I’ve only used a Pi so can’t speak from experience. The main virtue of an NAA is to be minimal. If there is any difference in SQ between them then I would expect the device which runs fewest processes and has the quietest power supply to be best. Not sure which one that would be.
I have read somewhere that the Pi shares it’s ethernet with usb. Just wondering if this would have an impact on SQ.
I’m sure Jussi will give us the answer.
Great to have those images, that’s going to make trying an NAA a lot more straightforward.
The million dollar question though for whoever’s tried it - did things sound better or worse using a Pi/NAA compared to the same settings direct from your HQP machine? Care to share the differences?
Also, anyone tried a ‘cheap’ LPS from eBay. I ordered one from China just for fun - when I do get round to trying an NAA/RAAT, I want to cover a few bases to see what’s possible compared to a high-end (I.e. much more expensive) streamers like the Aurender I demod.
I replaced a 5m Belkin USB cable with 5m Ethernet/Pi/1m Mapleshade Clearlight. The sound was more immediate and dynamic. Better leading edge on transients. I had some soft crackling and pops which I tracked down to how the Mapleshade cable was routed. It needs to be kept well away from power conductors as it is unshielded.
That’s the main difference, Pi uses SMSC’s USB ethernet adapter. BBB and CuBox-i/HummingBoard have on-chip ethernet. The AM335x used on BBB has gigabit ethernet, but the speed is limited by the LAN8710A 100 Mbps PHY. The NXP (originally Freescale) i.MX6 used in CuBox-i/HummingBoard has also on-chip gigabit ethernet and PHY is capable of gigabit ethernet.
For ethernet, there are still typically limitations due to internal bus speeds, like CuBox-i spec page has foot note “Note that due to internal i.MX6 buses the 1000Mbps interface speed is limited to 470Mbps.”
The i.MX6 is the most powerful of the three (also costs more), available in single, dual and quad core versions. NAA cannot really utilize more than dual-core. So the CuBox-i2eX is kind of cost-optimized best fit for NAA. I have only couple of the i4Pro models I use for various different things.
Each of the devices have their sides, hard to put things into any particular order.
I would like to give this a try, I purchased the HQPlayer license over Christmas. My only concern, since I’m not very familiar with Linux, is if any special drivers are required for the DAC? I’m hoping that Linux is not like windows and more like MAC in that respect,
Thanks Jussi…Awesome job
+1 Fantastic work Jussi. I had the RPi2 image up and running in minutes without any issues - playing music directly via the HiFiBerry amp module (to my very low-fi garage setup).
One question - is WiFi possible with the RPi builds? It would be great to have WiFi NAA as an option.
Hi Scott,
The images provided by Jussi are, I understand, Raspbian Lite with ALSA libraries and NAA installed. So with some Linux knowledge and a terminal (like PuTTY) you can configure a Wi-Fi USB adapter for use with the Pi2.
There are some N adapters about but I havent seen any AC ones supported by Raspbian. I’ll try to provide some links when at my PC again.
Edit: This looked like a reasonably straightforward tutorial about setting up Wi-Fi on a Pi.
Hi…
It was with some mixed feelings I noticed that Jussi now finally have provided images for the Raspberry Pi’s (and BBB). I had spend hours searching for such images before Christmas without any luck. I then spend 4 long evenings playing around with the Pi and building a minimal Linux NAA myself.
I guess I should thank Jussi anyway since I did learn a lot about Linux doing this exercise…So thank you 
Anyway, to the point…
I did manage to build a working NAA on my RP2 but soon after noticed that there where some minor clicks appearing every now and then (apx. a click around every 3-5 seconds). Sound almost like Vinyl noise.
It happens when outputting DSD256 but it seems to be absent on DSD128 or only happening rarely here…
I have heard about the RP USB problems because of the NIC sharing the same USB-bus. I thought that was a likely reason for this problem but hoped there was some kind of solution by increasing some buffers or so…
I now tried Jussi’s image for the RP2 but the problem is still there.
Am I really the only one experiencing this on DSD256 playback on the Pi?
By the way my DAC is an Ifi iDSD Nano with newest available firmware installed.
I have also used an older IBM ThinkPad laptop with Windows 7 as NAA, and from here DSD256 plays perfectly fine. That should prove that my clicking problem is indeed the Pi and not the HQP-pc or the network.
Regards…
As above, I did hear soft vinyl like pops using the Pi2 and upsampling Redbook to DSD128. One or two pops or clicks every minute or so. The popping continued whether powering the Pi from mains or a battery.
