File Testing

[DiscJocky]

Does any one know of any decent file testing software that can be used to determine the quality of recovered files?

[mediaman]

probably a bad answer but we put them on a ftp or http site and let the client browse them, they then have a 7 day guarantee to report bad files that we help them with. works fine.

[cgallery]

How would you write a program that would be able to determine the quality of a file?

[kovacsp]

As I didn't find a complete solution for this problem either, I use different tools.
For image files, I just let Windows to create thumbnails for the files. As soon as it finished, the images not having a thumbnail are usually bad. (I've actually tried one app that was for testing JPG files, but didn't find the corrupted images )
For Word, Excel, etc files, I've just written a macro to load all the files and test if they're ok.

kp

[mediaman]

In my quite limited experience good thumbnails don't mean good files.

In never understood how jpg's could thumbnail ok (showing entire image) but that same image could be corrupted, like half appearing. Does the structure of a JPG have a it's own thumbnail embedded in there?

[cgallery]

Does the structure of a JPG have a it's own thumbnail embedded in there?

Yes, many JPG's have embedded TN's.

[beto]

At some times if u did a soft wich verified an file for example "Office document" u algortihm could check the header, but what happend if on the middle of hex code , was damaged on the extraction process i.e bad sectors, bad media, etc your soft or any soft could, said its ok, where´s not, its hard , and what about the owner programs, for example here on mexico , many enterprise use an accountan program wich its only used on this country, the file wich its generated, for accounts, its very complex and doesnt has a header or a structure , so u cannot got a way to determine if file its correctly or no until, open it or mount it with the owner program, or maybe with an algorithm CRC, wich could be added while was saving and after that check the CRC?? well its an idea


Regards

[BlackST]

For JPGs and media files I have written some software that can also fix the mess keeping the other data, like shoot date, camera model etc. , for other kind of files sometimes manual check is the only solution. It is a service the customers pay for. I have made my custom tools during the years.

[scratchy]

There is some software that can test the headers of files, but like you say it can only give an indication of how whether the file is good or bad.

http://www.datarecoveryengineer.com

It looks like it is still in beta, so could be worth testing.

[beto]

i had check this program, but its the same, only check headers, im think its impossible to check the integrity fully without open it , by that reason, the header could be good, but on the middle or end of code, bad code

Regards

[kovacsp]

You're right, thumbnail-based testing is not accurate, but it gives a rough estimation of the overall quality of the recovered files. That's usually enough to see if I have to try harder or not
In case accurate results are needed, I'd try scripting around some good command-line image processing tool (like imagemagick), or better, code a small utility based on libjpeg / libpng / whatever, that does detect and warn about decompression errors returned by the library functions.
As for unknown file formats, there's no magic, at least I don't know any. I think you cannot detect if a file is correct in case you don't know the criteria of correctness for that file format.

[rchadwick]

Agreed. I went to that website, read the description, and was completely baffled. I'd love to find a tool that can test, and even repair batches of .JPG files.

[DiscJocky]

I think your right beto, any program is not going to be 100% accurate but if you can get something that is going to be in the high nineties most of the times then perhaps the header testing is the way to go.

I can take mediaman’s comments on board but when you are having to take time out for a lot of customers, to test files that they are not sure they know the names of or locations of the files that they want tested then it takes a lot of time which could be better spent working on other recoveries. Random testing is a long way from accurate testing so that doesn’t help.

Yes you can say to the customer here’s your files please test them, but then you are relying on the customers honesty and if they are paying for it after they have the data then there’s not a lot you can do if they claim a good file is bad.

[DiscJocky]

Thanks Scratchy, I'll check it out.

[bnice]

File verification tools.

A zip verify tool can be small K files, can verify zips quickly. The same is for near (99.9%) of every other file, find the decompression/verify tool, the job is made easy. We could use the site files area to store and share such files, that are found to work. With many such small K size tools needed, a gui upgradable. For any new tools added via simple script command line or better.

[mediaman]

IMHO WINRAR has the best built in zip tester

[bnice]

Refering to small cli tools the could be used via batch (.bat), cli or gui. For gui click and check all files, that it is capable of checking the recovered files for fitness.

Winrar cannot check jpg, xml and other files, this idea needs many small cli executables. Although the cli rar tool can verify archives that it supports. It is also smaller in size (as needed) than winrar and can be used with a cli, batch, gui.