Tír Na HiFi

jesuscheung wrote:

Aleg wrote:6.15 avx2 has vinyl-like tick.
...

have to agree.

i think that's because of the affinity tweak by mqncontrol. it alters the texture. restructures again by mqnplayer.

1. try restore all affinity of all processes, you would hear terrible texture. (texture is worse than any other software except for jplay which is unbelievably bad)
2. reenable the affinity tweak by mqncontrol, the sound is tidied up.

makes sound dirty and cleaning is not the way.

this is last time i say this, so tired of repeating myself.

JC

You don't have to repeat at all, because your remarks don't make any sense because you don't know what I'm doing in my setup.
I'm not looking for an amalgamation of distortions that make it sound 'nice' again, so I usually ignore your recommendations unless they make sense in some way.

My impression is that most of your tweaks are not repeatable except in your situation of listening directly from a mobo-dac and applying a whole load of 'tweaks' in combination.

So I'm very carefull in applying your recommendations as it is easy to mess up a good functioning/sounding setup.

But if it works for you, by all means continue your trials.

Cheers

Aleg

I'm interested in how far wdmks can be taken, think wasapi had reached a limit as 5.17 was best version, but some can't play it.

7.02 is quite good and hopefully can be made much better, it already has a more natural flow than wasapi, though the detail retrieval of 5.17 is very good. Can adjust the volume, so 7.02 not bypassing the mixer.

Using the memory play/minimalist exe of mqn made quite a difference, so good to have that validated.

Aim is to use portaudio to set up the ks pins and filters and to have the streaming code optimised and local to MQn.

thought wasapi sits on top of wdm/ks so how can an OS that can play wasapi not have wdm/ks ?

aleg, i bet you never even heard mqnplayer without any mqncontrol affinity tweak have you?
try it and comment again


----
don't see 5.17? or 6.17

Gordon, how about all sample rates milestone versions of one of the recent MQN's eg 6.15 or 5.14 to keep the music playing while the best KS version is being developed. I'm sure a lot of users of 16 bit MQN regard it as their player of choice but for 24/88 or 24/192 there's currently no recent player.
Anyone agree?
KM

Ken Moreland wrote:Gordon, how about all sample rates milestone versions of one of the recent MQN's eg 6.15 or 5.14 to keep the music playing while the best KS version is being developed. I'm sure a lot of users of 16 bit MQN regard it as their player of choice but for 24/88 or 24/192 there's currently no recent player.
Anyone agree?
KM

Agree.

Personally I like 6.17 SSE2 ( which I think Gordon is meaning not 5.17) out of all the recent ones and would really like 5.14 SSE2 to compare - as do several others, I gather.

Jonathan

sbgk wrote:I'm interested in how far wdmks can be taken, think wasapi had reached a limit as 5.17 was best version, but some can't play it.

7.02 is quite good and hopefully can be made much better, it already has a more natural flow than wasapi, though the detail retrieval of 5.17 is very good. Can adjust the volume, so 7.02 not bypassing the mixer.

Using the memory play/minimalist exe of mqn made quite a difference, so good to have that validated.

Aim is to use portaudio to set up the ks pins and filters and to have the streaming code optimised and local to MQn.

thought wasapi sits on top of wdm/ks so how can an OS that can play wasapi not have wdm/ks ?

This is great news. KS in theory has to be better than WASAPI as you have said a few times, as it is stripping out a layer of driver getting in the way (WASAPI that is).

Portaudio is also mature code and you used it to great advantage with JLP.

Thanks

Jonathan

jesuscheung wrote:i apologise for my recent comments

JC - thanks for the apology.

I do appreciate your energetic and enthusiastic approach to improving MQn.
Perhaps when you guess (wrongly in my case) that others are lazy or not prepared to put in effort themselves to contribute to better SQ, you get a bit carried away without knowing anything about what they have done.

Just to put the record straight, I have spent considerable time & effort over the last three years tweaking Windows XP (CMP/CPLAY), Windows 7, Windows 8 and WS2012 R2.
I have tried Minimal Server mode of WS 2012R2 but did not find any appreciable benefit and certainly some downside. So reverted to GUI mode. I applied the CAD Optimisations which were of great benefit.I have applied several other WS2012 OS tweaks. My DH61DL BIOS though offers no ability to tweak it at all.

My view is that when you get to a certain stage of tweaking the OS, the benefits are outweighed by the risks of making it worse and the time and frustration in trying is certainly not worth it.
Putting it another way, as I said before, 'The Bastard' Windows is such a complex (and unsuitable for audio) OS, and the permutations to be tried run into millions.
Also, as you have demonstrated, it is mostly if not entirely trial & error, as predicting outcomes is impossible.
When you add in the different code behaviour of MQn and its interaction with the Windows OS, the situation gets beyond comprehension.

As I posted, on my set-up, I am finding 6.17 SSE2 truly remarkably good. So good, that I am willing to spend my time listening to music rather than spending fruitless time chasing that elusive Registry change that might make a positive or a negative effect.

Ken Moreland wrote:Gordon, how about all sample rates milestone versions of one of the recent MQN's eg 6.15 or 5.14 to keep the music playing while the best KS version is being developed. I'm sure a lot of users of 16 bit MQN regard it as their player of choice but for 24/88 or 24/192 there's currently no recent player.
Anyone agree?
KM

Ken
Completely agree.

It is a shame the 6.15 to 6.18 can only play with ticks on some/most setups (appears to be again a difference between Amanero drivers and Thesycon-drivers).

I would need 32-bit container versions 24/44, 24/88 and 24/96 as a minimum for playback of most of my music.
Higher sample rates would be welcome but not strictly necessary for me.

And 5.14 is the only one that qualifies IMHO.

Cheers

Aleg

Ken Moreland wrote:Gordon, how about all sample rates milestone versions of one of the recent MQN's eg 6.15 or 5.14 to keep the music playing while the best KS version is being developed. I'm sure a lot of users of 16 bit MQN regard it as their player of choice but for 24/88 or 24/192 there's currently no recent player.
Anyone agree?
KM

By all means! I still play all of my music converted to 24/96 with MQN 3.39, as there have not been any recent hires versions that better it. I suggest a full range of versions on the basis of a recent MQN like 6.15 or 5.98, which indeed sound detailed and balanced, and yes, 5.14 as I often tend to agree with Aleg's perceptions. I would love to have AMD AVX 24/96 versions of these two.

I think it would be a good thing to kind of "complete" this long Wasapi MQN quest with all of us getting one or two milestone versions in the needed format for use as reference in the next expedition into KS.

jesuscheung wrote:aleg, i bet you never even heard mqnplayer without any mqncontrol affinity tweak have you?
try it and comment again

Of course I have, and I prefer to have mqncontrol on its own core as well, though the effect of this isn't as marked as with mqnplay on its own core.