Short interruptions in playback en very tiny hickups since y3 and y4

[ivo]

Hei, I also did notice this in my setup, I would even say that gapless playback is gone sometimes.

Cases:
1) Between the 2 tracks there is a short drop out, however it is the same album.
2) At the end of album, dropouts occur in the last tracks only, the first tracks play well.

When I did HDD defrag in my Vista (used alternate software) the dropouts became much smaller. It makes me think this all is connected to the way XX works with files on the HDD. It has become more sensitive to HDD access since y4.

[PeterSt]

Ivo, thank you for sharing this. And yes, it occurred to me that Ed is experiencing this not right from the start, or anyway it gets worse lateron. This well can be when memory is involved, and knowing that the used memory will always grow somewhat in the beginning until it has reached a stable level. In the mean time though, other processes may have more problems (because of less available memory) opposed to the beginning of playback.
So it is the art here to recognize what actually is causing this bugger.

Anyway, that 0.9y in general is more sensitive to this, I experienced myself too and is even recognizeable in the program to a certain extend.

Well, for those who may think that supid Peter only destroys at some times, here's another experience from just the other day :

I was copying a disk for backup during playback. At a certain stage I thought "did I hear a click ?" At a next stage it became clear that clicks were there which shouldn't, and again later (think in terms of an hour) it became crazily annoying, and ten minutes later I had to shut off the music.

Doesn't this sound similar ?
In this case the memory did not increase, but sure something was more and more degrading sound (puh, call that degrading while in the end it is merely like trrrrrrrrr).
This, while before I could do these kind of copies unnoticed.

The (or my) point is, there is no way I can cause this explicitly, while some change of mine certainly now incurs for it. It is that backup "causing" it, whereas before it caused nothing.

[ivo]

Peter,

Well, it seems you are in even worse situation. I have pretty well optimized my laptop and actually I can browse web, copy files from hdd to my usb stick and do similar stuff during playback and I do not hear such clicks at all. Actually I use to do nothing when listening to music, but just giving example of what I can do. And the most interesting is that when I do not do anything just listen I hear short dropouts as I posted earlier, however it is ok with latency and it is ok with HDD (it is fragmented for 1% only!).
Peter: what is your latency as reported by DPC latency checker?

[PeterSt]

Don't laugh ... a steady 1000.
This means there's some fixed holding up somewhere, and I never could find what it is. But, it never jumps out of that 1000 either, and officially it's okay.

But to keep in mind :

Some have the experience of this reported latency *dropping* when XX plays. Go figure.
(although back then I could reason a bit how this can happen).


PS: I have an (HTPC organized) XP machine that shows 30-40 or so. Not much helpful, but that thing is loaded with everything you can imagine.

[ed linssen]

Hi Peter,
Back again.
First I have to tell you that the Atom I am using is the 330. This means it has two real cores of 1.6 each. It has even the possibillity of switching on hyperthreading. That means two extra virtual cores. I realize myself now, maybe I better can switch that off. Maybe coreapp. can't work with that properly. I have to try that.
I have worked with the single Atom to, and that one worked fine withe the older XXhsses.
Next..
I have done all the settings you asked for. Starting with Four times Upsampling Arc pred', Core app. 3, 120MB, Unatt. etc.
I started a disk. 1: 4.17min, 2: 4.01min, 3: 5.12, 4: 2.51, 5: 3.26, 6: 6.07, 7: 3.26, 8: 3.04
Music started after 25sec (15sec, end of third track with no upsampling 4xchecked), end of 7th track.
Irregularly very short flashes. Now once in tw seconds, then in half a second.  Not predictable! More seldom than these small onces, irregular too, longer flashes. Some seconds after the occurrance, music can halt for half a second orr even three or four seconds.
I tried this with 4x, No Upsampling, and had more regular flashes folowed by stops.
Think you'r getting a little crazy now!
Will stop here for now,


Ed

[ed linssen]

Hi Peter,
Maybe good to know.....
The dropout behaviour is unchanged when I uncheck double and quad.
Strength!

Ed

[ed linssen]

Hi Peter,

..and forget about the ticks. I am affraid they were not related to the stops. They didn't show up the last two days or so.

