XXHighEnd

I'm not sure if this is the best place to put this post, but...

Having replaced my previous 2x 30GB 'slow' SSDs (100MB/s sustained write - one for OS and one for XX) with a single 'fast' 90GB SSD (250MB/s sustained write - for both OS and XX), I now have a very specific problem:

When I record from vinyl, the recording software saves 16/44.1 rates as 24/44.1 files. With my previous setup, I had no problem applying QAP to these 24/44.1 files. But now, I get a loud tick coming through the left speaker at the beginning of playback and continously throughout, whenever data is transferred to memory. It doesn't matter if I use 'Copy to XX drive' or not. Nor does it matter if the split file size is lowered right down to 12MB.

To put this into context, it is NOT a big a deal at all - I don't need to get these files working with QAP - QAP works perfectly well with all 'normal' 16/44.1 files. But I'd like to know what has changed that could be causing this... Before, I could apply QAP to these 24/44.1 files and it worked fine.

Mani.

So Mani ... you say that this now happens only because of using a single SSD ?


Like in : each 30 seconds or so with a low Split File size ?

If this is so, it is totally crazy.
What happens if you direct your Swap File (virtual memory) to one of the other disks ... ehh, can that be done anyway ? ... If it can not, what if you shut it off ? (either just for testing).

Interesting ...
Peter

It doesn't have anything to do with your hardware (IMHO). I'd put the blame on your system settings (I assume that you reinstalled your OS). Try to disable virtual memory and set processor scheduling to background services. I'd also reinstall sound card drivers as well as review XXHE settings.

Greets,
Marcin

I think you're right Marcin, it doesn't look like a hardware issue. I've just installed another 90GB 'Vertex 2' SSD; so there's now one 'fast' SSD for the OS and another 'fast' SSD just for XX. And the loud ticks are still there!

(I had already done the things you suggest, and of course 'tweaked W7 to death'...)


Yep! Every ~9s with a SFS of 12MB, and every ~45s with a SFS of 60MB. Sometimes, they're quite hard to hear when there's a lot of HF energy (cymbals, etc) in the music. But they're clearly visible on the SW/HW 'level meters'.

I still have the 'slow' SSDs that I replaced from the PC. I'll put these back in and see if the ticks disappear. If the ticks are there, then I can only assume that I missed them before.

Of course, 24/44.1 files are almost non-existant in the 'real' world, so this just shouldn't be an issue. However, I remember hearing exactly the same ticks (from my left speaker) when 'Decode HDCD' was on and a 16/44.1 HDCD file was played in XX with QAP. My understanding is that hdcd.exe converts 16/44.1 HDCD files to 24/44.1 files. The fact that there were ticks with these seems too much of a coincidence to me...

Mani.

What is your power scheme setting? Set it to performance.

Already done!

I think I've optimised things as much as possible in the SW.

And I'm not sure I can improve things any further in my passively cooled, no-moving-parts, totally silent PC... FWIW, here are my 'Windows Experience Index' scores:
Processor 6.5
Memory (RAM) 7.6
Primary HD 7.4

But if you have any ideas, please let me know.

Cheers,
Mani.

(Latency attached - playing a 16/44.1 file with QAP.)

To Mani,

Didn't you told here a while ago that you not using all CPU cores ?

Could this be a part of the problem ?


Roy

PS: Nice DPC value (should be around 10-15 range when xx plays)

PPS: Maybe do a BIOS update.
PPPS: Did you properly flashed the SSD's

Hi Roy,

Thanks for the suggestions.

I tried 2 and 4 cores, with all the different appointment schemes in XX. But the ticks remained.

I think the easiest way to figure out what's going on is to simply try a 24/44.1 file with QAP on another PC. I'll hook the FF800 to one of my other PCs and see what happens.

Actually, anyone who downloaded my files from the '16/44.1 vs 24/176.4' thread can help here. Do you have the ticks when you play the "native-16/44.1" file (which the recording SW actually saved as a 24/44.1 file) with QAP?


How do I do this?

When I installed the SSD for the OS, it was a 'clean install'. I'm assuming the install initialised and formatted the SSD using NTFS. For the second SSD, I initialised it and formatted it as exFAT.

In any event, a WEI score of 7.4/7.9 suggests that the SSDs are working OK.

Mani.

Hey, didn't hear that one for a long time ! Sounds better than the past time I listened. Haha.
Anyway, no ticks. Scheme-3 and the 2nd core stays "empty" (see first pic).
With OAP the 2nd core has to do a few things (2nd pic), and I posted it because of the downgoing slope in the 2nd core while as far as I am concerned this should stay equal. Hmm ...

