Tír Na HiFi

Aleg wrote:

sbgk wrote:....
Just to recap there is now nothing between the source data in ram and the device driver, so there's little that MQn is doing to influence the sound. It's an interesting insight into what affects the sound and largely confirms my thoughts.

The rest is down to your system and the choices you've made.

This suggests that every player that uses KS would sound the same because "there is now nothing between the source data in ram and the device driver, so there's little that MQn is doing to influence the sound.".

So I can go back to Foobar as along as I use KS?

don't know if you're being deliberately obtuse, but use whatever floats your boat.

Aleg wrote:

sbgk wrote:....
Just to recap there is now nothing between the source data in ram and the device driver, so there's little that MQn is doing to influence the sound. It's an interesting insight into what affects the sound and largely confirms my thoughts.

The rest is down to your system and the choices you've made.

This suggests that every player that uses KS would sound the same because "there is now nothing between the source data in ram and the device driver, so there's little that MQn is doing to influence the sound.".

So I can go back to Foobar as along as I use KS?

I am not a coder, but I do think your assertion is a leap too far. Not all coders are equal - I can tell you that having employed many of them at our software company! Also there are many ways to skin a cat and as we all have heard with our own ears, one seemingly trivial code change or even compiler setting can affect SQ in a material way.

The fact that Gordon has been ingenious, clever, studious, persevering and downright focussed on one outcome, which he has now apparently achieved though not listened to it myself yet, is a testament to the individual. Whether the Foobar collective, JPlay or any other KS players coders have been so successful in achieving that direct commection between RAM and the device is open to question. I would doubt it, but perhaps you could enquire of the Foobar Community and let us know their answer?

Putting it another way, how do you compare Foobar KS to MQn latest version in SQ? The outcome of that comparison should guide you to answering your own question.

Aleg wrote:

sbgk wrote:....
Just to recap there is now nothing between the source data in ram and the device driver, so there's little that MQn is doing to influence the sound. It's an interesting insight into what affects the sound and largely confirms my thoughts.

The rest is down to your system and the choices you've made.

This suggests that every player that uses KS would sound the same because "there is now nothing between the source data in ram and the device driver, so there's little that MQn is doing to influence the sound.".

So I can go back to Foobar as along as I use KS?

The way I see it...I could easily be wrong....MQn is now pretty much "blameless", the same cannot be said of Foobar which will have negative effects. If Foobar could use v51 then maybe it would be improved but it would still be stymied by the "blamefull" Foobar code.

Clive wrote:

Aleg wrote:

Clive wrote:Unfortunately the 5.14 comparisons are a bit esoteric as not many can run it.

I don't believe anybody had a problem with 5.14 apart from those with old mobo/cpu noy supporting avx.

But it is not esoteric because sq is still superior to the rest of the 100+ versions that came after

Around that time I couldn't run any AVX but now I can for some reason. If 5.14 will run on my ivy bridge I'll give it a go except it's not in the archive, unless I've missed it.

Clive check back a few pages I think Cvrle59(George) gave a dropbox link for it.

The way I see it wasapi is an api that sits on top of ks and allows an easy programming interface to render data, so not only do you have wasapi overhead, but also ks as well, hence why wasapi latency can't get below 3ms.

normal wavecyclic method of processing is to load data into a buffer and then submit that data as packets to the output pin. MQn now instead of having that buffer and load process just passes the ram address of the loaded PCM data to the output pin.

I guess this could be done by other players, but it does require a ram player of the type of MQn where all the data is loaded/converted into ram first.

Typically a player would convert formats etc on the fly which requires buffers eg portaudio converts from 24 bit to 32 etc at the point of loading the data into the buffer.

The normal method has several packets to help smooth out latencies in the buffering and MQn still thinks it is dealing with 4 packets. If I can get this down to 1 then it may improve things as the data is already there in ram. Also need to understand more about how the audio device interacts with the player in terms of timing.

Potentially you could get lower latency by setting the buffer size for your device as that will in theory be the size of data the driver gets from the pin, not sure how that works yet, though - the amanero driver doesn't have any settings so maybe time to get the mf vlink back up and runnning.

Still lot's to learn and also think there would be an improvement if the portaudio dll was removed as dlls always seem to soften the sound.

Running 7.36+v51 with Tasker I find the sound very clear, clean and natural. Very 3D with instruments hanging well in space. I've no qualms about balance bass/mid/treble. Decay is excellent and timing pretty good too. These comments refer to an untweaked laptop running Tasker with no internet access. I've searched for the link to 5.14 but haven't found it so I can't make any comparison with that.

Clive wrote:Running 7.36+v51 with Tasker I find the sound very clear, clean and natural. Very 3D with instruments hanging well in space. I've no qualms about balance bass/mid/treble. Decay is excellent and timing pretty good too. These comments refer to an untweaked laptop running Tasker with no internet access. I've searched for the link to 5.14 but haven't found it so I can't make any comparison with that.

Gordon - thanks for the really clear and helpful explanation of WASAPI v. KS when it comes to MQn. I thought I understood the difference before, but now I realise I didn't really, but now I do.

Aleg should now be aware of what you have done and can find out if Foobar (Full memory play) can and does do the same. If so, it will be an 'Apples with Apples' comparison on SQ, which he is well equipped to make.

I know where I would place my money.

Jonathan

cvrle59 wrote:

Clive wrote:Running 7.36+v51 with Tasker I find the sound very clear, clean and natural. Very 3D with instruments hanging well in space. I've no qualms about balance bass/mid/treble. Decay is excellent and timing pretty good too. These comments refer to an untweaked laptop running Tasker with no internet access. I've searched for the link to 5.14 but haven't found it so I can't make any comparison with that.

Thank you. I've tried 5.14, it won't run on my ivy bridge. There's a possibility I need a different control, the one in my wasapi avx folder is dated 29/4. Is there a particular one I should use or does 5.14 AVX live amongst the versions I cannot run...

Clive wrote:

cvrle59 wrote:

Clive wrote:Running 7.36+v51 with Tasker I find the sound very clear, clean and natural. Very 3D with instruments hanging well in space. I've no qualms about balance bass/mid/treble. Decay is excellent and timing pretty good too. These comments refer to an untweaked laptop running Tasker with no internet access. I've searched for the link to 5.14 but haven't found it so I can't make any comparison with that.

Thank you. I've tried 5.14, it won't run on my ivy bridge. There's a possibility I need a different control, the one in my wasapi avx folder is dated 29/4. Is there a particular one I should use or does 5.14 AVX live amongst the versions I cannot run...

Clive - it is an AVX2 version of 5.14 which only runs on Haswells not Ivy Bridge (which is what I also have).