JLP wdm-ks player

[sbgk]

Now we have a name for the modified squeezelite and portaudio software (fully compliant with licencing, in case anyone asks), I thought it would be a good idea to split it out from the MQn thread.

[sbgk]

I can confidently say that 13/14 is the best version so far, at least on my system.
I can hear for the first time benefit of 24/192. Meet me London is like a dream.
Gordon, I don't care what they say about you, but you are the genius of Computer Playback...LOL.

did you try 13/13 ? It has more oomph. 13/14 seems a lot quieter and needs the volume turned up, so am undecided about whether the change I made to PA is the right thing to do. The way they coded it it only happens for ds, asio and wasapi and I have removed all of these, so is it required for ks and they just didn't code it because it was included in ds/wasapi ? more investigation required. If I put it in sqeezelite there is a different sound again, a bit more refined than 13/13, but not as quiet as 13/14.

Don't know about genius, there are much more capable and cleverer developers out there, but they tend to go by measurements/functionality rather than ears, but tweaking pa/ squeezelite to make the JULF player seems an obvious thing to do, but no one has tried it before, so to some extent my efforts are pioneering/opportunistic.

[Julf]

fully compliant with licencing, in case anyone asks

Very good to hear! Thank you!

How can I get the source code?

[Aleg]

fully compliant with licencing, in case anyone asks

Very good to hear! Thank you!

How can I get the source code?

Did you check the google drive refered to?

-
Aleg

[nige2000]

got to say the latest version sqlite v13/pa v14 of the JULF Player is the best thing I've ever heard.
Couldn't have imagined that that amount of detail and sq would be available from a windows streaming player playing 16/44.1.

v14 is fabulous, I know all depends on many things (gear, interconnects etc.) but my #1 is opt100x0 - I can hear more details (not noise). v14 is a great alternative (more open soundstage and analogue but still lacks some of my beloved details). That's just my opinion.

OK
im guilty of not trying every version but i preferred 13/13 over 13/14 just because 14 is flat
more importantly 10/10 0x0 is far more detailed than 13/14, 13/13
i have missed a few versions in between so i cant speak for them
absolutely certain im not listening to just noise, it is unmistakable
still not as sharp or crisp/clear as 2.7x but we will get there.

[Julf]

Did you check the google drive refered to?

I have now - and I am very happy to see that the source code is now provided. Many thanks, sbgk!

[sbgk]

[quote="nige2000
OK
im guilty of not trying every version but i preferred 13/13 over 13/14 just because 14 is flat
more importantly 10/10 0x0 is far more detailed than 13/14, 13/13
i have missed a few versions in between so i cant speak for them
absolutely certain im not listening to just noise, it is unmistakable
still not as sharp or crisp/clear as 2.7x but we will get there.[/quote]

Thought 10/10 0x0 was maybe too 3d, bloated in the soundstage, whereas 13/13 sits back a bit more in the speakers.

14 is very flat, what is going on to produce such an effect, produces an alternative view of detail though.

Think ks has a smoother delivery, so the sharpness you are missing may not return and I think it was possible an artifact of wasapi, which is just a wrapper for ks. The minimum latency I got with wasapi was 3ms, ks is running at 1ms.

The streamer functionality limits the ram play so may end up doing an MQn version that uses pa. Not sure how much the layers of code pa has on top of ks affects the sound. Still, early days.

[nige2000]

Thought 10/10 0x0 was maybe too 3d, bloated in the soundstage, whereas 13/13 sits back a bit more in the speakers.

14 is very flat, what is going on to produce such an effect, produces an alternative view of detail though.

Think ks has a smoother delivery, so the sharpness you are missing may not return and I think it was possible an artifact of wasapi, which is just a wrapper for ks. The minimum latency I got with wasapi was 3ms, ks is running at 1ms.

The streamer functionality limits the ram play so may end up doing an MQn version that uses pa. Not sure how much the layers of code pa has on top of ks affects the sound. Still, early days.

the 13/13 is very very smooth which i can see why some would like it, for me the bass isnt as tight and the resolution/detail not near as good as 2.7x which seems to have a pin point accuracy in comparison that makes it seem live/real
10/ 10 0x0 seems more in the right direction for me and obviously john dot too

i remember in jplay terms ultrasize used to help with this

[Clive]

I must admit I've not been playing enough attention recently. What's the strategy now? Is MQN on the backburner due to the Squeezelite software using KS and therefore having the potential to sound better?

[sbgk]

Thought 10/10 0x0 was maybe too 3d, bloated in the soundstage, whereas 13/13 sits back a bit more in the speakers.

14 is very flat, what is going on to produce such an effect, produces an alternative view of detail though.

Think ks has a smoother delivery, so the sharpness you are missing may not return and I think it was possible an artifact of wasapi, which is just a wrapper for ks. The minimum latency I got with wasapi was 3ms, ks is running at 1ms.

The streamer functionality limits the ram play so may end up doing an MQn version that uses pa. Not sure how much the layers of code pa has on top of ks affects the sound. Still, early days.

the 13/13 is very very smooth which i can see why some would like it, for me the bass isnt as tight and the resolution/detail not near as good as 2.7x which seems to have a pin point accuracy in comparison that makes it seem live/real
10/ 10 0x0 seems more in the right direction for me and obviously john dot too

i remember in jplay terms ultrasize used to help with this

still have a feeling that one or more of the optimisation settings is causing the smoothness, just need to work out which it is. Getting the affinity set will improve things.

Looks like pa wdm/ks was designed to work with mmcss and if mmcss is unavailable it sets the sytem timer to 1ms (which I have removed). It uses the mmcss pro audio setting. I thought the sound without mmcss was a bit thinner sounding on win 8.1, at least.