luchoh

Too much hassle to return it from where I currently reside :-) I'll probably give it away as a present to my nephew.

@FooFighter, I have Hifiman EF400 and Zen Stream and can probably help here. To start with, MacOS is correct - the EF400 DAC can only go to 192KHz. If you know a way to make it go higher, I'd be happy stand corrected. Subjectively, the EF400 DAC itself isn't all that great, compared to Topping D90 (I find it resolving, but dull) and Mojo 2 (the best of the bunch, though doesn't have a dedicated line out). The pianos with EF400 in particular are quite irritating. The EF400 HeadAmp is subjectively worse than Topping A90, which I also find dull, though the combination of EF400 DAC + A90 sounds livelier than EF400 on its own.

Thank you @Miska, I'll check with the spectrum analyzer and report back. I believe you're correct though - the firmware shouldn't lie.

@Miska do you mean the music is audibly slower? That's not what I hear. It plays at normal speed. I think I should be able to hear 8% difference in speed... is there a way to be certain? This is what the log says when inputting 96KHz: NAA output set sampling rate: 6144000 (6144000) Automatic output rate: 6144000 Rate or blocksize change triggered Rate: 96000, block size: 7680, frame size: 1280 Block size: 7680 (sample: 3) The DAC however shows 11.28MHz - probably because it is higher than DSD128 and this is what the firmware developer decided to show when > 5644800

@Miska after installing Embedded 4.15.2 (your kernel 5.4.22-jl+) and NAA 4.1.1 suddenly my DAC is recognized to support quite a few more rates: + 2020/04/03 13:42:00 NAA output connect to 192.168.20.55:43210 [ipv4] 2020/04/03 13:42:00 NAA output network format: 44100/24/2 [pcm] 2020/04/03 13:42:00 NAA output network format: 48000/24/2 [pcm] 2020/04/03 13:42:00 NAA output network format: 88200/24/2 [pcm] 2020/04/03 13:42:00 NAA output network format: 96000/24/2 [pcm] 2020/04/03 13:42:00 NAA output network format: 176400/24/2 [pcm] 2020/04/03 13:42:00 NAA output network format: 192000/24/2 [pcm] 2020/04/03 13:42:00 NAA output network format: 352800/24/2 [pcm] 2020/04/03 13:42:00 NAA output network format: 384000/24/2 [pcm] 2020/04/03 13:42:00 NAA output network format: 705600/24/2 [pcm] 2020/04/03 13:42:00 NAA output network format: 768000/24/2 [pcm] 2020/04/03 13:42:00 NAA output network format: 2822400/1/2 [dsd] 2020/04/03 13:42:00 NAA output network format: 3072000/1/2 [dsd] 2020/04/03 13:42:00 NAA output network format: 5644800/1/2 [dsd] 2020/04/03 13:42:00 NAA output network format: 6144000/1/2 [dsd] 2020/04/03 13:42:00 NAA output network format: 11289600/1/2 [dsd] 2020/04/03 13:42:00 NAA output network format: 12288000/1/2 [dsd] 2020/04/03 13:42:00 NAA output network format: 22579200/1/2 [dsd] 2020/04/03 13:42:00 NAA output network format: 24576000/1/2 [dsd] So it seems now I can play x48 with Topping D90. It was HQPe/NAA (or kernel) issue all along. What happens however is very interesting ... when i play a x44.1 file first and the engine is initialized, I can then go to a x48 file in my playlist. 2020/04/03 13:50:37 NAA output set sampling rate: 12288000 (12288000) 2020/04/03 13:50:37 Automatic output rate: 12288000 2020/04/03 13:50:37 Rate or blocksize change triggered 2020/04/03 13:50:37 Rate: 96000, block size: 7680, frame size: 1280 2020/04/03 13:50:37 Block size: 7680 (sample: 3) 2020/04/03 13:50:37 Oversampling: sinc (1M taps) 2020/04/03 13:50:37 Modulator: adaptive fifth order 1-bit ec 2020/04/03 13:50:37 Integrator: IIR 2020/04/03 13:50:38 Playback engine ratio: 128 However, if I change settings in HQPe, then playing directly x48 produces silence (the log file shows that it's playing - NAA output network engine started at: 12288000 and Roon tracker is moving). If I keep changing things at some point it will directly refuse to play with: No suitable output rate for 96000, stop If I skip to x44.1 track and then back to x48 track - it plays fine... It's not the end of the world for me as I found a workaround, but I figured it may be helpful to you. I'm using CUDA offload btw. I can supply the config file should you need it.

