Corruption in RLE compressed TIFFs?

[Jose_Cao-Garcia@adobeforums.com]

Hello,

I like to use layered tiffs as my format of choice but I'm discovering that with Photoshop CS, that I get a *lot* of corruption with these tiffs. I will get a horizontal band of missing data in a specific channel, or the data will move to a different point in the channel. The problem doesn't seem to affect the flattened part of the tiff (if I open up the tiff in Imageready or Preview the TIFF looks file).

Oddly enough the corruption (seems) to occur when I am working in Illustrator or InDesign - not when I am in photoshop. Might this have to do with the fact that I am using my media drive as a scratch disk? I always have but it has never caused a problem before.

How I'm saving my TIFFs: RGB Tiff with layers, no general compression for the tiff, but RLE compression for the layers.

My Computer: G4 MDD FW800 Dual 1.25, 2GB ram, OsX 10.3.3, 120 GB RAID 1 for media (not on a PCI card), Creative Suite Professional (with all updates applied).

Thanks,
-Jose

[Chris_Cox@adobeforums.com]

No, nobody else has reported problems.

At least RLE tends to allow for more error recovery (LZW and ZIP don't).

This sounds suspiciously like the scratch drive or RAM is bad.

[Jose_Cao-Garcia@adobeforums.com]

Though I suppose one can never be 100% positive about this stuff, I am fairly certain that neither the scratch drive or ram are bad. (I don't use cheap ram, the raid is 2X seagate Barracuda 120 GIG drives that both test out fine with TTP4 and Disk Doctor.) I have also used this machine for Audio Production work and have had not had any file corruption problems.

If it were a hardware issue it would seem that other applications and filetypes would be affected, this problem *only* manifests itself in Layered Tiffs that use RLE compression, only in PhotoShop CS.

This is the only case of file corruption I've had the entire time I've owned this machine and I haven't changed ram or drives in a long time, the problem began to appear when I started using Photoshop CS.

[Toby_Thain@adobeforums.com]

I agree with Chris - it does sound like a RAM problem. I suggest running Apple's extended Hardware Test a few times.

[Jose_Cao-Garcia@adobeforums.com]

I have tested the ram and hard drives with Apple's Hardware Test as well. The ram and drives are both fine. As I said before, this is the first file corruption problem I've encountered.

[Ramón_G_Castañeda@adobeforums.com]

Jose,

RAM testing utilities need to be run continuously for days at a time, and even then they often fail to uncover bad RAM in certain cases.

I had a case like that myself once, when one out of three RAM sticks passed all RAM software tests and eventually turned out to be the culprit in an intermittent but frequent crash I was experiencing when using the ColorVision Spyder. It never caused a single problem otherwise.

I'm not saying your RAM is bad necessarily, but at this point you can't rule it out.

[Graham_Phillips@adobeforums.com]

I had a bad RAM stick last year, and no amount of testing with the Apple Hardware Test showed up the fault. But after a day or two the computer would kernel panic.

[Jose_Cao-Garcia@adobeforums.com]

Ok, well I guess I should maybe find a more rigorous RAM test and see if It can find any problems (though I'm still pretty skeptical). I can set it going on Friday night and take a road trip this weekend - What RAM testing utility provides the most rigorous/complete test you can think of?

[Buko@adobeforums.com]

What RAM testing utility provides the most rigorous/complete test you
can think of?




Photoshop.

Yank all your RAM but one. does it still screw up?

next stick.

[Jose_Cao-Garcia@adobeforums.com]

Hey Buko - that's a good idea, I'll give it a shot. Only problem is that this is not a consistently occurring problem, it happens to some files, some of the time (but they're always RLE compressed Layered TIFs). So the test may take some time ... I can probably speed up the process by making the scratch disk go nuts with some huge files/processing tho.

Will give that a shot later on ... thanks, good idea.

[Ramón_G_Castañeda@adobeforums.com]

What RAM testing utility provides the most rigorous/complete test you
can think of?




None (other than extensive use of Photoshop :)). Follow Buko's advice; it's the only way.

[Jose_Cao-Garcia@adobeforums.com]

Ok thanks y'all - I'll post back to this once I've had time to run the test.

[Chris_Cox@adobeforums.com]

The only good way to test RAM is to remove it or swap it out (short of using a $10k hardware RAM tester, and even I don't have one of those).

[Jose_Cao-Garcia@adobeforums.com]

Yeah, that's what I'm beginning to find out. I'm trying to see if there's a computer store or technology place around here that has one, but that is a wildly distant possibility so I will perform the test described earlier. I'm also trying out memtest:

<http://www.macupdate.com/info.php/id/14004>

A number of people have told me that running this utility continuously over several days has helped to identify bad ram sticks.

[Toby_Thain@adobeforums.com]

memtest does not test all of RAM, but I've successfully used it to detect RAM issues. As far as application level testing goes, running a compiler (CodeWarrior or gcc) tends to quickly and dramatically reveal RAM problems. Try a "make -j 8" on a large project :)

[Jose_Cao-Garcia@adobeforums.com]

Thanks Toby - is it best to remove all but one stick of ram, and then run memtest 5 or 6 times, replace the ram and repeat (to figure out which stick if any is bad?) or if you run it with the whole kit n' kaboodle inside will it identify the bad ram stick if it encounters any errors?

[Toby_Thain@adobeforums.com]

You might be able to work out which stick from the reported address. But I'd first try it with all RAM installed just to see if it turns up anything. If it comes up clean, and you are still suspicious or have time to kill, change the ordering of the sticks and run it again in case it missed a fault within the memory it has to reserve for the O/S while running.

Two things are really needed for a solid tester on Power Macs: test patterns and regime that is designed to trigger faults - apparently this is a black art - and a tester that can be booted directly (like Apple's HW test) so that ALL memory can be tested. I keep seeing reports that Apple's test is not missing RAM faults, which is very worrying. Yet memtest runs within the O/S so it can't be exhaustive. On Intel there is memtest86 but there doesn't seem to be anything like that for the Mac.

[Jose_Cao-Garcia@adobeforums.com]

Yeah Apple's HW ram test and Techtool haven't turned up any problems for me - thanks again for the info. I will post my findings once I've run al the tests. I am also seeing if I can find a way to have my ram tested in a hardware ram tester.