New Holo Audio Cyan DAC/AMP: PCM or DSD module?

Discussion in 'Dedicated Source Components' started by doraymon, Sep 28, 2017.

?

Would you be more interested in the PCM or DSD module for the Holo Audio Cyan DAC/AMP?

  1. PCM

  2. DSD

Results are only viewable after voting.
  1. gr8soundz
    @Energy thanks for all the posts about DSD transcoding. I did up to DSD512 in the past with an iDSD Micro, Foobar2000, and i7 4765T (with c-states, turbo, and hyperthreading disabled for less jitter) at about 40% usage. Also have an 8-core AMD FX machine but sounds like neither cpu has enough cores/cache to max out HQPlayer's settings.

    Haven't tried HQplayer yet but your posts got me thinking about my future upgrade to a Holo Audio DAC. Was also planning to get the Pink Faun i2S Bridge and avoid USB altogether but afaik that card doesn't do DSD.

    Assuming I got a DSD Cyan (or Spring), do you think DSD512 over USB/i2s via an SU-1 or Matrix X-Spdif 2 would perform better than PCM over a direct i2s to i2s connection?
     
  2. Energy
    Yes Jussi tells me memory access speed and amount of cache is critical.

    I never got the chance to use the ThreadRipper with HQPlayer. At the time I figured the 7900X would work just fine and so I was more focused on the aesthetics of the build but now all i really want is for poly-sinc-xtr to with DSD512.

    I booted my computer with power savings off and ran the 7900X at a consistent frequency base clock of 3.3GHz and the temperatures were around 60C. When HQPlayer used the CPU standalone (no CUDA offload), it rocketed up to 80C which made the music almost listenable/playable. It played without stuttering when I enabled CUDA offload which sent the filter to the GTX 1080 Ti for processing so due to splitting the loads, it kept the CPU temperature around 68C which wasn’t all that bad. The problem is that I didn’t want to have to rely on the graphic card just to process DSD. It would be awesome being able to play games while listening to music so I wanted my graphic card un-used by HQPlayer.

    Since the frequency was consistent, turbo boost didn’t work for some reason so I couldn’t see if the CPU would work standalone since the temperature was already 80C and any higher frequency would probably send it soaring up to dangerous levels.

    Long story short I made an order for the ThreadRipper 1950X (16 core) once again and have returned the 7900X. With 32MB worth of L3 cache and so many cores surely it will work this time around, can’t wait. I’ll see if it works at stock frequency. Ultimately I will probably end up overclocking it to 4.2GHz for gaming purposes.

    You’re welcome mate!

    I think hyperthreading and turbo-boost enabled should be fine. I think some CPU’s have issues going into idle/standby instead of operating at standard or turbo frequency which explains the stutters and why my 7900X didn’t work. That as well as having low L3 cache. This is probably why you managed to get yours to run smoothly after disabling the power saving modes or the CPU’s frequency throttling. There’s a better way around this though, one more natural and I think it has to do with what CPU is used. For example the 6950X used by Jussi I’m sure runs at standard clocks and idles to save on power consumption and lowers temperature. But miraculously when he plays music his cores functions properly and doesn’t drop into idle frequency states. I feel like certain old or too new CPU’s aren’t properly supported with HQPlayer or there may be various reasons with Windows or BIOS that prohibits it.

    AMD FX is fairly old, around 2011 if I recall and only a generation ahead of the bulldozer model. I don’t think it will be strong enough to handle an xtreme filter (non-2s) at DSD512 even when overclocked. Even at 5GHz my 7900X failed to do it but then again most of it’s cores throttled down to 1.2GHz so I don’t know anymore. I think the safest bet is to just get a CPU with lots of cores and L3 cache like the AMD ThreadRipper and call it a day.

    You know that USB isn’t bad right? It’s just the noise and jitter behind it partnered with the need for galvanic isolation which again falls in the category of noise. Regardless if Pink Faun does I²S it’s still tied to the dirty power source of a computer which takes 12V from the PCI-E or Molex connectors. Even when filtered is actually not that clean because it’s connected to a computer high speed circuit which is populated with plenty of high frequency noise that will in turn disrupt the audio band.

    I’m a little biased due to equipment I have but I did go this direction from sheer technical facts so my advice to you would be to get the SU-1 and pair it with a cheaper cable like the Apollo AV Lightning V2 0.5M for $15. So yes, I highly recomend you get the level 1 Spring to save yourself money. I’ve heard all the levels and can’t say the price justifies any sonic improvements unless you get one of the only 10 units made with 4 pole Jensen’s which is rare. Anywho, later down the line if I were you I’d incorporate a iFi iGalvanic 3.0 (or ISO REGEN) into the mix to separate the SU-1 from computer as it offers galvanic isolation alongside reclock, rebalancing, and regeneration of a cleaner USB signal to pass onto the SU-1 prior to USB to I²S conversion. Another extra if you have the money would to power the SU-1 with a LPS-1 by Uptone Audio but if $395 is too expensive for you, I’d recommend just changing out the internal regulators within the SU-1 for ones with lower noise to decrease further jitter generated by the Crystek HD-575 crystal oscilllators. I am one of the ones who started these mods over at Computer Audiophile forums and can help you out anytime in the future if need be.
     