@Miska there must be a setting that eludes me. If I uncheck 48k DSD and the stream rate is a multiple of 48KHz, the DAC refuses to play. And it's not only D90 - I wasn't able to do it with any of the other ones. I suspected it should be possible to do the resampling only in HQP, but ... obviously I'm missing something. I'm getting & 2020/01/28 13:24:01 Play 2020/01/28 13:24:01 Offload: resampler=enabled convolution=disabled + 2020/01/28 13:24:01 Playback engine running 2020/01/28 13:24:01 No suitable output rate for 96000, stop ... I suppose I needed to uncheck Auto rate family...

@Miska thank you! I would make one more post so that people who search for info have some reference. 1. When I upgraded the Odroid XU4 to Armbian with 5.4.6 kernel and NAA 4.0, the native DSD is ON. 2. Unlike the x64 (or maybe NAA 3.6.1 or maybe NAA image vs Armbian), the NAA 4.0 doesn't advertise the 48k DSD support, see below: + 2020/01/28 12:03:31 Connect to 192.168.20.55:43210 [ipv4] 2020/01/28 12:03:31 Network format: 44100/24/2 [pcm] 2020/01/28 12:03:31 Network format: 48000/24/2 [pcm] 2020/01/28 12:03:31 Network format: 88200/24/2 [pcm] 2020/01/28 12:03:31 Network format: 96000/24/2 [pcm] 2020/01/28 12:03:31 Network format: 176400/24/2 [pcm] 2020/01/28 12:03:31 Network format: 192000/24/2 [pcm] 2020/01/28 12:03:31 Network format: 352800/24/2 [pcm] 2020/01/28 12:03:31 Network format: 384000/24/2 [pcm] 2020/01/28 12:03:31 Network format: 705600/24/2 [pcm] 2020/01/28 12:03:31 Network format: 768000/24/2 [pcm] 2020/01/28 12:03:31 Network format: 2822400/1/2 [dsd] 2020/01/28 12:03:31 Network format: 5644800/1/2 [dsd] 2020/01/28 12:03:31 Network format: 11289600/1/2 [dsd] 2020/01/28 12:03:31 Network format: 22579200/1/2 [dsd] I will hope for future firmware update, but won't hold my breath. @Miska one last question: a workaround for this issue is to resample outside of HQPlayer (in my case - Roon). Should I upsample all 48x rates to 705.6KHz and then feed it to HQPlayer or selectively for each rate? Lucho

@Miska, thank you for the tip - it worked! I booted one of the MacMinis with your x86 appliance and I have partial success. Now the 44.1 rates work with DSD and 48 - don't - white noise. I did check 48k DSD option - this is how I've been using the other DACs previously. This is what I'm getting, suggesting it should work at DSD256x48: 2020/01/28 08:48:03 Network endpoint: D90: USB Audio (hw:CARD=D90,DEV=0) + 2020/01/28 08:48:03 Connect to 192.168.20.56:43210 [ipv4] 2020/01/28 08:48:03 Network format: 44100/24/2 [pcm] 2020/01/28 08:48:03 Network format: 48000/24/2 [pcm] 2020/01/28 08:48:03 Network format: 88200/24/2 [pcm] 2020/01/28 08:48:03 Network format: 96000/24/2 [pcm] 2020/01/28 08:48:03 Network format: 176400/24/2 [pcm] 2020/01/28 08:48:03 Network format: 192000/24/2 [pcm] 2020/01/28 08:48:03 Network format: 352800/24/2 [pcm] 2020/01/28 08:48:03 Network format: 384000/24/2 [pcm] 2020/01/28 08:48:03 Network format: 705600/24/2 [pcm] 2020/01/28 08:48:03 Network format: 768000/24/2 [pcm] 2020/01/28 08:48:03 Network format: 2822400/1/2 [dsd] 2020/01/28 08:48:03 Network format: 3072000/1/2 [dsd] 2020/01/28 08:48:03 Network format: 5644800/1/2 [dsd] 2020/01/28 08:48:03 Network format: 6144000/1/2 [dsd] 2020/01/28 08:48:03 Network format: 11289600/1/2 [dsd] 2020/01/28 08:48:03 Network format: 12288000/1/2 [dsd] 2020/01/28 08:48:03 Network format: 22579200/1/2 [dsd] 2020/01/28 08:48:03 Network format: 24576000/1/2 [dsd] Any ideas? I tried NAA 3.6.1. Maybe 4.x? Thanks a lot for your quick help! Lucho

