mp3 corruption, from out of nowhere... Options

[Digga]

hi there.
I just listened to some of my mp3s and to my surprise I heared some real nasty loud clicks in a few of them. they where fine the other day, I'm sure of that. now I'm kind of in fear my whole collection might be or become partly corrupt
what might have been the cause for that?! the only cause I could think of would be some faulty hardware... maybe the hd, wouldn't have a clou though, never experienced something like that before.
I have to mention that all my mp3s are 'mp3gained' through v.1.0, I couldn't find any hints in the forums it's the apps fault though.

any help, other experiences or hints would be more than welcome.

some information on the system I use for playback:
MSI KT3 Ultra2 (VIA 333)
WD 80GB 7200 rpm 2MB cache
SB Audigy

WinXP Pro SP1
Winamp 2.81

[coy-itys]

Listen to the songs on another computer, maybe it's just a software error.

[magic75]

Try scanning the files with mp3utility for instance. If there is anything wrong with the files mp3utility should detect that. If it doesn't it is probably something else that is wrong.

[Jan S.]

The obvious cause for this would be a bad rip... how are you sure that it wasn't there before?
Try playing it in foobar (reports errors in console AFAIK).

[Digga]

Listen to the songs on another computer, maybe it's just a software error

hmm, Im gonna try that, though I wonder what kind of installed sofware / software conflicted would cause audible errors to mp3s...

Try scanning the files with mp3utility for instance. If there is anything wrong with the files mp3utility should detect that.

shoudn't they then have been sounded crap from the first moment on? I'll do it anyway.

The obvious cause for this would be a bad rip... how are you sure that it wasn't there before

bacause I had them for years and listen to them often. I guess I know my collection pretty much by heart

Try playing it in foobar (reports errors in console AFAIK).

ah, well, so I'm finaly tricked into installing foobar2k eh? I'll give it a try.

thx for your replies,
I'm gonna do as you suggested and report back then.

This post has been edited by Digga: Oct 7 2003, 17:21

[Digga]

reporting in.
(I only refer to one file, as I deleted the others already, gonna get them back flawlessy)

mp3-trim did not find any errors, everything looks fine.

I experience the same click when playing on another machine

foobar detects clipping in the song, but otherwise no errors accured.
that clipping couldn't be the source of the corruption I guess, because heavy audible distortions just appeared now and weren't there before...

does someone can suggest further steps to take in order to find the reason for this realy odd stuff? ta.

edit: clarification

This post has been edited by Digga: Oct 7 2003, 17:48

[Snelg]

(I only refer to one file, as I deleted the others already, gonna get them back flawlessy)

I experience the same click when playing on another machine

foobar detects clipping in the song, but otherwise no errors accured.
that clipping couldn't be the source of the corruption I guess, because heavy audible distortions just appeared now and weren't there before... 

does someone can suggest further steps to take in order to find the reason for this realy odd stuff? ta.

(I can't see an appropriate "embarassed" smiley in the list)

Actually, there is a chance that MP3Gain 1.0 is the culprit. If it is the rare bug I'm thinking of (which was fixed as of version 1.1), then the bit that confuses me is that the click is audible in Winamp 2.81...? As far as I knew, when that bug only made clicks that were audible in earlier versions of Winamp, and in a few other less-common players (no clicks in foobar).

Can you e-mail me the one mp3 you have left so I can check if I'm right? My hotmail mailbox doesn't have much space, so send it to glen@moosebutter.com

-Glen

[nwn]

- Mp3utility - Ok .. http://www.geocities.com/mp3utility/

- but who knows mp3guessenc ? mp3guessenc http://shibatch.sourceforge.net/

- and ;;; MP3Check http://freshmeat.net/projects/mp3check/?to...=19%2C114%2C123

- and also .. mp3Test ... http://www.shivi.de/MP3Test/

what about these softs?


Encspot of course ..

..

This post has been edited by nwn: Oct 8 2003, 07:18

[magic75]

Try scanning the files with mp3utility for instance. If there is anything wrong with the files mp3utility should detect that.

shoudn't they then have been sounded crap from the first moment on? I'll do it anyway.

Well, you said that the files were OK a while ago. That could have been an indicator that the file had been corrupted for some reason quite recently.

[Digga]

Try scanning the files with mp3utility for instance. If there is anything wrong with the files mp3utility should detect that. If it doesn't it is probably something else that is wrong

uhm, I confused mp3utility with mp3 trim, sorry bout that. I also checked the file now with mp3utility: no errors found...

@ nwn: I'm gonna try tha apps, thx.

@ Snelg: email sent, thx for the support.

