Which HQP Filter are you using? [2024]


DAC correction on Spring 3 KTE
Using non 512+ ASDM7EC-fast for DSD 256
Overall nice for low volume listening

Great for Redbook resolution rock and progressive music :musical_notes:
DAC correction :level_slider:on

3 Likes

I find this pair matches well for Rock, Pop, fast transients for Redbook and High Resolution material:

3 Likes

Hey @jussi_laako

Happy holidays. Do you have any block cycle recommendations for 9800x3d? Manual says higher is better since i have the 3D Cache. Does the max option of 16 sound right?

I have CUDA offload checkmarked using a 5070ti. Using default for CUDA as I briefly remember reading it is recommended over selecting the exact GPU?

Cheers, happy holidays and than thank you as always.

1 Like

Happy Holidays!

The automatically detected value is probably good, but you can also try with 16 and see if it drops your loads without causing drop-outs.

Yes, it will automatically pick your GPU. If you have only one Nvidia GPU it will anyway pick the same GPU regardless if you explicitly select the GPU or use “Default”. Selecting the GPU explicitly only makes difference if you have more than one Nvidia GPU and want to select which ones are used.

2 Likes

For whatever reason (HQPlayer updates or MacOS Tahoe but NOT any action on my network) I can now enjoy AHM7EC8B @ 1024 without (or very rare) static as per packet loss. Best sound ever with my May. Kudos and thanks

2 Likes

Congrats!
Probably a network setting got changed, in this case for the good.

Now don’t change anything nertwork wise! :laughing:

That 1024 setting is different, soundwise, i found it works with certain genres better than others. I still prefer 256dsd.

thank you. Which genres ? soundwise it’s different from my souvenirs (and I could have minutes without glitches) : might be different. I remember a beefy sound, while now I enjoy a tight precise sound with impressive soundstage. Most impressive with a pianist’s left hand that gains focus and energy

I kind of feel, and I think I may be repeating similar experiences by others in saying, the higher the DSD rate the softer the sound. It’s by no means bad or worse.

I found that I liked 1024 with classical more than other genres. As with almost all things hqplayer, YMMV.

beefy, softer, “analog”, “organic” etc yeah that might have been the appeal of AHM7EC8B when I first tried it. It now sounds quite the opposite and is quite mercilessly vivid (closer to live sound). I have often wondered if the taste for filters that would drive CPU/GPU combos to their limits comes not from the filter itself but from the hard driven computer sort of making a paradoxal MP3 effect. Could explain the difference between my perception now and then.
Alternatively, AHM7EC8B @1024 could be no better or that different than 7EC Fast and what I enjoy at the moment could only be due to a low degree of hygrometry thanks to the cold and dry weather (and heating). A coincidence then. Anyway, AHM7EC8B works fine and I wanted to thank Jussi.

From what I remember, AHM7EC8B measured very flat. It may be one of the best choices for 1024.

I shared a measurement for Spring3 in the dsd direct thread, some week ago. It gives DSD1024 very good performance under 100kHz

Then the (fixed) DSD noise sky rockets. But it is above 100kHz so user can be the judge about if this is audible lol

But its a great demonstration of how much the 1-bit modulator can affect measurements

1 Like

But it is still over 10 dB less than for example SMSL D-6 puts out with regular modulator at DSD256
 And it drops faster, by 1 MHz it is already in the analog noise floor, while D-6 at DSD256 reaches analog noise floor by 5 MHz.

3 Likes

True but D-6 has been surpassed by a few newer affordable products. Plus at least when I last measured, DSD direct looked just like non-direct on the Pico, which it shouldn’t be. Maybe just my unit?

SMSL D200 is better! :grinning_face_with_smiling_eyes:

I don’t think AH90 was any better in that respect at DSD256. I don’t know how D200 performs. This thread is too long to find anything
 :sweat_smile:

With AHM5 the ultrasonic peak noise would be probably below level of inherent noise floor of what RedBook has in audio band.

May be time to start a 2026 thread.

3 Likes

I concur! 2026 thread time :smiley:

Quick one for @jussi_laako

After 6 months of DSD filters I’ve switched to PCM on my Holo May and it just feels better. I have Graham speakers and Bryston amplification and they’re very analytical so perhaps too much of a good thing.

Anyway I’m really happy with poly-sinc-xtr-short-mp with LNS15 at 1536 KHz

When I first initiated this it took ages to get going. Now it just changes tracks with no delay at all and doesn’t see to be taxing my Mac mini

I want to repurpose the Mac and get a more basic one still M4 but I’m confused. Is this demanding filter or does it just do some mojo when it sets itself up the first time?

Thanks!

Remember to set “PCM gain compensation” in HQPlayer to -6 dB. Otherwise your PCM will be notably louder (6 dB) than DSD, and louder can make it sound “better”.

Some of the filters take some time to initialize. HQPlayer caches some of the initialization data in RAM, so subsequent starts with same conversion ratio are faster. As long as HQPlayer is not restarted.

Thank you.

Yes I have PCM Gain compensation set correctly so there’s no volume difference.

I have the latest Mac mini M4 Pro with 48GB RAM. I can run poly-sinc-xtr-short-mp on PCM no prob, and can run that on DSD with 7EC Super for high res content, but for rebook quality it hangs, the memory usages slowly increases, and after 5 mins I force quit. I can run -2s version for redbook

Should I be able to run the redbook one without the -2s?

Also, what’s the best modulator for this / closest to the PCM sound?

I do hear a difference of course, the DSD can be a little too glare or sharp if that makes sense.

Ta

Simon