Will your new DAC accept higher DSD rates than 128?
I will perhaps in a few weeks get the Phison preamp with the new dac card also based on ESS9028 PRO and that will accept DSD 512 over usb.
Your impressions regarding DSD from Roon is something that sounds really intresting as I asume Roon upsampling to higher dsd rates is less demanding on the computer than HQ Player?
Firstly, I think you are right about attack with PCM. I used the term weight because I hear it most in the mid-bass, but it is the leading edge that we are both talking aboutā¦
I have listened now for good long periods upsampling Redbook to:
DSD 512 (8 x 48), xtr 2s, AMSDM 512 fs+ in HQP
PCM 384, xtr, NS5 in HQP
PCM 352/384, Precise Linear, 7th Order CLANS in Roon.
These are all linear phase filters and the first surprise for me was that I generally preferred linear phase over minimal phase. I do like minimal phase for tuned percussion (Chinese gongs, vibraphone etc.), but otherwise I find the soundstage more focussed with linear phase.
The second surprise was that I like long filters. Xtr are the longest that HQP does and some of the longest you will find outside a Chord. I donāt appear to be troubled by pre-ringing and I suspect that I like the euphonious thickness of ringing generally. Long filters enable steeper band rejection and it seems that I prefer ringing to higher leakage of ultrasound.
In terms of comparison, I tend to leave it on one of the above for lengthy periods of time rather than change it around. I think, however, that each is suited to different material as follows:
is my current setting (again). This has the best detail and expansive open soundstage but it loses some impact on the attack. Itās not my favourite for classical or percussion for those reasons. I like it for serious listening to most modern music, female vocals and especially for acoustic guitar (which seems counterintuitive, but there you go);
is good for classical and percussion ālisteningā. Better attack, slightly narrower and more forward soundstage;
is more like a āparty modeā. Good for pumping dance music, male vocals, upsampling internet radio and when playing background music socially. It may be that others would be happier with a shorter minimal phase filter for those tasks, but you can only take so much of my soundstage away before I get quite bolshie.
Edit: Also thanks very much for linking the Dustin Foreman article in this post.. I think itās the best explanation of filters that Iāve read.
DSD vs PCM. DSD is thicker, slightly less precise and more relaxed for me, but Iām only working working with DSD128. PCM seems more detailed to me but that could just be more high end energy. PCM sound stage is more precise and everything has more punch and impact.
Iām going to stick with PCM->PCM via HQPlayer for a few days and see how I get accustomed to it.
Very interesting discussion. Until I got my new system dialled in I preferred upsampling to DSD 512 in HQ Player. I liked the smooth analogue sound.
However now that my new system has finished burning in (with some new NOS tubes) I prefer upsampling PCM to PCM384K using poly sinc xtr (when upsampling to DSD512 I am limited to the -2s poly sinc filters due to my PC)
Both options sound amazing - I just prefer the extra bite staying with PCM and the xtr filter.
But in a harsher/brighter system I could easily imagine preferring DSD. I suspect the best solution is system dependent.
I like the sound of the poly-sinc-txt filters very much, but I noticed I would experience frequent dropouts or extended periods of silence with those filters engaged whenever my VPN (Netshade) was active.
I used the xtr filters, both for pcm and dsd (128). I play from audiolinux to microrendu to lampizator amber.
Usually I was satisfied about the sound, apart from high and forte string passages in classical music. The strings seem to have a sort of phasing sound. Of course this phasing effect is naturally due to the different intonation of the players, but I felt it as more than normal.
Yesterday I experimented with the FIR , the MinringFIR and minringfirmp filter. For PCM I used the gauss1 as dither and upsample to 88.2.
Are these FIR and minringFIR filters actually okay. They sound better and more transparent in my ears. And the phasing sound of the strings seems more into natural proportion.
Does anyone use these filters, I dontāread anything about them.
I have certainly seen users comment that they prefer the minring FIR filter. If you imagine a notional scale in HQP with (short, minimal ringing, shallower cutoff) at one end and (long, ringing, steeper cutoff) at the other then the minring FIR filters are towards the short end (polynomial is shortest) and the XTR filters are at the long end.
A shorter filter with less ringing yields a higher fidelity transient response. A steeper cutoff pushes out of bandwidth noise (antialias or DSD artifacts) lower which I believe results in better preservation of spatial cues. So ringing is the price we pay for steeper cutoff filters.
The HQP filter palette has a wide variety of different compromises between the various constraints imposed by mathematics. I understand Jussi listens to poly-sinc-mp as his preferred filter. There isnāt a lot of writing about the sound of the filters because people are still discovering what they like and are hesitant to give opinions. There are no musically right or wrong filters, just different compromises. Similarly to speakers. Sometimes people prefer a euphonious sound to fidelity, I suspect I am one of them.
Thanks for the insight about a phasing sound in the forte strings. I think this could be a consequence of ringing.
Even though no one has posted here for awhile, I have actually read this a few times recently for ideas on HQP filters.
FWIW, I just upgraded the firmware on my Lampi Atlantic (or, i should say Lampi did) so that it could play DSD 512 with my Microrendu. Using the same xtr-2 filter that i had used in 128, the sound was initially very exciting, but then quickly turned very fatiguing. Very. It seemed like the ringing created by that filter was more audible. Gave me quite the headache.
So, that got me playing around with the filters. I am currently spending time with the minringFIR. I am really impressed. Amazing separation between instruments and vocals, very natural sounding, great presence and quite punchy. All without the fatigue.
Another vote for minringFIR. I will give it another listen.
At the moment Iāve been taking a holiday from DSD and listening to PCM 768 using the xtr filter which I find quite enjoyable. However Iāve also been mainly listening āon the goā while doing something else in another part of my apartment (cooking, cleaning, reading etc.) so I will have to sit down and compare things properly. Maybe I will end up with different configurations according to what I am doing.
Volpone. Just last Friday, Amanero released the combo384 firmware update that permits use of Linux NAA to play DSD 256 and 512. No issues whatsoever with the update, has played glitch free since last Saturday.