With time and after rerouting my USB cable away from power conductors those pops disappeared almost completely for me. I played through the Pi for about 2 hours (night before last) and thought I may have heard one pop.
It may be the Pi, your test with the Thinkpad indicates that, but my experience was the popping diminished with time and cable routing.
Edit: still hearing pops and clicks. It seems to come and go. I was hearing 1 or 2 clicks a minute last night. I don’t think it is cable routing.
I think I will check out a Cubox-i.
Thanks Andy.
I just did a few experiments. First I changed the USB cable to the little short blue one which came with the iDSD Nano. Then I tried a different RP2 device and ethernet cable. In both cases I first thought the clicks were gone but after a while I started to notice them again. They are not always easy to hear and it seems some music track disguises them more than other tracks…which after all is not surprising. Changing the modulator from ASDM7 to DSD7 256+fs somehow created more clicks or made them more noticeable at the same samplerate.
I then tried falling back to DSD at 6.1440 MHz and here it’s much better. I can hear many track without hearing any artifacts and I’m actually not sure if the occational tick I do hear is from the recording it self.
I can see you are using the very interesting Auralic VEGA DAC which unfortunately doesn’t support DSD256 so you cannot test this rate on your DAC.
Can anyone else try DSD256 and rapport back your findings.
If someone else do confirm clicking problems at DSD256 rates we could start looking for a possible solution.
Of course we could also just accept that the RP’s limit is DSD128 but for people with DSD256 capable DAC’s this would be somehow unsatisfactory.
I can easily find another device for a NAA but I do think the Raspberry Pi’s are important for a lot of people who would like to try out HQPlayer. Especially now Jussi have made prebuild images.
They are cheap and readily avalable at least here in Europe and many “nerds” have a Pi already.
Before Christmas I also considered buying a Cubox-i but these costs 3-4 times as much as a Pi and had 2 weeks delivery time. Not ideally for an HQPlayer test.
1 Like
That was the problem that plagued the very first RasPi version I had to the extent I discarded the entire RasPi as useless. Now I have B+ and 2 and those are better. Probably the add-on boards work even better since they don’t use USB. I shortly tested RasPi B+ with iDSD Nano and didn’t notice problems during my short testing, but I didn’t spend my time with it.
For me, CuBox-i has been the most robust ARM-based NAA platform. But also BeagleBone has improved over the years and works better with recent kernels (I ended up using BeagleBone with M2Tech hiFace (1) plugged straight to the USB connector, somehow was a good fit for it)…
It is not Raspbian, all images for the three devices are based on same custom OS.
I could make PC builds too, but the problem with PC hardware is widely varying set of hardware so it is complex to define hardware where it would work. But I could have a builds that I know would work on Intel Galileo and MinnowMAX boards, or some other limited and well defined set of hardware.
Thanks for clarifying Jussi,
Is the custom OS Linux based in a way that would enable the user to configure Wi-Fi as per this tutorial ?
Edit: I can’t connect to Jussi’s image through a terminal, so we can’t configure Wi-Fi this way.
Jussi, any plans to build an image for the Wandboard Quad? Thanks, Hammer.
When I first experimented with the Pi2 as a concept, I had a lot of issues with pops and crackles using squeezelite - and I could never get rid of them. Feint, occasional crackles, just like Devialet’s Air issues if anyone knows what that sounds like, probably as described above like feint vinyl noise.
Interestingly, using Volumio, the sound was rock solid and I never heard a single noise, but it wasn’t an interface/setup I was interested in, so I abandoned the Pi (until now, I want to experiment again!). All the research I did at the time said it was either the shared USB/ethernet on the Pi, or the Pi being underpowered (i.e. not enough current from the power adapter). No idea if either are the real reason - seems odd that Volumio was solid.
Apologies for going a bit OT, but should anyone think that the CuBox-I is the bee’s knees, it works great - most of the time. It does however have click/pop problems with DragonFly v1.2 DACs that can only be rectified by putting a hub or a reclocker (like Schiit Wyrd) in the loop. Possibly some other DACs for all I know.
So robust - yes. Problem free - no.
Yeah, I just checked these boards out. Looks really interesting for €100 (Dual-core) to €127 (Quad-core) including a case. If Jussi haven’t got time one of us users probably could create a working image.
Anyway…I still really would like if a few more of you guys would try DSD256 on your Pi and rapport if it’s “rock solid” for you…or if you are also hearing faint clicking noise. Please 