    Last edited: Oct 17, 2017
  3. gr8soundz
    Threadripper may indeed be the best option if cores and cache take priority. Look forward to reading about your results with it.

    Only reason I'm looking at i2s is because it's the most direct feed possible. I've used USB for both DAC power and audio data for years and built my audio PC from the ground up with all fanless components including a low ripple Seasonic psu and no moving parts (other than the optical drive). My MSI motherboard even has isolated 5V power and USB ports specifically for DAC use. I also use an Jitterbug and iPurifier2 so my USB can't get much cleaner. However, each piece changes the sound a bit (for the better imo) and I'm wondering if all the conversions/passes the signal takes before getting to the actual DAC chip are having a negative effect.

    I looked at further USB upgrades but my current DAC has built-in galvanic isolation so I don't need the iGalvanic3 or Iso Regen (not to mention each of those costs about the same as my DAC). I also prefer internal components and less cabling to save desktop space and reduce clutter so the Pink Faun card is enticing.

    I've found that a proper R2R (or multibit) design negates some need (for me at least) to transcode to DSD. However, I'd still like the ability to upsample when desired. Also curious to know if HQPlayer can do DSD512 using 48kHz files and whether the Holo Spring/Cyan can do DSD1024? Only know of 2 DACs that officially handle 1024 (including the Matrix X-Sabre) but would love to know if the Holos are capable. Extreme, yes, but so is DSD512 and building super-high powered machines specifically for audio processing. In for a penny.....
     
  4. zolkis
    Call me interested... only if they would make an option without an on-board headphone amp, and an option without PSU (using external PSU).
     
  5. Jimi Zine
    Any more info on the Cyan?
     
  6. Yeang
    Hi Thanh @Energy ,

    Any progress with experimenting HQplayer DSD512 xtr with the Threadripper yet? I am placing all my hopes on you! I really don't want to have to build a 7920X or higher system.

    Fingers crossed that Threadripper can pull it off. Remember to use the highest speed DDR4 you can get for highest memory bandwidth. Jussi said that the memory speed/bandwidth is an important part, and he was excited about the 4 way interleaving available on Threadripper (like i9-79xx series too).

    Keep us updated!
    Yeang
     
  7. Energy
    Just because you use an i2S connection from the computer to the DAC doesn't make it a good connection though.
    Direct feeds only work well for cable implementation where the shortest distance is the superior one.
    When it comes to digital corrections however, reclocking is fine and does not influence negative feedback if done properly.

    There is no latency delay when all the bits are received and ultimately converted over to i2S when using such devices like Singxer SU-1.
    The Pink Faun use dirty generated computer noise from the 12V power supply from the computer. Despite single or multiple 12V rail, it's still very very noisy regardless of how many filtering the Pink Faun card has on it. Generating from a different line of voltage is better in this case scenario. The i2S output from the Pink Faun I reckon is also inferior to the Crystek HD-575 available on the SU-1 and I doubt the former has much digital isolation and power isolation stages as compared to the later. This is why direct doesn't mean it's good. It may remove the need to improve USB signal integrity, but a line out i2S from the computer to the DAC is a pretty long distance that's full of jitter unless you can somehow keep it within 0.3-0.5M and even then, the noisy power from the computer will just simply induce the jittr count even more so. I2S is better when it's being more regulated and this usually is only offered by external devices.

    I don't recall the Soekris dac1101 having galvanic isolation on it's input, especially at that price. I do know the new generation USB cards for the Schiit DAC's do, but again, that's USB.

    I can't agree to the statement where R2R design negates the need to transcode DSD. The Holo Audio Spring as you know has two separate ladder for PCM and DSD. Playing the same PCM file music transcoded to Native DSD to let the DSD R2R Ladder decode it simply sounds way better. This can either be due to the improved filter/modulation that gets applied to these files when using HQPlayer or there may be some anomaly with this DAC in particular to how well it decodes DSD versus PCM. Either way, it's amazing and so I would heavily recommend R2R Ladder (DSD) as well as encoding PCM to DSD to gain from the improved filters that matches the level of degree as the Chord Dave.

    I had a friend that can play DSD512 on 48Khz so should confirm. As far as I can tell, the highest the Spring can currently do is DSD512.
    The stock firmware for the Singxer SU-1 does only DSD256 and needs the new v2.20 firmware update to support DSD512.

    I do wonder how DSD1024 would sound like. I've had Matrix X-Sabre as well as the Pro (later release) and can say I prefer the Spring miles ahead.

    I will give you an update this weekend! Setup has been completed but the CPU doesn't get here until Friday (11/3/2017).
    I will test it at stock clocks to see if that works right out the box with upsampling PCM to DSD512 using the poly-sinc-xtr filter.

