I am finally getting to try a few different filters now my UPS and HDPlex PSU are in place and running powering my Core, Switch and NAS. I mostly listen to electronic/rock rather than anything classical so I tend to go for minimum phase filters based from what I have read previously.
I am currently using poly-sinc-short-mp, DSD7 256+FS, upscaling to 44.1 256 DSD to my Mytek Liberty DAC. Sounds really good.
I am not sure I am the best at trying different filters though. I tend to jump around quite a bit, listening to different tracks when I probably should stay with a select few and I seem to have no aural memory but that is probably due to my terrible testing methodology!!
I updated to HQ Player Desktop 4.1.1, and this is the first time I have gone through and updated since buying HQ Player earlier on in the year. I just uninstalled on my NUC, and then ran the newly downloaded version. It automatically registered, so all good.
Anyway, I finally had a go with the new EC variants which @dabassgoesboomboom alluded to. I tried both the ASDM7EC and ASDM5EC and both cause my NUC i7 to jump to 100% utilisation and although music will play it jumps, so I am back on ASDM7 for now, still sounds good.
I wonder if I should start thinking about another server machine to replace the NUCâŚI might just build a desktop PC this time and not go the NUC route, probably windows 10 to run both HQ Player and the Core.
To run the new EC modulators at DSD256 with ext2 filter, none of the NUCs will cut it. My previous NUC could do it at DSD128 but no chance any can do DSD256.
For anyone building/buying a new PC, i7-9700K and i9-9900K are good options.
Thereâs not a lot of feedback from HQP users on the AMD chips so far (specifically with new EC modulators at DSD256) so Iâd stick to these Intel chips.
My i9-9900K works really well, running both HQP Embedded and Roon Server without any hiccups. Runs cool too, stays under 60 deg C at all times.
I recently purchased an mscaler and had been using HQP. So for fun I ran at the same time:
roon-wifi-mbpro/hqp: poly-sinc-xtr, no dither, 768 to Hugo2 usb to utopia and
roon-wifi-digione-mscaler-dual coax(DX) to Hugo2 to utopia
after syncing up the source and switching between usb and coax in on the H2 I listened to an entire album.
By the end of the album I had no idea which I was listening to and had to refer to the lights on the H2.
To my non golden ears they sounded practically identical. Of course the hardware solution has proven much more stable and glitch free.
You could also try the sinc-M filter (sinc function with 1 million taps) and poly-sinc-ext2 filter - both high quality linear phase filters. With Gauss1 or TPDF dither.
What glitches did you experience?
I run HQP Embedded on a headless (no keyboard/monitor/mouse) machine, which also runs Roon.
Never had a glitch since I (finally) setup everything (server, NAA/endpoint, network) properly âŚ
If you can share some info on the glitches, we can try to help.
That was my fear! I have just been reading a thread on audiophilestyle which basically says the same thing, in fact, I think I recognise you on there too!
I have outlayed quite a bit lately - new speakers, stands, HQ Player, Home Audio Fidelity Filters, acoustic panels, UPS, HD Plex PSU and my next planned purchase would be a new streamer, one with NAA in-built. Given this, I am wondering if there is another way for me. The desktop PC I am using now is a i7 9700k (o/c to 5ghz) and so I am wondering if I should look to shift HQ Player to this machine although I am not sure that is possible?!?. The original idea of the NUC was so I could use it without having to use my desktop but I have built this PC quieter with water cooling so it wouldnât be too bad to leave it on.
I was pointed to poly-sinc-xtr by another user as the closest to the Chord filter for the test. I have tried many of the HQP filters over the year I spent with it.
The glitches during my year of use were usually on startup as the mbpro was over wifi. Once running it was usually glitch free.
The mscaler is definitely the best option for me as it is less hardware and better foot print. I was just surprised how close it was sonically to my old ears.
I had asked on several forums for a comparison between HQP and Mscaler to chord dacs but never got a reply ecept from Miska who said not to use with Chord.
I had the hardware so I did the A/B and was surprised at the result.
I donât want to go off-topic for too long but just briefly, does the M-Scaler add much difference to your Chord DAC alone? Or the difference is small?
I demoâed an M-Scaler with both Dave (a Blu2) and later with my Hugo2 and didnât hear a significant difference. The small difference I heard I actually preferred without M-Scaler. Just my own experience.
I might do that, however I am just trying to decide if there are any downsides to moving the licence before I contact Jussi. The only negative I can see is needing to leave on my main PC when I want to listen in the lounge, in addition to the NUC which is an upstairs cupboard. The plus side, the desktop PC is powerful and well cooled so should be fine for the task, it probably need a new graphics card, but I could offload some to my current nVidia 970. The NUC should also run cooler, so I could adjust the fan settings if I wanted to.
I just wonder if network gremlins would be better or worse in terms of HQ Player âseeingâ NAA on the USBridge, but I guess I wonât know until I try it. hmmmm.
Well that was a fun 24hrs! I downloaded a trial version of HQP Desktop as discussed above, so far so good, however I was still struggling to get the software to see the endpoint as normal with my network. Anyway, fiddling around, and then realised I was suffering from drastic fluctuations in internet speed from 1 meg one minute to 80 meg then next and back again. It affected the whole network and after a conversation with my internet provider they have sent 4g hub and an engineer visit next week. However I preceded to take the network apart last night testing as I went looking for the culprit, just in case the problem was my end. After much switch and cable testing I finally narrowed it down to my new ups. It has ethernet protection function (I wasnât sure about it when it arrived), basically you run your cable connecting it to the ups and when you do this you should be protected in terms of surges etc. Well, it destroys you internet connection too! So, after removing that step, the network is back and running, anyway, I digressâŚ
Back to my trial copy of HQP, I managed to get it to find the Allo USBridge and I set-up ext2 filters along with ASDM7EC. Running this on my main pc, i7-9700 with 16gb of ram and a nvidia 970, I could play music but I was getting small drop-outs. However the PC is âonlyâ running at 75% ish, so there looks to be some head-room. ASDM5ED runs no problem. Hmm, any ideas?? Maybe the 970 doesnât have the power for the CUDA offload.
Yes, it is the same if I uncheck the CUDA offload. The PC is about 75% used at around 4.8ghz, so their does look to be some headroom. No problem with the ASDM5EC.
Are you using multicore in auto mode? Greyed rather than checked. My system is far below a 9700k but multicore auto is the key to success with EC modulators for me, especially after the recent round of optimizations. Certainly a GTX970 could bottleneck such a powerful CPU but I canât imagine it not being up to the task even without CUDA.
Hi @lorin - thats it! I did have the box ticked rather than greyed. When its greyed the CPU is around 40-45% load running poly-sinc-ext2 and ASDM7EC. If I then grey the box for CUDA I perhaps gain a percent or so, but probably not worth having it on to be honest. However I am running custom filters from Home Audio Fidelity (x4) via the pipeline, so I am not sure how that plays into the equation really.
Awesome, glad to hear it. If you ever try closed-form filters, the 970 would come in handy for that. CUDA offload is extremely efficient for that family of filters.
