Yes, but in this case it might be the hardware itself that’s causing the problems (that static when playing PCM768). I’m also waiting for Allo’s final answer before I decide what to do next. Anyway, thank you for your input!
1 Like
Thanks anyway. It’s back at Allo and being refunded today. Allo produced a great product which worked well 99% of the time, but a few Chord DAC users have had issues.
1 Like
As far as I can see, that image doesn’t work on the USBridge Signature. It doesn’t seem to boot. The player does not get an ip address from the router nor does the activity led blink.
I just recently picked up a used Chord Dave. It’s a big improvement over the Hugo 2. The other thing is I find the USB to sound better than the optical. This is the reverse of my findings with the Hugo 2. I’m sure it has to do with the galvanized USB port.
So now that USB works better than optical it made testing HQPlayer easier against Dave’s internal upsampler. I still found I preferred the internal upsampling vs HQPlayer. It all comes down to tone. Dave has a warmer tone and vocals sound better to me. HQPlayer seems like it may have slightly better separation but also because it has colder sound. And the Dave already has amazing separation. So I’m sticking with just the Dave and maybe getting the mscaler down the line.
Optical is 100% galvanically isolated and all inputs to Dave are totally immune to jitter. If you tested optical while you had the USB input still connected to your source you compromised your test.
I am sure, all else being equal, toslink presents less RF intrusion and thus presents the music with its native sonic signature.
Well I’ll try again but Rob also mentioned that with USB it uses the internal clock as well. Mind you, I’m not as knowledgeable on all of this. Listening was personal preference.
Think I have issue with my optical cable and the Dave. It won’t play back 192k files. It also seems weird on Dave is optical doesn’t snap into place like on the Hugo 2. Going to have try another cable on it to test optical.
I got it to work. Optical port in iMac mini needed to be blown out. Also get cable to lock in place. And yes, now optical does sound better than usb.
The inability to play 192k via toslink is to do with a lack of proper insertion. Yes to the hard click …so the laser light and sensor are at the right distance. Check to ensure you cable ends are not bulky preventing full insertion.
I was still getting little pops at 192k so not completely working right. Think need to clean out port better. Have no problem locking cable to Dave with a hard push. So for now I’m using usb and do find it sounds excellent as well.
@jussi_laako does your latest RPi3 NAA image support Allo USBridge Sig?
Or is this list of RPi overlays up to date?
No it doesn’t…
Yes, that is up to date list.
1 Like
So if finally tested HQPlayer sinc-m LNS 15 and the xtr2 LNS15 against the Chord Hugo Mscaler on my Chord Dave with Focal Utopia. To beat to the punch, the mscaler won.
I’m not a fan of the sinc-m filter. It does some things the mscaler does like open up soundstage and tightens the bass but seems to strip the texture and changed the tone. I’d rather listen to the Dave without out it
The ext2 had better tone and texture but still falls short of the mscaler. There’s just something more satisfying and enjoyable about the mscaler. And for me it comes down to the texture retrieval. Mscaler is just so real and HQPlayer is too smooth. And believe me, if HQPlayer sounded as good or better, I could easily sell the mscaler. I’ve already paid for HQPlayer.
I’m allergic to “digitus” or the harshness of many digital systems. Likely M-Scaler is non-apodizing leaving the aliasing errors of most modern source content in place which tends to manifest itself as over-bright harsh and edgy sound, which tends to be mistaken for “detail” (ESS Sabre and Cirrus Logic chips especially shine on this front). So you could likely find closed-form filters sound better since these are strictly non-apodizing, there is also poly-sinc-hb which is non-apodizing. By “xtr2” do you mean poly-sinc-xtr or poly-sinc-ext2? Former is almost non-apodizing while latter is apodizing.
But if HQPlayer sounds smoother it is indication that the the reconstruction is closer to the original smooth analog waveform with less digital artifacts.
sinc-M is not trying to be replacement or “sound like” Chord’s M-Scaler filter, it is just my take on a million tap filter, in my way.
With lot of modern recordings it is essential for the upsampler to fix the errors created by the ADC and/or mastering stage.
It was ext2 filter that I tried based off your recommendation. And think might not have been clear in how I’m hearing the sound. The sinc-m just didn’t have the texture that the mscaler had. The mscaler has no harshness or digital sound. So it’s the vibration of the cello strings thar I hear that sound smoothed over with the sinc-m filter. The ext2 came closer to mscaler in capturing this but still fell short for me.
I can’t really go into the technical aspects of any of this. This was just my own personal comparison using the equipment mentioned.