    Do you know anything about modulation? I'm not sure which modulator is the best. Jussi has his set to ASDM7 while I had mine at AMSDM7 512+fs this entire time. Bit rate is at 44.1 x 512 for me.

    And yes. I am using G.Skill Trident Z RGB at 3200MHz Quad Channel (32GB) which is four sticks of memory at 4-way interleaving.
     
  8. vizion
    Energy, can the filter DSD512 poly-sinc-xtr-2s utilize more than 4 cores using a less capable CPU but with many cores ? No GPU offload will be used.
     
    Last edited: Oct 31, 2017
  9. Energy
    Definitely. It can utilize up to 27 cores as long as the CPU you have is year 2013 and above.
    I would recommend POLY-SINC-XTR-MP-2S since it has better transients but entirely up to you!

    No CUDA Offload needed. 2S doesn't take much processing which is awesome. Try it and do share your findings :]
     
  10. gr8soundz
    Thanks @Energy for those details. Guess internal i2s won't have the advantage I imagined. I'll stick with USB for now but good thing the Cyan and Spring have a ton of input options. Still difficult to choose between the Cyan DSD vs. PCM though.

    I was looking into iFi's iGalvanic3.0 when I decided to email Soekris directly. Already have an iPurifier2 and the iGalvanic costs the same as my dac1101. Here's the reply I got regarding if the dac1101 has galvanic isolation?:
    "Yes, the USB interface is fully isolated, both signal and power."
     
  11. Energy
    Success! DSD512 works with POLY-SINC-XTR!

    I ran through a lot of problems such as the ISO REGEN that was used in-between the Singxer SU-1 broke on me. It made me think there were driver issues or that the computer had CPU issues.
    Long story short, both the ISO REGEN and RAM sticks did not function properly, but with those eliminated and accessed, everything works now. Quad channel isn't really needed as long as there was at least two sticks of ram for dual channel.

    Yeang, do you know anything about modulators? I've been using AMDSM7 512+fs and see Jussi using regular ADSM7. I'm still quite unsure which is better. I switched back and fourth and am quite sure they sound slightly different.

    Did you plan on using the Cyan as your all in one DAC/AMP? If that's the case i'd use a computer for processing and get the DSD version. Only considering you have a strong enough computer for the processing of course. If the Cyan has I2S input, I would probably pick the Singxer SU-1 over the iGalvanic in terms of priority. For me if I could only pick one, having the SU-1 has more of a sonic benefit than using the ISO REGEN (or iGalvanic).
     
    gr8soundz likes this.
  12. oneguy
    This is with CUDA offload or no offload?
     
  13. Energy
    No offload. Just purely the CPU.
     
  14. Yeang
    Hi Thanh @Energy ,

    That’s amazing news! I was waiting anxiously to hear.

    Can you give a little more details on setup and problems with ISO REGEN and RAM?

    Was the TR on default energy saving settings, did you have to disable C states, use Throttlestop etc?

    Also do you think that the 7900x might have worked if you solved the ram and iso regen problems or do you think that the 16 cores are really helping?

    I apologize as I have little experience playing with the modulators. I have only focused on the filters so far. I have no doubt it will make a difference!!

    Congrats again!
    Yeang

     
  15. Energy
    ISO REGEN ISSUE
    - ISO REGEN stopped working after Windows update. I uninstalled all the audio drivers and updated to the Windows Fall Creators Build 1709, then I reinstalled the Holo Audio Spring 3.40 driver and the ISO REGEN still did not work, Singxer SU-1 would not be recognized by the computer. Because of that I thought the ISO REGEN was broken. For some strange reason today I decided to it throw back into the equipment chain and it detected the Singxer SU-1 and plays nicely now without pop noises or detection issues so I'm glad I don't need to send it back.

    RAM ISSUE
    - The G.Skill Trident Z 32GB Quad Channel I bought did not run stable with it's rated X.M.P profile for 3000MHz. Right now I have it at 2133MHz but it's still good enough for xtr filters.

    PC Settings
    AMD 1950X.png

    - No ThrottleStop, Disable C-States, or Energy Settings is needed.
    All you need is the CPU at stock speeds with at least dual channel ram at 2133MHz. USB Streaming mode default "Reliable" is okay. No need for Extra Safe.

    - ISO REGEN problem did NOT affect why HQPlayer still stuttered on the Intel 7900X. The CPU simply does not have enough L3 cache. HQPlayer is a program that utilizes more L3 cache which is why the Intel 6950X worked and not the 7900X. I think the cheaper AMD 1920X should be good enough for the job with the extra 2 cores and more L3 cache.

    MODULATOR

    AMSDM7 512+fs sounds more weighty to me than the ASDM7 so I prefer it. ASDM7 was a little more thin sounding. It makes sense since AMSDM7 512+fs takes more CPU power so I assume it's the better performing modulator. You can experiment with it and let me know your results.
     

Share This Page