I'll download the apps tonight (have to work today) and see what they say about this file...

[Lev]

Whether they say "File(s) = Good" or "File(s) = Not Good" will not change the fact they sound crappy.

So, have you re-ripped any yet? Do they still feature loud clicks?

[Snelg]

@ Snelg: email sent, thx for the support.

It's not a problem caused by MP3Gain. The rare bug I mentioned produces loud clicks where there should be absolute silence.
That's not what's happening in your file, though. Based on the one mp3 file you sent, my first guess would really be a bad initial rip.
But if you're 100% certain that the problem appeared recently, then my second guess would be that somehow you used a program that cleanly snipped a frame or two out of the middle of the mp3. That is not a problem that MP3Gain has ever had. (Really early versions would occasionally cut a file short, but never drop pieces in the middle )

An alternate (though less likely) theory would be that previously, you had a few frames in the mp3 that were just a little strange. Your mp3 player was able to recognize it as a corrupt frame and "smooth out" the sound between the surrounding frames.
But then you ran some mp3 cleaning program that said, "Hey, this is a corrupt frame. I'll just clip out this garbage data." So now your mp3 player does not see any corrupt frames and thus makes no attempts to smooth out the sound.
...or something like that.

-Glen

This post has been edited by Snelg: Oct 8 2003, 15:17

[madah]

I recommend that you use fsum or similar to calculate md5 checksums of your files. That way you will now for sure if the data is changed.

This post has been edited by madah: Oct 8 2003, 19:08

[Digga]

Whether they say "File(s) = Good" or "File(s) = Not Good" will not change the fact they sound crappy.
So, have you re-ripped any yet? Do they still feature loud clicks?

I have reripped the few songs that had clicks in them, now they are allright... again. so that means the source was allright, the encodingprocess also, as the last time. it must have something happend lately...
but this can't be the solution, as it appears to me stuff like that could happen any time again in the future, though I don't know why!
the source of one song is no longer available to me, so it seems that there is no rescue for it...

mp3-test made my system 'crash' (weird thing: it seems it changes something in the windows-settings, every time I clicked anything an windows error sound was heared, and I couldn't browse folders anymore, as the desktop was shown above all other windows...) I slowly get the feeling that my sytem isn't in the best shape... or mp3test is reeaall buggy.
all other apps recommended by nwn came to the conclusion that the file's o.k.

It's not a problem caused by MP3Gain. The rare bug I mentioned produces loud clicks where there should be absolute silence.

well, that at least (hopfully) rules out one possibility for the corruption, again thx for your support.

But if you're 100% certain that the problem appeared recently...

I am.

then my second guess would be that somehow you used a program that cleanly snipped a frame or two out of the middle of the mp3

I coudn't think of such an app. beside mp3gain and encspot I didn't use any program on my files...

I recommend that you use fsum or similar to calculate md5 checksums of your files. That way you will now for sure if the data is changed.

uah, command line utility *shiver* forgive me my ignorance, but did I get it right that the tool checks if the data has changed from timestamp a to timestamp b?
if this is the case, then it would be useless, as the data (according to me) has already changed... timestamp a would thus be after the corruption and the app would not reveal any changes. please correct me if I'm wrong.
anyway, I tried the tool (I hope I did it right...) and no errors were found.

This post has been edited by Digga: Oct 8 2003, 22:07

[M]

I recommend that you use fsum or similar to calculate md5 checksums of your files. That way you will now for sure if the data is changed.

uah, command line utility *shiver* forgive me my ignorance, but did I get it right that the tool checks if the data has changed from timestamp a to timestamp b?
if this is the case, then it would be useless, as the data (according to me) has already changed... timestamp a would thus be after the corruption and the app would not reveal any changes. please correct me if I'm wrong.
anyway, I tried the tool (I hope I did it right...) and no errors were found.

If you dislike command-line utilities that much, you could try MD5Summer (a GUI-based MD5 program). However, generating MD5 sums for the files you already know have been corrupted is a useless task; the program does not check timestamps, but it does check the actual binary data within the file. If you are going to use MD5 sums, you should generate these after encoding/tagging your audio, on the final version of the files (that is, if you are ready to archive them to CD or something similar, and are not going to edit the tags/change gain/alter anything else about them ever again, since any of those thinigs would change the resultant MD5 sum!). In other words, you cannot use MD5 sums to magically reveal if/how your current files were corrupted... but if you encode new files, you can use MD5 sums to verify whether they remain intact.