I think it is worth checking a non-apodizing filter like closed-form, poly-sinc-xtr or poly-sinc-hb for comparison, gives a bit of idea if it is related to that technical aspect. Or it could be the about million times (120 dB) higher HF leakage (lower attenuation) of the WTA1 filter, but for that aspect I don’t have a filter to compare against.
With classical it is usually hard to evaluate harshness directly, it is better demonstrated with rock drum kits. One of my test tracks is Porcupine Tree - Anesthetize (from album Fear of a Blank Planet), especially when the intro section ends at 6:20. The cymbals for example from position 7:50. They shouldn’t have any “messy glossed over extra sheen” but just clean and precise. And the sound shouldn’t become fatiguing even when played at high volume levels as it is supposed to. From 11:10 the song becomes relatively dense where you can test separation of things and the cymbal keeps clinging nicely in back.
Another test track I use is Opeth - Eternal Rains Will Come (from album Pale Communion).
Of these two, the Porcupine Tree is typical problematic digital track while the Opeth is clean of clipping etc. So good comparison in that way too. Former benefits notably from apodizing filters while with latter they don’t have much effect. (this of course applies to the CD rips I have, there may be different masters out there)
Both actually have the same guy behind - Steven Wilson. 
Steven… Prog god. I can’t count how many times I’ve listened to that track (Anesthetize). I used the ripped dvd version 24/48 with sinc-m and the cymbals retained their sheen and separation from the hi-hat throughout.
Hi Jussi,
I would be really interested in more of your test tracks, and what you look for when you are testing different things. This might really help me get my head around some of the different settings and it would be really enjoyable to do it. I have Qobuz, so I should be able to get at the relevent tracks.
Cheers
Ross
I have an extremely well isolated setup: digital source components in an RF isolation chamber; optical networking to a distant switch; optical signal to my DAC. I also have nearfield reference studio monitors (Focal SM9) and everything on batteries. I have gone to this extreme only to be able to really ‘hear’ the sonic signature of HQPlayer and hardware DAC combinations.
A/B comparisons are not possible when too much of the chain changes. You can’t compare Source->MScaler->DAC vs Source(HQPlayer)->DAC when the signal connection is galvanic and radiated RF noise (across open air) from the Source’s digital machinations impinges on DAC. You are then hearing the myriad subtle perturbations of the DACs final D/A: clocking stability and reference/ground fluctuations.
As an industry, we have seriously underestimated the impact of RF noise on a DACs ultimate transparency.
I have a Blu2/DAVE combo. Can I confirm what you are saying please?
If I feed the Blu2 a 705.6/768 signal via USB it will pass thru’ the Blu2 unfiltered?
Given that there are issues with the ground planes between Blu2 and DAVE (as admitted by Rob Watts and in addition the technical ‘issue’ that underpins the market for heavily ferrited BNC dual interconnects between the 2) would not one obtain comparable/superior results going NAA -> DAVE USB direct with the appropriately filtered 705.6/768 signal rather than ‘transparently’ thru’ the Blu2?
On a recent podcast Rob noted that in his opinion listening to 768kHz PCM was noticeably superior to lower rate ‘hi res’ signals
1 Like
Hello. The point I was making in regards to comparing HQPlayer vs FPGA WTA1 is to make the comparison using similar hardware paths. So if you send 44.1k via USB to your BLU2->DAVE from a source -vs- 705.6k via USB to your BLU2->DAVE from a NAA running HQP. Ideal would be the same source=NAA with HQP upsampling turned off/on. When BLU2 receives 705.6/768 from USB it just does pass-through to DAVE. I think Rob Watts does some attenuation of ~3dB but no other filtering is done. By going through BLU2 (vs direct to DAVE as you describe), you maintain a matched hardware path and your subjective comparison listening is valid.
I did such a comparison over a year ago using source->MScaler->TT2 using PCM poly-sinc-long-lp + NS9 and I thought it was close. Since then, Jussi has added a 2M tap filter (Sinc-L) and 15th order noise shaper (LNS15) and its leaps above the hardware WTA1. I am fully on board with Rob (and other converts) that a high tap count SINC filter that is appropriately noise shaped produces the most realistic portrayal of stage depth. Its not a fake echo/reverb thing, I think it best recreates the original analog …what it sounded like in the recording venue. IMO, HQP is pretty much end-game in this regard.
Now …please be aware that if you have a galvanic link between your BLU2 and DAVE (via metal coax cables, regardless of ferrites) then your comparison will be muddied by conduced RF noise. If you have your BLU2 and DAVE sitting in close proximity (less than 10’) then your comparison will also be muddied by radiated RF noise. With Rob’s DACs you are rewarded by pushing the boundaries of reduced RF noise from all paths.