla through nVidia Control Panel. So there is enough room to double the memory usage. Other filters play fine with this modulator so it is unlikely that it is a CPU issue (I9-12900k). What does HQPlayer look for when it decides not to play a file ?
  3. I have a couple of questions for @JohnSwenson . I have come across the linked post while reading through this thread and I am curious about two points mentioned in it : First: Leakage Current "is formed by capacitances in the PS between AC line and DC output generating a voltage between AC line and DC output. When a low impedance connection is formed between the DC output of one PS and the DC output of another PS these voltages form a current flow " My first questions is: Would the use of symmetrically balanced power reduce these voltages which cause the leakage currents with one phase counteracting the action of the other phase? Second: "The current itself is not the problem, it is the voltage developed by the current flowing through an impedance. So decreasing the impedance of part of the loop can decrease the voltage resulting in less effect on audio if you do it right. OR you can block the leakage loop altogether." My second question is : Is it it possible to decrease the impedance in another part of the loop other than at the power input of the audio equipment by connecting them to the same power strip? More specifically, would the products designed to connect the signal grounds and/or chassis grounds of the equipment in an audio system to a common point achieve the same effect (e.g. Shunyata Altaira, Puritan Audio Route Master, Synergistic Research Ground block, etc). If this has the same effect, it could open the door to using multiple isolation transformers of different sizes for different equipment in the same system or starting with smaller ultra isolation transformers for the sources and line level equipment if a large enough one is not available right away.
  4. @Miska I am considering buying an older GPU with high FP64 processing capacity to see whether it will help with the more demanding filters if I can get one cheap enough to justify the uncertain outcome of this experiment. I do not remember seeing that route being explored while browsing HQplayer forums. I would like to know if the minimum required CUDA Compute Capability level of HQ Player is going to change from 5.2 in the near future.
  5. Thanks Rajiv, If I remember correctly from another review, you had a clock that synchronized the SOTM and Uptone Audio network switches and a USB interface. Did you remove it after adding the Taiko server?
  6. I would like to know what is the lowest power CPU that can run HQPlayer at DSD 256 upsampling without any restriction on modulator or filter type. I tried to check the posts on the mac mini M1 and, while I understand it will do DSD 256, I am still not sure if it can keep up with all modulator and filter combinations. Also, is it correct that in case of using a MAC, an NAA may be required because MAC can only output DSD over PCM (and many DACs will not accept the sample rates required)? Has anybody tried the intel i7 "T" processors which run at a base frequency of 2.0Ghz for use with HQplayer? They are used in many very compact computers, which is the reason I am interested in them.
  7. I guess this means that we are better off looking for DACs that can do DSD256 without clock dividers rather than waiting for CPU technology to catch up with HQ Player's EC modulators DSD512 processing needs for real time playback. Because if we are lucky enough to get 20% increase in processing power every 2 years it is still going be a long wait (that's hoping for a new architecture every two years that can deliver at least 15% improvement and some minor improvement in between). May be if we get together we can commission T+A or Lampizator to build such a DAC 🙂 I am curious though why the processing speed was only 0.55x. Cache not being able to feed the rest of the processor fast enough? ..other thread dependencies? ..thermal throttling? Something that can be addressed in future versions of HQ player?
  8. First , thanks to Audiophilestyle for this great opportunity. I have a few things that I am curious about: - Where does Bruno think class D is heading and where are the improvements going to come from? There are companies which now use gallium nitride transistors and are pushing the switching speed into the MHz region. Does Bruno think this is the way forward? - Class D has achieved very low levels of distortion, but is it possible for class D amplifiers to continue their evolution into something close to a straight wire with gain, i.e. minimal phase shift in the audio band? - Are you using the same feedback scheme which was developed for ncore in the new amplifiers? - Does the high damping factor of class D amplifiers negatively affect the subjective sound quality of some types of speakers as some reviews seem to suggest? Finally some requests: A few things that I did not like in ncore development and I hope will be better with Eigentakt: - It took so long to get ncore amplifiers with a variety of power ratings, particularly for the lower power parts and when they came they seemed to be positioned below the nc400/nc500 and nc1200 . - Did we really need a halo part with, frankly, an eyebrow raising price (I am looking at the nc1200)? - Please no capacitors with low capacitance stability vs frequency in the output filter in the lower end models. The polyester caps used in the nc400/nc500 apparently acted like capacitors with variable tolerance when faced with the variable switching frequency of the amplifier creating additional errors for the feedback scheme to correct. Some users have replaced those caps with metalized polypropylene caps (space on the board did not accommodate normal polyp. caps) and reported noticeable improvement in sound quality. I think the coming 5 to 10 years will be great for class D amplifiers. There are more companies developing and adopting the technology and Bruno was one of the instrumental figures in achieving a well earned acceptance for switching amplifiers in the audiophile word. I am very excited to see what Purifi has to offer for both amplifiers and speaker drivers.
