Jump to content


  • Content Count

  • Joined

  • Last visited

About audiofool

  • Rank

Recent Profile Visitors

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

  1. Thanks for the info on Noise shaping. Looks like 32-bit and LNS15 is the way to go. Snippets from the previous song occur in both Embedded and Desktop using HQP playlist, I'm also using NAA. Both flac and dsd sources behave the same, I noticed using Sinc-M. Hit stop in the middle of a song and then click on another on in the playlist.
  2. If I understand this correctly, it is a tough decision between lowering the number of bits to deal with linearity or sending 32-bits to give the internal DSP the best precision? Does noise shaping also benefit 32 bit data? DSD seems so much simpler. When I click on another song in the playlist and I get a snippet of the previous song playing, is this something that can be fixed or expected behavior with long filters? Thanks Miska
  3. Makes sense, delay is long enough. One issue using Sinc-M is if I hit the stop button in Desktop version and then click on another song in the playlist I get a snippet of the previous song playing. I emailed FiiO, they are looking into my dsd bypass mode request. I grounded the power supply and I can hear a -135db 1khz tone with more sensitive headphones using 32-bit and TPDF dither. What does the AKM4493 do with 32 bit data since state of the art is in the low 20 bits? Are there cases where it is technically better to send 32-bits instead of for example 22-bits?
  4. Looks like the K3 and the K5Pro use the same ID
  5. I find noise shaping dither to always sound brighter than TPDF for some reason, though I Iike sinc-short-mp with LNS15 at 18bits but with sinc-M I prefer TPDF. sinc-M is interesting, any chance for the an extreme 16M apodizing version for PCM? Weird thing with Archlinux and dsd. I incorrectly assumed because Arch uses the latest kernel that FiiO would be supported in the kernel. I am using NAA so not an issue for me, but if anyone is using a XMOS 208 dac which is unsupported and require dsd support, this is what I did to create an alias to the XMOS ID: lsusb -> to get VID and PID of FiiO XMOS is 20b1 FiiO K5Pro is 2972 sudo modprobe -r snd-usb-audio sudo modprobe -v snd-usb-audio "quirk_alias=29720047:20b10047" cat /proc/asound/card*/stream0 should show -> SPECIAL DSD_U32_BE I emailed FiiO requesting a firmware update to enable dsd bypass mode - not holding my breath. Also, the phase inversion is working correctly on the latest embedded version, Thanks Miska.
  6. The EVGA is on my radar for when I build a new PC. Since the below results for the EVGA were similar to the FiiO results I used your EVGA measurements as a starting point for the FiiO linearity. So, I created a 192khz 32-bit float 1khz sine wave at -110db in Audacity. HQPlayer set to pcm and 0 gain with LNS dither and volume on the dac set to maximum. Can just hear the tone above a very low noise floor. I was looking for the lowest noise with the cleanest tone. 18 or 19 bits seemed the best, anything higher including 32 bit seemed to make the noise floor not as clean. smd output mode was clean but the noise floor was higher than pcm, so a no go. Is this a valid alternative to listening to fade outs to determine dither settings? Thanks Miska
  7. They replied saying DSD is converted to PCM internally so that DSD's volume can be adjusted. Really strange since they show this diagram. Very slight clicks at end of dsd songs in sdm mode, no clicks in pcm mode. 512sdm has the same gain as 256sdm but needs to be power cycled to switch back to pcm so something is different with 512 sdm mode in the dac. I was looking at this for linearity, maybe it doesn't mean anything and AKM is best at 32 bits?
  8. No dsd options https://fiio.com/k5pro They are using a NJW1195 chip for volume after the dac, seems like they are converting sdm to pcm in the xmos chip since sdm output is louder than pcm output for a sdm file. Thanks for the details on sdm vs pcm volume, sdm mode has about a 6db gain vs pcm instead of -3.5db in sdm mode, which was confusing me. I tried your NAA image on my HP ProDesk, DSD_U32_BE shows up. Something missing with the Archlinux kernel so not a hardware issue. I am now trying to decide between TPDF at 32 bits or LNS15 at 19 bits at 768k. Thanks Miska
  9. I'm on 5.7.7-arch1-1 I think it was status or get-status that shows a 1 in Desktop when inverted but always 0 in Embedded. I think it sounds better in sdm mode, wish there was a way to know if it was working in bypass mode. Also, when sending sdm to sdm the volume is approx 6db louder than pcm to sdm and sdm to pcm. I have dsd_6db="0" is this normal with sdm to sdm upsampling?
  10. Trying a Fiio K5 Pro and have some dsd issues 1. DSD works from my cubox using NAA image but doesn't work from my HP Prodesk, PCM 768k work fine. The HP shows Format: SPECIAL but is missing DSD_U32_BE. I am running Archlinux with the latest kernel. Anything else I need to load or is it a hardware limitation from the usb ports? 2. Maybe a bug, hqp-control2 localhost --set-invert 1 doesn't seem to work with Embedded version 4.18.0. Also is dsd wavpack support still on your todo list? 3. I am trying to maximize efficiency since I have a slow i5 cpu and the Fiio K5 Pro doesn't seem to do 48k dsd, it appears to work but plays static. I understand this is common, but I will email the company about a firmware update. On the plus side it seems to mute between sample rate changes, etc. 4. I don't believe the Fiio runs in dsd bypass mode so would PCM 768 or SDM 256 be the best choice for the AKM 4493 not running in bypass mode? I think someone said the SDM 256 measures better than 512 for the AK4493.
  11. My current computer has a M.2 slot for a SSD, would it be possible to use something like this with an external 12 volt supply https://www.amazon.ca/dp/B081CCVSYW?tag=duc12-20&linkCode=ogi&th=1 Anyone used anything similar? Good idea on the mini-ITX, a slippery slope though, next I'll be building a supercomputer.
  12. @MIska, I'm still considering the FiiO K5 Pro since it uses the AKM4493 chip, even though it appears to not use the DSD bypass mode. I believe the RME ADI-2 measures almost as good as dsd using PCM at 768k where I assume it cannot run in bypass mode. Would dsd 256 or 512 from hqplayer bypass the onboard filters with the FiiO K5 Pro or is the bypass mode a necessity for the best performace from the AKM4493? Also, does the latest NAA image for the cubox work with the XMOS XUF208?
  13. Anybody spec'd out an EVGA NU Audio Card with a cheap intel NUC or custom motherboard as a NAA end point? The FIIO K5 Pro almost fit the bill for me but doesn't appear to run in DSD bypass mode from what I could gather from the company.
  14. Excellent, I was hoping to try this in the new version. Is it usable now with the bug? How does one calibrate the cross-feed? Is this also planned for the desktop version?
  • Create New...