@Miska, first, let me thank you for your fantastic software - I'm using it daily for the past 3 years. I just got a D90 and the NAA cannot recognize it (the native DSD that is), I'm sure it's not NAA's fault, but wanted to mention it here, before I hit the other forums and ultimately - return the DAC if there is no solution. I have two other DACs here (not counting DSC2 which is patiently waiting for output transformers)L Khadas Tone board and iFi iDAC2. Both are being recognized properly - see HQPlayer log: 2020/01/27 17:19:47 Network endpoint: iFi (by AMR) HD USB Audio: USB Audio (hw:CARD=Audio,DEV=0) + 2020/01/27 17:19:47 Connect to 192.168.20.20:43210 [ipv4] 2020/01/27 17:19:47 Network format: 44100/32/2 [pcm] 2020/01/27 17:19:47 Network format: 48000/32/2 [pcm] 2020/01/27 17:19:47 Network format: 88200/32/2 [pcm] 2020/01/27 17:19:47 Network format: 96000/32/2 [pcm] 2020/01/27 17:19:47 Network format: 176400/32/2 [pcm] 2020/01/27 17:19:47 Network format: 192000/32/2 [pcm] 2020/01/27 17:19:47 Network format: 352800/32/2 [pcm] 2020/01/27 17:19:47 Network format: 384000/32/2 [pcm] 2020/01/27 17:19:47 Network format: 2822400/1/2 [dsd] 2020/01/27 17:19:47 Network format: 3072000/1/2 [dsd] 2020/01/27 17:19:47 Network format: 5644800/1/2 [dsd] 2020/01/27 17:19:47 Network format: 6144000/1/2 [dsd] 2020/01/27 17:19:47 Network format: 11289600/1/2 [dsd] 2020/01/27 17:19:47 Network format: 12288000/1/2 [dsd] and 2020/01/27 17:22:55 Network endpoint: Khadas Tone Control: USB Audio (hw:CARD=Control,DEV=0) + 2020/01/27 17:22:55 Connect to 192.168.20.55:43210 [ipv4] 2020/01/27 17:22:55 Network format: 44100/32/2 [pcm] 2020/01/27 17:22:55 Network format: 48000/32/2 [pcm] 2020/01/27 17:22:55 Network format: 88200/32/2 [pcm] 2020/01/27 17:22:55 Network format: 96000/32/2 [pcm] 2020/01/27 17:22:55 Network format: 176400/32/2 [pcm] 2020/01/27 17:22:55 Network format: 192000/32/2 [pcm] 2020/01/27 17:22:55 Network format: 352800/32/2 [pcm] 2020/01/27 17:22:55 Network format: 384000/32/2 [pcm] 2020/01/27 17:22:55 Network format: 705600/32/2 [pcm] 2020/01/27 17:22:55 Network format: 768000/32/2 [pcm] 2020/01/27 17:22:55 Network format: 2822400/1/2 [dsd] 2020/01/27 17:22:55 Network format: 3072000/1/2 [dsd] 2020/01/27 17:22:55 Network format: 5644800/1/2 [dsd] 2020/01/27 17:22:55 Network format: 6144000/1/2 [dsd] 2020/01/27 17:22:55 Network format: 11289600/1/2 [dsd] 2020/01/27 17:22:55 Network format: 12288000/1/2 [dsd] 2020/01/27 17:22:55 Network format: 22579200/1/2 [dsd] 2020/01/27 17:22:55 Network format: 24576000/1/2 [dsd] However, try as I might, D90 is only recognized as PCM and can only play DoP. The "lowly" Khadas is totally fine at DSD256x48, while D90 chokes at this as it can't go above 768KHz to encapsulate DSD256x48. 2020/01/27 17:21:05 Network endpoint: D90: USB Audio (hw:CARD=D90,DEV=0) + 2020/01/27 17:21:05 Connect to 192.168.20.55:43210 [ipv4] 2020/01/27 17:21:05 Network format: 44100/32/2 [pcm] 2020/01/27 17:21:05 Network format: 48000/32/2 [pcm] 2020/01/27 17:21:05 Network format: 88200/32/2 [pcm] 2020/01/27 17:21:05 Network format: 96000/32/2 [pcm] 2020/01/27 17:21:05 Network format: 176400/32/2 [pcm] 2020/01/27 17:21:05 Network format: 192000/32/2 [pcm] 2020/01/27 17:21:05 Network format: 352800/32/2 [pcm] 2020/01/27 17:21:05 Network format: 384000/32/2 [pcm] 2020/01/27 17:21:05 Network format: 705600/32/2 [pcm] 2020/01/27 17:21:05 Network format: 768000/32/2 [pcm] I am using Odroid XU4Q as NAA (Debian GNU/Linux 9 \n \l, Linux DietPi 4.14.66+ #1 SMP PREEMPT Thu Aug 23 05:59:33 UTC 2018 armv7l GNU/Linux). I have plenty of other SBCs (BBB, RPi3 and Odroid C2) and a couple of old (2011-2012) MacMinis that I could use if you think it might help. I can also install some Armbian with a newer kernel if it would make a difference. I'd also be happy to do alpha/beta testing if you are interested. Finally, sorry if this is off-topic: are you going to release an embedded version of HQP to match NAA4? Thanks, Lucho