Ed

[PeterSt]

..and forget about the ticks. I am affraid they were not related to the stops. They didn't show up the last two days or so.

But I hope you know what you did to let them get away ! (notice that a reboot can "do" that).

[PeterSt]

Irregularly very short flashes. Now once in tw seconds, then in half a second.  Not predictable! More seldom than these small onces, irregular too, longer flashes. Some seconds after the occurrance, music can halt for half a second orr even three or four seconds.
I tried this with 4x, No Upsampling, and had more regular flashes folowed by stops.

Ok Ed, thank you for your time on this.

So, notice that with 4x upsampling and 120MB Split Size, the really only I/O XX is causing will be at the 1 minute mark. One next a 2:10 and then at around 3:10 etc.. A few tiny I/O's may buzz around that larger I/O (with large I mean half a second, a bit depending on the speed of your (I/O) system) but always right ahead or after it (maybe one second distance.

All other I/O is not to be there.

Question : Did you perform this : How I tweaked my Vista virtually dead ?
And notice that the one thing which may bug you for hours and hours (even to infinity) would be the "Indexing" of the drives. This is a checkbox at the bottom of the property page of the device (rightclick on it in Explorer), and if you really have many albums, Vista will indeed be busy with that forever. This is the only explanation I see for so many I/O's, outside spyware (!!). Anyway, shut that indexing off if you have it on, and denote "all dubdirectories and files" (similar) as well. *IF* you have done this before, but forgot the subdirectories and files, redo it by briefly switching it on, and off again. Only then you are able to shut off the subdirectories and files. Do this for each disk (partition) you have, including the OS disk !

If this is not it, unplug your network cable and see if it stops. If so, you'll be having spyware.

Lastly for now, I don't know anymore whether I asked to you shut off the virtual memory, but if not, do so for testing; I mean, those I/Os also can come from Vista putting your music files into virtual memory (hence the swap file on disk), and that causes the highest priority I/Os and will cease sound for sure when performed at the wrong moment. If you have it On for more disks, shut it off for all of them.
Please note that I don't think this is the culprit, because the I/Os coming from there won't be "continuous" and only at track or file part boundaries. However, only when virtual memory has been shut off, you can be sure the I/Os you see don't come from there. Otherwise you will be in the blind.

Important : This effort is only worthwhile if you have these "drop outs" within the around 1 minute boundaries. If you have them at those boundaries only, it still may help, but chances are little then. All depends on what you mean by "irregular too" and your interpretation of the longer lasting flashes". Otoh, if such an irregular longer lasting flash comes together with the "flash" I need for loading the track (part), you may well be in trouble because of *that*. It is all a complex thing and about "balancing" priorities with regard to necessary processes, keeping up the good sound, and presenting some sound at all. Nothing is allowed to disturb that, but in your case something (or many things if you never performed the Tweak Dead subjects) does.
Regardless of 0.9y changes, what you have, should not be.


PS: It doesn't look like your Atom is in trouble, which was about your first test : when does music start playing. It is good you added the "no difference with no upsampling", because that only prooves it more. It is the I/O (and the priority of it) doing it to you, and notice that "priority" can also mean the processor stalls because it *has* to. This happens with virtual memory swap activitities (some times).

[JohanZ]

So, notice that with 4x upsampling and 120MB Split Size, the really only I/O XX is causing will be at the 1 minute mark. One next a 2:10 and then at around 3:10 etc.. A few tiny I/O's 

I do recognise this interval on mine system when i have the external diskacitivity. Every minute I hear the external disk spin-up, mostly followed with a break of 1 or more seconds in the music. When Ed uses external drives that are very quit is it possible that we have the same problem?

Ed you mentioned earlier that you use two seperate discs? Using "Copy to XX drive" And when you switch your external drives off after a break in the sound, do you get error messages?

