Help addressing " special mode " problem

[stefanobilliani]

Hello Peter and All ,

From my current and limited system ,

all works fine itself , except when I try to listen to the special mode .
At a given buffer size , say 1024 , scheme 4 , sfs 150 mixed contiguous  or straight contiguous  ,well I just could listen one song at a time correctly .

When XX goes to the next track in the play list , distorsion occurs as very low pitch and rumors .
If I play stop and play again the "new" song it can be reproduced well .

This just in special mode .

What could I do to address the problem ?

Thanks

Stefano

[PeterSt]

Hi there Stefano,

To be honest, I never heard of such a thing. But let's first try something else :

Do exactly the same as what you are used to when it goes wrong at a track boundary, but now play Unattendedly (so, no XXHighEnd UI). This MUST work, unless you are able to tell me that sometimes it goes wrong in the middle of a track as well (but use something like a 10 minute track for it).

Unattended does not know about Track Boundaries, so there it can't happen.
But *if* it does, it should tell me something ...
Btw, do not use tracks of a different format for this test (in the Playlist Area), because then indeed "track boundaries" are dealt with explicitly in Unattended.

Let me know !
Peter

PS: How is your "DAC Needs" setting ? 32 bits or 24 bits ?

[stefanobilliani]

Hi Peter , I am back after a test :

Dac needs are 32 bit , and I confirm that the playlist area is composed with just one album of the same format . Then let's try .

Buffer Size =1024
Q1=14
PP=low
TP=normal
Sfs=150 , SC

I am able to tell you that sometimes it goes wrong also in the middle of
a track .
@Unattended the same problem happens , and even during the playing of the single track , then I have to stop by the means of hot key .
Sometimes the system plays the full track completely well ( it depends probably on the memory settings imo )

Looks to me like the system is close to its limits , since it tends to "loose synchronization" also when I -touch the player- ( it does means when I scroll the playlist menu or the settings menu during the playback ) .

In Adaptive mode , everything works perfectly also with very long playlists .

Thanks again

Stefano

[stefanobilliani]

Hi Peter , I am back after a test :

Dac needs are 32 bit , and I confirm that the playlist area is composed with just one album of the same format . Then let's try .

Buffer Size =1024
Q1=14
PP=low
TP=normal
Sfs=150 , SC

I am able to tell you that sometimes it goes wrong also in the middle of
a track .
@Unattended the same problem happens , and even during the playing of the single track , then I have to stop by the means of hot key .
Sometimes the system plays the full track completely well ( it depends probably on the memory settings imo )

Looks to me like the system is close to its limits , since it tends to "loose synchronization" also when I -touch the player- ( it does means when I scroll the playlist menu or the settings menu during the playback ) .

In Adaptive mode , everything works perfectly also with very long playlists .

Thanks again

Stefano

I may add one thing : it *sometimes* happens that when Unattended is not pushed , and soon after the request for "play" the player gui disappears from the screen , and can't be recovered . It will be still in the processes , then the music starts . After some minutes , but not always , the gui returns as if it was minimized on the bar .

[PeterSt]

Looks to me like the system is close to its limits , since it tends to "loose synchronization" also when I -touch the player- ( it does means when I scroll the playlist menu or the settings menu during the playback ) .

Yes, for sure it looks like that (after your thorough examination and report - thanks !);

To me it looks like your SFS of 150 is way too high. So, try 80 to start with, and try gain. Unattended would be the best, so any anomalies from the otherwise controlling XXHighEnd (which really controls playback at Attended !) are avoided to be sure (and anomalies regarding this sure exist; mainly when the system is not quite "up to speed").

The "out of sync" as you perfectly described it, has never been reported before (that I recall) and is a bad thing; I mean, this sure looks like sheer static, and of it is not, it sure can turn into that; when it does not, this is IMO just coincidence. But also ... maybe you use a preamp and attenuate analoguely. Well, a good thing when static relly emerges !!

When you push the system into an out of memory like you do here (and really, I recall that an SFS of 150 can work, at least with 8GB of memory), this is not really seen as an out off memory (and in the end it isn't really) but the OS needs too much time to find usable memory blocks and the implied "real time operation" in order here goes down in the "kernel" wrong. It is here where the synchonisation is lost, which in the end is about being too late to "match" two subsequent playback buffers. This by itself is because they don't start/end at "audio words" (which is a 1-channel complete sample) but at bytes (which is less than that word). So, all it needs now is skipping a complete buffer (no time for it !) and there you go - static I'd say.

but ...

When you are not able to cause real static (which would be windows blowing when at full gain), then there's the situation that the OS has run behind, and can't catch up anymore. Now it starts to skip samples, and although it brings distorted sound, it is no static. It might even start to be chipmunks stuff ...

Lastly, the fact that this happens with KS Special Mode only, is because, well, Special Mode is Special. It makes use of undocumented tricks, and ... the OS doesn't take those tricks into account herself when things come down to the matters (this is my fault of course, but, I assume a normally behaving OS).

Long story short : I think lowering the SFS significantly (at least at first) should show it helps.
In the mean time, let me know whether you (sometimes) can receive real static (which is no musical note at all), or that it's just distorted music always. So, no matter the lower SFS helps, please let me know.

Thanks and regards,
Peter

[stefanobilliani]

Thanks to you  Peter 

Recently , I changed the RAM bank in this old Laptop of mine .
For long had the suspect that the little system suffered some "bottle neck" effect from some side .
So that , changed the 2 gig to a more comfortable frequency ( higher than the previous ) and it started to manage well some obvious things , like the better audio and video .


The " non up to speed " systems mentioned in your reply above , reminded me to try again the " power settings " which are in the laptop ( every laptop? ) and under W7 , by pushing the battery icon .
And what I found? ... that the cpu was working at almost 40% of its possible speed . I never pushed to full speed before , simply because the old ram bank - by my point of view - was not "interacting" well for the system .
Now , with the CPU at 100% XXHigh End  to an SFS of 80 , has been able in Unattended mode to let me listen almost the complete album with only one interruption .
It's not perfect yet , but the sound is starting finally to go out through the speakers , more depth and bass and detail , and ...  in Special Mode .
Well a good achievement for today , and let say the problem is "solved" for now . 

Nice learning experience , and looking forward for a new PC anyway .


Stefano