Hi, I don't know who this should be addressed to - @Miska or DietPi or this forum. In a nutshell, when playing DSD256 (DoP) from HQPlayer USBridge behaves as if it can't consistently lock to the signal: it would play 1 second, then no signal for a second, then play again. Playing PCM 764KHz is OK. Playing DSD256 from Roon is OK. DietPi on BBB Clack/Green acts the same as Sparky. Odroid XU4Q is flawless in the same environment (with DietPi). The DAC is RME ADI-2 DAC (DSD via DoP). Playing native DSD256 with any of those SBC devices to iFi IDAC2 and DSD512 to iFi IDSD is also without problems. Any help would be appreciated.

@crenca Mixture mostly, predominantly 16/44 PCM. Unfortunately I'm having an Internet outage, just after I started A/B-ing, so I can't tell yet what sounds better. I need to clarify that the MQA -> iDAC2 is served not via Tidal, but Audirvana. I haven't tested if there is a difference between Tidal and Audirvana yet... so many things to do - so little time.

@AMR/iFi audio I think if you call this tune/remap rather than "firmware upgrade", people would be less bothered. I personally have iDAC2 and iDSD BL and I don't plan flashing the BL, while I flashed the iDAC2 immediately. Still comparing the results of MQA/iDAC2 vs HQPlayer@DSD512/iDSD. All via RCA outputs, to minimize the headamp differences. Finally, I am grateful for your efforts and as a software engineer understand why you had to make sacrifices to retroactively implement something to a system it wasn't designed for in the first place. Sorry it somewhat backfired on you.