  9. Interesting. I think this simple calculation sheds some light on the reported performance of various processor designs with the more strenuous HQPlayer algorithms. It makes sense not to waste many of those clock cycles over latency, hence the importance of cache size, or other unrelated processes, which is probably part of the reason, beside larger total cache size, that some processors with a large number of cores are reported to perform ok with HQPlayer's more difficult loads despite their lower clock speeds as they have more free cores to use. One question I have is whether the newer non-inclusive victim L3 cache design that Intel have been using with Skylake X processors is still useful for HQPlayer more demanding loads. In other words, would one be (at least theoretically) be better off using a high core Broadwell generation processor with its larger L3 inclusive cache than going with Skylake X?
  10. Hi, I would like to know if Miska or any HQ Player user have experience they can share about running 512x DSD upsampling on Ryzen, or Threadripper CPUs and whether there are any limitations on the upsampling filters that can be used in such a setup. I am considering building a PC around AMD's latest generation CPU's but I am a bit worried because they are behind Intel in terms of AVX implementation, though these CPUs sometimes seem to do quite well in compute tests.
  11. Quick but relevant question for Miska. Do you think if the upcoming E5 Skylake Xeons have AVX 512 instructions or a subset of the AVX512 instruction set enabled as expected, that would allow HQ player to required lower minimum base frequency requirement? My understanding is that AVX 512 should allow doubling the floating point operations executed per clock cycle.
  12. A detailed spec table was published in this recent anandtech article NVIDIA Posts Full GeForce GTX 1070 Specifications: 1920 CUDA Cores Boosting to 1.68GHz Apparently nvidia decided to deliberately cripple FP64 performance in Pascal consumer cards and limit it to 1/32 of the FP32 performance, i.e. the same ratio as Maxwell except that Maxwell did not have much FP64 hardware in its silicon to start with. I have been expecting this to happen and I have a strong feeling that the Quadro variant will be partially disabled as well, except may be for the most expensive ones. FP64 is a specialized need and nvidia will make sure those who need it pay for it. Now this gives the GTX 1080 a theoretical maximum throughput of slightly more than 278 Gflops. To put this into perspective (the unknown effect of graphics RAM speed aside), this is less than half of the theoretical peak FP64 throughput of Intel's latest integrated GPU (Iris Pro P580). I know you said before that Open CL is not suitable. However, I hope you will consider using it at least to partially do what you could do with CUDA.
  13. Just a quick question, does HQ player support acceleration via Intel integrated GPUs?.
  14. I never found it clear how Lampizator does DSD conversion. Their website says "THE LAMPIZATOR DSD DAC HAS USB PORT BUILT IN, SOLID STATE DIGITAL FILTER, PASSIVE DISCRETE ANALOG FILTER AND ACTIVE DISCRETE TUBE FILTER. " Now, the only digital filter I have seen implemented for DSD in my limited knowledge is a "moving average" digital filter with different implementations. I understand it is built in the Burr Brown DAC chip which ifi uses, and is also implemented in Accuphase DACs via a combination of DSP and the multichannels of the ESS DAC chips. (Is this also the filter implemented in the DSC 1 with discrete components?). So for Lampizator I am not sure whether some DAC chip is involved or not in their DSD DACs. It is also not likely a no-dac approach (which I understand as being just simple low pass filtering). All of this is of course of little importance and more a matter of curiosity as apparently their analog implementation is superb and their DACs sounds great.
  15. Yes, they say they have developed their own 1-bit DAC and according to the available photos there are 4 of them inside the DAC8 DSD. It seems they follow a purist approach for their DSD processing inside the DAC. From the user guide: " In the past it has been standard practice to apply digital and analogue filtering processes in an attempt to reduce DSD noise, but such solutions are never entirely without side effects on sound quality. For the DAC 8 DSD we have developed two special techniques designed to eliminate the sonic disadvantages: 1. The True-DSD technique, consisting of a direct digital signal path without filtering and noise-shaping, plus our True 1-bit DSD D/A converter 2. Analogue reconstruction filter with adjustable bandwidth "
×
×
  • Create New...