[ivo]

Hei, some news from me:
Tonight I did listen (unattended) to 2 albums (mp3 192 kbps) and did not have any interrupts in music at all with latest XX. Music files sitting on the same drive where XX sits, just different folder.
My XX settings (see also signature):
scheme-2
nothing
high
Next I mention only then if ticked:
Allow Format Change
24 bits
simple search
fast rendering
stop services
100 MB
start Engine3 during conversion
tooltips

[ed linssen]

Hi Johan, Hi Peter,

I did kill vista/w7 like peter advised. In fact, it is one of the first things I do after a good working fresh install. In my case I have eventually put some more disc-activities from automatic to manually.
Yesterday late night I did put the latest bios in my little Intel-board, fiddeld with the PCI latency settings, switched on Hpet in the bios, and some more crazy things. But no gain! Too much disc activity still! Causing stops. I tried without Hyperthreading as well. No effect!
Johan, my new setup is one 2.5 inch sata disc with both XXhi-end and W7 Ultimate 64bits on sata 0. A second sata-disc with the music on it on the second sata connection on the board. So both intern!
My momentary dac , 24/192, is the PCI audiocard from ASUS, the Xonar essence st, works pretty good, even with the four times upsampling and Arc Prediction.
Today I even played with a seperate USB mediaplayer as extern musiccontainer disc, made regrettably no difference in the dropout behaviour.
In the meantime I have forgotten what I did more, instead of getting to bed on a reasonable time. But I fancy playing around with the toys, so.....
Hope you(and I) will find a solution eventually.
Keep your ears open!!!
Regards,
Ed

 

[Telstar]

Hi Johan, Hi Peter,

I did kill vista/w7 like peter advised. In fact, it is one of the first things I do after a good working fresh install. In my case I have eventually put some more disc-activities from automatic to manually.
Yesterday late night I did put the latest bios in my little Intel-board, fiddeld with the PCI latency settings, switched on Hpet in the bios, and some more crazy things. But no gain! Too much disc activity still! Causing stops. I tried without Hyperthreading as well. No effect!
Johan, my new setup is one 2.5 inch sata disc with both XXhi-end and W7 Ultimate 64bits on sata 0. A second sata-disc with the music on it on the second sata connection on the board. So both intern!
My momentary dac , 24/192, is the PCI audiocard from ASUS, the Xonar essence st, works pretty good, even with the four times upsampling and Arc Prediction.
Today I even played with a seperate USB mediaplayer as extern musiccontainer disc, made regrettably no difference in the dropout behaviour.
In the meantime I have forgotten what I did more, instead of getting to bed on a reasonable time. But I fancy playing around with the toys, so.....
Hope you(and I) will find a solution eventually.
Keep your ears open!!!
Regards,
Ed

 

You dont have any of the following:
-antivirus
-defrag (of all brands but the ones that are not resident such as defraggler)
-any of the cr*ppy asus utility (es. asus probe)
-some of the autoupdate stuff from creative

Go to  your control panel and check for those. also post a list of the running processes. There must be something bugged (my first suspect is asus probe).

[ivo]

Ed, please do the following also:
1) turn off pagefile
2) set to disable or to manual the majority of your services (be carefull here!!! I have only 36 "STARTED" services out of 146 on my vista home premium)
3) turn off scheduled disk defrag
4) check dma settings for HDD, DMA helps a lot
5) turn off aero or similar UI stuff, stick with default or standard scheme
6) turn off all UI effects
7) turn off update features of all your software (including windows update, it can be turned on manually when needed)
under vista there are still several deeper things to turn off or set to manual (just do not remember all now, take a walk over Control Panel)
9) set to disable in bios and in OS all hardware which you do not need (built in audio, comm ports, etc)
10) turn off HDD indexing stuff, this one is really IS NOT NEEDED

[ed linssen]

Hi Telstar,

Thanks for your remarks!. Asus drivers are reasonable weak, the 64bit drivers, the very latest though still beta, exelled over the vista 32 drivers. The exe update utillity works smooth, but prone to virusses of course. I wil do a thorough viruscheck tonight, hopen to find some bug or whatever. I must say, I had the same problems with the former Vista-setup. The dropout thing....
The computer is not online and not connected to my network or the internet. But it has been for a short time for updating W7 after the istall!
Most of the drivers and stuf I needed was brought into the cumputer with a USB-stick.
The sevices-list changes in a weird way, it grows and get smaller all the time with a lot of doubles. Looks like infection indeed. I dive in it right away and let you know what I find.
Thanks for your advice,

Ed