But seriously what do you expect to learn from ASR community in regards to HQP? SNR measurement on selection of filters we rarely use with conclusion that all HQP benefits are in ultrasonic range and there’s no sense in using it? And certainly test with Topping, most likely on ESS chip, that would not make sense.
You may also search ASR to see a huge Jussi’s disagreement with their key measurements, saying it’s not representative enough and they are not taking bandwith into account. In return community picked on Jussi as he does not want in his own full right to disclose his measurement rig, so ASR says his measurements are not credible… frankly its just funny because ASR just picking on words and not taking into account the core argument. Quick summary from ASR - oversampling is stupid and not needed.
If you search on the net for “archimago hqplayer” you will find a very decent review from Archimago with measurements already, And if I’m not mistaken his measurements has been also discussed on this forum. Quick summary from that - HQPlayer does quality job (especially with DSD/SDM)
And for the difference in 5EC vs 7EC vs 5ECv2 and 7ECv2, well I think non-bias blind testing with sufficient statistical data may answer the concerns for those who doubt that there is a difference.
PS: Saying above I have a huge respect to the work that Amir is doing, it is immense, I have rebuild my system pretty much completely based on his research and I’m “sponsoring” ASR, but you really need to take some things from them with a pinch of salt,
EDIT: Here. I found discussion on this forum already
Congrats on this! I envy you for many wonderful discoveries ahead with all sorts of filters, modulators and rates !
You may want to check a link I have just posted above to Archimago’s research on HQPlayer, he visually shows the difference in DSD64 vs 128 for example. With 128 being undoubtfully superior in many senses.
Hehe, same here. I asked at their forum how they could consider it science when there is no proof that everything that we hear can be measured (which is a base requirement for drawing any sort of conclusions from measurements).
Agreed, I don’t mind the measurements either, in fact, they are very important in a lot of design and manufacturing of audio equipment. But Amir’s conclusions that he draws from some simple 12khz jitter measurements are often laughable.
And how many times have we read from Amir “There is some [insert problem here] but its not audible”. Audible for whom? In what system? In what room? Those kinds of statements are the very opposite of what real science is.
Amir’s view tends to be that there is one relevant set of measurements, perform well on those and that’s it. Jussi and HQ Player shows us a) that so much of digital audio is about compromises or tradeoffs and b) that how we each are likely to choose differently when we, not the equipment designer, is allowed to choose which tradeoffs we prefer.
Photoshop taught us that about digital photography – my preferred level of sharpness and saturation may be quite different than yours, but I still want a camera that provides me with files that represent the maximum dynamic range and number of pixels (captured through high quality lenses) so that when the software goes to work it has great underlying data to work with.
I see how HQ Player operates with digital music data the same way – you want to feed it with data captured through the best hardware (DAC,power supply, nrtwork) you can start with and then let the software do its magic.
” And for the difference in 5EC vs 7EC vs 5ECv2 and 7ECv2, well I think non-bias blind testing with sufficient statistical data may answer the concerns for those who doubt that there is a difference.”
Have you seen or done such a test? I have never seen any type of blind testing performed on HQ Player. I use HQ Player myself, and I have “decided” that it sounds good, but knowing how easy it is to trick oneself I’m still a bit skeptical. I’ve raised the question before in this thread, where people rave over huge differences in filters and modulators. I’d like to see them pick a winner in DBT…
This composition “Journey of Life” starts with a combination of drum and voice with middle east/north african type of rythm. My ‘default’ playlist setting was ‘gauss-xla’ and I cought my self that it did not sound punchy enough. Next I tried “ext3” and for this composition it was noticably better with drums getting more ‘body’. But the winner was Sinc-LL
The realism of the drums in the above combination is insane.
To your earlier comment - could I guess the difference blindly? I don’t know… but I invite you to give it a try, especially the difference between ‘gauss-xla’ and ‘Sinc-Ll’ is very pronounced, i think…
Interesting. Have you tried the poly-sinc-gauss family? In my case, I have 1x = poly-sinc-gauss-xla, Nx = poly-sinc-gauss-hires-ip. I listen to lots of modern jazz, including other Tord Gustavsen albums, and I don’t feel a lack of percussion punch. Just listening to that track (thanks Qobuz) and the drum/voice balance feels just right, the drum overtones and decay complementing the voice texture so organically. But it could also be down to differences in gear. Setup: NAA endpoint>Intona isolator>Holo Spring 2 KTE>DNA Stellaris Special>ZMF Atrium LTD (koa).
BTW, if you are enjoying that track, I suggest you take a listen to Amina Alaoui (Alcantara, Siwan, Arco Iris) or Dhafer Youssef (Birds Requiem, Sounds of Mirrors, Diwan of Beauty and Odd).
This was not to say poly-sinc-gauss family is not good, it is in fact great! And without knowing that you can get a tiny notch differences - “xla” would’ve been perfect! In a last few days I’m driven away by testing my new discovery, that is DSD@1024. My comparison over that particular composition happened between the following:
poly-sinc-gauss-xla with AMSDM7EC 512+fs@1024x48
poly-sinc-ext3 with AMSDM7EC 512+fs@1024x48
sinc-Ll with with ASDM7ECv2@256x48
So not only the filters, but also SDM rates and modulators. For the other compositions on the almum the difference is less striking if anything. I was also surprised that on my system I was not able to get stable Sinc-Ll@512, it took 11G of RAM, btw.
But, voila, i’m still convinced Sinc-Ll is a winner in the above three choices (ok, not really apple to apple, but still…) even @256. I may revisit it later today dropping ‘gauss-hires-…’ in a mix out of curiosity and also comparing at different rates. But this is the whole point - there are sweetspots and when you can find one its a bliss!
Another difference in what we hear may be the end-point. I listen on a floorstanders ProAc D20R. Room is with out correction so suppose I have a room factor as well.
But I would still insist that in “Journey of Life” Sinc-Ll gives a punch!
Thank you for this tips! Lately I’ve been listening to a lot of ECM catalogue. I will definatelly listen to these. I would throw also Anouar Brahem ina mix!
EDIT: just tried, non scientific, -gauss-hires-lp is not too bad at all !!!