Another option is the use of parity files, and I'm afraid I know far less about that procedure than others on this board. However, parity files would allow you to recover corrupted data (although, again, you are limited to generating parity files from your original, uncorrupted files; they will do nothing to help you recover what you have already lost).

Hope this helps.

- M.

[Digga]

However, generating MD5 sums for the files you already know have been corrupted is a useless task

that's what I thought.

Another option is the use of parity files

uhm... never heared of that kind of stuff, please englight me.

the idea of creating checking-methods for newly encoded files to see if they where corrupted after some time seems more or less pretty useless to me in my case (no offence meant). I just can't periodly check my files for errors and encode the faulty ones again and again. ah well, I could, but thats not the way it meant to be.
that would be a way to deal with the problem, but not to get rid of the problem itself!
besides, some mp3s I've got wheren't encoded by me, so I can't get them back easily (and I actually I didn't intend to...)

I realy appreciate everbodys effort to help me with my problem, I sorry if I might sound a bit rude (do I?) for expect a fast sollution for that, and to push asside well meant hints... but my music collection realy means much to me, and I can't stand the thought (actually the fact! ) that for some unknown reason potentially all my mp3s could be seriously corrupted...

[niktheblak]

This reminds me of an incident that occurred to me about five years or so.

At that time I was using mostly Win98's DOS. At one day, I connected a friend's HDD into my system in order to transfer some files to it. After performing the transaction and removing the other HDD, I noticed that virtually every file on my FAT system was corrupted. No EXE file would execute and after reinstalling the OS, I noticed that all of my MP3's, modules and images were badly damaged.

Even today I wonder what the hell happened on that day. I was pretty pissed off to my friend for some time, because I naturally blamed his HDD for the incident.

Maybe something similar has happened to you too

[Digga]

Even today I wonder what the hell happened on that day. I was pretty pissed off to my friend for some time, because I naturally blamed his HDD for the incident.
Maybe something similar has happened to you too

haven't had another hd in my precious little pc for quite a while (it has come to me, it is mine, nasty little hobbit )
exept for the above mentined issue after using mp3test my sytem ran stable for a whole while. all other files /docs seems to be in best shape.

my condolences for your accident though, must be awfull if it was a huge collection...

btw. I already scanned for viruses, nothing found...

This post has been edited by Digga: Oct 9 2003, 14:13

[madah]

This is a very good test to determine if your system is really stable:

1. memtest86. Burn an ISO and boot with it. Choose "Total (all tests)" and let it run overnight (test takes 5+ hours or so). If it fails then you have bad ram, only solution is to replace it ASAP.

2. Prime95. Let it run overnight or for a couple of hours. If it fails, then your system is perhaps overclocked? Install latest drivers for your motherboard/chipset. Try to lower the FSB in BIOS settings.

2.5. Additionally, you could try to compress a coupe GBs of .wavs with Monkeys Audio, in "Extra High" mode. Use the 3.97 version. Restart (to make sure the files are read from disk and not from cache). Then run verify on all .ape-files.

3. Then run a complete full surface scan with chkdsk.

For a stable system, all of the above should pass 100% with no errors.

Parity/recovery-files can be generated with QuickPAR. But of course it can't recover already damaged data!

This post has been edited by madah: Oct 9 2003, 14:19

[Digga]

I let Prime run recently, everything went fine.

I'll do mem tests etc soon.
but do you think that kind of errors described by me could be caused by bad ram?
the only issues I experienced and heared of so far with faulty ram where random turn offs, game chrashes and boot problems, nothing that I notice atm.
but then, why not test all available possibilities, it's by now not proven that it is not the rams fault afterall.

will post result then.

This post has been edited by Digga: Oct 9 2003, 14:36

[Digga]

all right, done everything suggested. prime95 ran perfectly the other day, mem86 didn't find any errors, checkdisk sais my drive's fine. ape files are also fine.

I just had a conversation with afriend of mine who also got an audigy 1, he had similar problems. could it be that my (maybe faulty) audigy card causes the permanent corruption?!

[/\/ephaestous]

but do you think that kind of errors described by me could be caused by bad ram?

Yes, operations that include rewriting the file (e.g. updating an ID3v2 tag), could damafe the file when the RAM is bad.

[Digga]

Yes, operations that include rewriting the file (e.g. updating an ID3v2 tag), could damafe the file when the RAM is bad.

that sounds logical, thx for the information.

coming back to my mp3 problem: I replaced my audigy with a new soundcard, I hope that this will solve the problem, as it might have been the cards fault...
but it realy could also be anything else... ah, well, I'll see.

I thank everyone that participated in this thread in order to tackle the problem!
I'll post here again if the situation doesn't change.