@Miska I'm having a weird issue with the embedded version and SMS-200: It won't be discovered. I pulled the desktop version and there is no problem whatsoever. I even went ahead and copied the line from HQPlayer.xml and replaced the respective line in hqplayerd.xml. No joy. I think what is happening is normally the NAA reports the sample rates, but with the embedded version and SMS-200 this doesn't happen. This is what the log says: 2018/03/24 10:44:49 Discovery from 0.0.0.0 & 2018/03/24 10:44:49 Discovered network audio: name='NAA' version='Signalyst Network Audio Daemon 3.5.2' @10.10.1.25:43210 & 2018/03/24 10:44:49 Discovered network audio: name='NAA' version='Signalyst Network Audio Daemon 3.5.2' @10.10.1.25:43210 2018/03/24 10:44:50 Network endpoint: iFi (by AMR) HD USB Audio: USB Audio (hw:CARD=Audio,DEV=0) 2018/03/24 10:44:50 Discovered 1 Network Audio Adapters 2018/03/24 10:44:50 Set channels: 2 (2) 2018/03/24 10:44:50 DAC bits: 24 ! 2018/03/24 10:44:50 clPlayerDaemon::CreateEngine(): clHQPlayerEngine::Initialize(): clNetMiniEngine::Initialize(): adapter not found - 2018/03/24 10:44:50 Network engine stopping... And this is the info from SMS-200: **** List of PLAYBACK Hardware Devices **** card 0: Audio [iFi (by AMR) HD USB Audio], device 0: USB Audio [USB Audio] Subdevices: 1/1 Subdevice #0: subdevice #0 Bus 004 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub Bus 002 Device 007: ID 20b1:3008 XMOS Ltd Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 003 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub Bus 001 Device 002: ID 05e3:0610 Genesys Logic, Inc. 4-port hub Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub 0 [Audio ]: USB-Audio - iFi (by AMR) HD USB Audio iFi (by AMR) iFi (by AMR) HD USB Audio at usb-1c1c000.usb-1, high speed iFi (by AMR) iFi (by AMR) HD USB Audio at usb-1c1c000.usb-1, high speed : USB Audio Playback: Status: Stop Interface 1 Altset 1 Format: S32_LE Channels: 2 Endpoint: 1 OUT (ASYNC) Rates: 44100, 48000, 88200, 96000, 176400, 192000, 352800, 384000 Data packet interval: 125 us Interface 1 Altset 2 Format: SPECIAL DSD_U32_BE Channels: 2 Endpoint: 1 OUT (ASYNC) Rates: 44100, 48000, 88200, 96000, 176400, 192000, 352800, 384000 Data packet interval: 125 us And Kernel Version 4.4.14-209.fc22.armv7hl (SMP) armv7l Distro Name Fedora release 22 (Twenty Two) The same DAC with BBB/Stretch with kernel 4.4.29 and Odroid C2/Stretch kernel 4.4.14 is discovered properly by both embedded and desktop. Any ideas?

@Miska thank you!

@Miska You are exactly right. I couldn't get the CUDA offload to work, but then I decided to "hack" the configuration file and add cuda="1" as such: <engine auto_family="1" channels="2" direct_sdm="0" dsd_6db="0" pdm_conv="5" pdm_filt="0" sdm_integrator="0" cuda="1"> Then, it started playing, but the log file was still showing 2018/03/21 10:22:51 Offload: resampler=disabled convolution=disabled however, nvidia-smi was happily reporting this: root@HQBuntu:/home/luchoh# nvidia-smi Wed Mar 21 10:24:35 2018 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 390.30 Driver Version: 390.30 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | |===============================+======================+======================| | 0 GeForce GTX 105... Off | 00000000:05:00.0 Off | N/A | | 53% 50C P0 N/A / 75W | 241MiB / 4036MiB | 78% Default | +-------------------------------+----------------------+----------------------+ +-----------------------------------------------------------------------------+ | Processes: GPU Memory | | GPU PID Type Process name Usage | |=============================================================================| | 0 15420 C /usr/bin/hqplayerd 231MiB | +-----------------------------------------------------------------------------+ While without the config parameter, it was showing +-----------------------------------------------------------------------------+ | Processes: GPU Memory | | GPU PID Type Process name Usage | |=============================================================================| | No running processes found | +-----------------------------------------------------------------------------+ What are you right about? The GPU is so slow, it makes even the DSD512 -2s stutter, while without CUDA offload, it is smooth. So now I need to find a suitable GPU, preferably below $300. Any suggestions?