But Mani ... I guess you're still outputting 24 bits ? If so, try 32 ands I'll bet you it will be away. In that case it is just my fault (and you didn't notice it before).

Let me add this (not much to my credits :)) : On to the next universe I have been working with a very very revealling amplifier, and I think I can fairly say that whatever happens in the system is just audible; loading a track piece in the first place. But nothing like real ticks of course.

Peter

Peter, we couldn't wait to get the new version of XXHE :)

Time to give up on this 'non-issue'!

I've just tried QAPing a 24/44.1 file with my Weiss AFI1... which is capable of outputting both 24 bits and 32 bits (the RME is only 32 bits). The ticks are still there, irrespective of the output depth.

BUT...

I've just tried QAPing a 24/48 FLAC file (from Society of Sound) and... NO TICKS WHATSOEVER!!! So, it really is an issue with XX QAPing 24/44.1 WAV files (I don't have any 24/44.1 FLAC files to try, and quite frankly, I'm not particularly inclined to now).

Peter, are you using a different version of XX? (Have you fixed any hdcd.exe issues in your version, because I think this is linked? EDIT: in the sense that hdcd.exe creates 24/44.1 files.) I'm pretty certain there's a problem with QAPing 24/44.1 files in 0.9z-2...

In any event, it looks like my HW is OK... phew!

Mani.

I've just tried QAPing an 16/44.1 HDCD file. With the SFS set to 12MB, I'm getting ticks every ~9s!!! And every ~60s with SFS set to 60MB.

The ticks are virtually inaudible. But they are easily seen on my DAC's level meter, peaking close to 0dB!!! On the left channel only.

I'd recommend that you DON'T use 'Decode HDCD' and QAP with an HDCD file... especially if you play at loud volumes!

Mani.

Mani,

Now it is not clear to me anymore when you have the ticks. I thought before you said it was about native 24/44.1 (hence no upsampling of any kind), and now it is about QAP ?

I will admit in advance it is me (as I offered right that haha), but since I don't have it in the first place, it's always good to know when things happen, EXACTLY.
Btw, was your Take5 a real example, or did you just think it would be ?

Sorry if I misinterpret things !
Peter

Sorry, I probably wasn't very clear to begin with.

I get ticks, as described, under two conditions:

1) when using my recorded 24/44.1 files with QAP
2) when using 16/44.1 HDCD files with 'Decode HDCD' and QAP

I'm pretty certain that this has nothing to do with my HW and that this is an issue with XX, for all the reasons I've cited above. Ordinarily, I wouldn't care too much about this. But if it affects playing back HDCD files with QAP, then I think it's more serious.

HTH.

Mani.

Ok, I can repeat it !

Indeed it is about 24 bits input, and it only occurs with QAP. So, not native, not DAP, not OAP, only QAP. With or without Peak Extend doesn't matter.

GREAT! I can sleep well at night.

If it transpires that there is a major issue with QAP and 24 bit files, one that impacts negatively on the SQ (as well as creating intermittent ticks), then I may have to totally re-evaluate my findings in the '16/44.1 vs 24/176.4' thread.

Mani.

But you could do that tonight, yes ? :) :)

The only thing I can tell now is that 24/88.2 QAP exhibits the same. So it really looks like Quad Arc Prediction is doing it, and not the resulting sample rate. Also I am sorry I never heard it, but I *couldn't* hear it, because OAP (which I always use) is just allright.

Since it is in the left channel only, I'd say I must be able to find some typo in the program somewhere. So, you'll just have a good night for now. Also because of this :

:offtopic:

I never came round to testing the "low Split File size" for its SQ merits, but today I used it to run into the ticks quickly. But I would swear it sounds better ...
I didn't pay attention to it at all, but it just occurred to me. So I thought "are those 24 input bits doing this to me ? nah ..." and then I thought of the small Split File size (12). Hmm ... I guess I leave it to that for now.

Thanks,
Peter

Peter, here's your homework for tonight: read http://www.phasure.com/index.php?topic=1368.0

Mani.

Ok, solved for 0.9z-3. And you know what ? I first solved 2 other bugs (but indeed typos) of which I thought they did it. How disappointed was I after the first solving not helping a thing.
After the third indeed did help, I went on ... and lost counting a bit. But 9 or 10 of those "typos" could be found (all different ones). Some without theoretical effect, but some with !

I even found one which really can (no, should)  influence SQ throughout if the file is 24 bits and upsampling (except AI) is in order.

I will admit that it is not the best thing that I (now for months) listen to different code than anyone of you, otherwise I sure will have noticed. As is the case with the OAP code of which I told I can hear each trackload with that very very resolving amp (and which I never heard before). But I still was ignorant enough to think "it just was so". Not anymore. So, on to that part now. :yes:

Peter