Who made your DSC3 ? And did they detail changes over DSC2?
Any photos of internals?
Exactly. it is about taste and perceptionā¦ and again, about the quality of components used. You may not take my word for it but, hand crafted, by experienced engineers , unique devices will always outperform mass production. When making a custom product, there is a possibility to use rare, NOS components, chinese factories will never have access to.
When i bought DSC2, i invited a friend over. He is so much into audiophile stuff, using stuff like Macintosh phono corrector, Macintosh amp, speakers priced like a flat and owns a collection of few thousands vinyl original discs. After couple of hours listening to hqplayer>DSC2>professional active speakers, he went home sad. Next day he asked about my setup and where to buy stuff i use.
1 Like
Thats my typoā¦ i have posted today earlier logs and configsā¦
Can you ask which model active speakers?
RCF NX932 with an 18AX subwoofer
maybe this error pops up because NAA is running wrong time? And there is no way of chaning timezoneā¦
naa-bd44e482:/home/dev# timedatectl list-timezones
UTC
nothing but UTC available
These are DSP active XO speakers.
Meaning if you feed them a DAC output, they convert the analogue to digital and then digital back to analogue.
This is not the way to evaluate a DAC in my opinion.
The last DAC you hear are the speakers internal.
I am aware of that, thank you. But if you are familiar with these speakers, you would know how transparent they are. Anyway long story short, i am perfectly satisfied with his setup, and everyone who took place at my parties, were really impressed. Now i am focused to deliver DSD1024 multi channel with convolution engine.
Well donāt get me wrong. I love my dsc2 and dsc3, without a doubt, the first day I listened to the dsc3 in my system, I listed and sold the cyan 2 the next day.
But thereās no getting away with the principles, you stuck a pair of transformers there, itās acting like a natural low pass filter, bass below 50hz are distorted and will sound soft
highs would sound a bit roll off with holographic, thatās what distortion does from the transformers. It sounds great to my ears, itās a very colored sound, far from neutral, and no way I would call the dsc2 a clean sounding dac
The same buddy who made me the dsc2!!
Here are a couple of pics of the internal and me testing between the Crystek 957 clock and the Ian Canada SC pure clock
Does anyone know why hqplayer choses lowest rate on spdif input ?
I am feeding hqplayerd a stream from pi with fifopi q7 hat + transportpi AES hats via toslink to the input socket of the SoundBlaster Z SE installed in the PC running hqplayerd. The output can be 44100, 48000 or 96000, but hqplayerd always choses 32000 ..
2025/07/29 20:11:54 ALSA input access mode: RW_INTERLEAVED
2025/07/29 20:11:54 ALSA input PCM format: S32_LE
2025/07/29 20:11:54 ALSA input PCM bits: 32
2025/07/29 20:11:54 ALSA input PCM physical width: 32
2025/07/29 20:11:54 ALSA input PCM rates: 32000 - 96000
2025/07/29 20:11:54 ALSA input PCM channels: 2 - 2
2025/07/29 20:11:54 ALSA input active PCM channels: 2
2025/07/29 20:11:54 ALSA input rate available: 32000
2025/07/29 20:11:54 ALSA input rate available: 44100
2025/07/29 20:11:54 ALSA input rate available: 48000
2025/07/29 20:11:54 ALSA input rate available: 88200
2025/07/29 20:11:54 ALSA input rate available: 96000
2025/07/29 20:11:54 ALSA input DSD not supported
2025/07/29 20:11:54 ALSA input channels: 2 (2)
+ 2025/07/29 20:11:54 ALSA input backend initialized
2025/07/29 20:11:54 Audio transport: rate=0 channels=2 format=auto buffer=100 (normal)
2025/07/29 20:11:54 ALSA input sampling rate: 32000 (32000)
2025/07/29 20:11:54 IntegratorM: FIR2
& 2025/07/29 20:11:54 Play (-1/0)
2025/07/29 20:11:54 Web switch input: Sblaster
+ 2025/07/29 20:11:54 Playback engine running
2025/07/29 20:11:54 IntegratorM: FIR2
2025/07/29 20:11:54 Engine reinit, rate or blocksize change triggered
2025/07/29 20:11:54 Rate: 32000, block size: 3408, frame size: 426
I have tried all possible options documented in the readme or examples from hqplayerd.xml
Maybe it is alsa downsamplingā¦ cant get to the bottom if this.
<input channel_offset="0" channels="2" device="hw:CARD=Creative,DEV=1" name="Sblaster" period_time="100" samplerate="0" sync_element="SPDIF Rate" sync_hctl="1"/>
This should work, but it doesnt. For the sake of experiment, i plugged the coax of the pi stack into ADI-2, while hqplayerd is running with tosslink input. ADI-2 detects sample rate successfully.
nwm, had to spend a day for another typo 
Time zone shouldnāt matter. NAA OS and HQPlayer OS run on UTC. And now includes NTP sync. In particular if you use net trigger, then time zone doesnāt matter. But also the time trigger is based on UTC regardless of time zone settings.
How accurate is NTP sync ? Will it be possible to stream to four pi NAA, connected into the same network switch ? for multi channel purpose of the same data.
Considering all four pi + reclocker hats are located in the same case, and its possible to share 1 mclk signal.
It depends on how long itās been running and the hardware you have. NTP takes a while to stabilize clocks on a computer. But in many cases net trigger is better solution.
Given the way many audio drivers work, set Buffer Time to 1 ms for most accurate start.
Eeek, you were to fastā¦
Thanks for the answer. I suppose if one pi cant handle 4 reclocker hats, using a pi board for each hat is possible, with a shared ocxo.
In theory, your NTP clients can have at least stratum 2 accuracy if you have a stratum 1 NTP server locally. My HQPlayer server and NAA client are all stratum 1 grade (with direct GPS + PTP time sync). FYI.
1 Like
It looks like NAA OS is not supporting boot from nvme device on pi 5. Even if i edit cmdline.txt, and change root partition to /dev/nvme0n1p2, it boots until ethernet adapter is detected and then goes into infinite loop of audit.
Yes, likely doesnāt work. microSD card is a good choice, since it is much lower power/noise than NVMe.
This sounds more like a power supply shortage. Usually if PSU is too weak, boot will eventually fail at some point, typically triggering reboot, so youād end up in a reboot loop. But it can also fail other ways.
already eliminated that causeā¦ if the power supply isnt delivering, pi 5 wont just wont boot, it will halt just after reading config.txt etc.
i tried an mean well 6.0A and a an overkill power supply from my notebook, its USB-C with PD.
Booting a clean flash for the first time, gives errors like fat32 partition wasnāt unmounted correctly, run fsck and other aberrationsā¦ will stick to microsd.