I just buyed a Lite-on 451s.

I tried to burn my first DVD, a +RW 2.4X TraxData made my Ricoh.
I burned same data for 7 times without visible problems (soft or hardware). After every burn I did a verify (with the included option from Nero or with a stand alone file comparer when I didn't use Nero)

- First burn with Nero 5.5 (last version from 5.5 series)
The file comparison gave me 8 differences between the files form HD and the files from DVD
- Second burn with Nero 5.5 (last version from 5.5 series)
4 differences
- Third burn with Nero 6 (last version from 6 series)
1 difference
- The last 4 burns was made with Nero 6, RecordNow 4.6 DX, Disc Juggler and again Nero 6
0 diferences for all the 4 burns

To sum up, 7 burns on the same DVD+RW Ricoh 2.4x, with the same files gave me file diferences on first 3 (8, 4, 1 differences) and no differences on last 4 burns.

What can be the problem? Why this behaviour?

Thanks,
Dumitru

I´m not sure I understand you right, but I can see from your post that you are trying to verify "burned data".

As an old Linux user my tool to verify critical data is md5summer. (You never tell us what tools you are using, or how you use them.)

For fun I just made a "real life test".
And this is how I tested and verifyed burned data.

i. I run md5summer on a 233MB ISO-file. (md5 hash-file saved.)
ii. Then I burned the same ISO-file with Nero 5.5.10.56 on a DVD+RW disk (DVD+RW:RICOHJPN-W11-001), as data.
iii. Copied back the burned iso-file to my HDD (another partition) and run md5summer again.

And, you know what... when I run compare in md5summer, the files (data) are exactly the same. Down to the last bit.

OK, I´m a practical man with some knowledge of verifying data. And I don´t care much about Nero´s way of data verification...
I just wanna same data back from my burned DVD´s at that day I need it, as they were originaly. Exactly the same!
Maybe some of our members will be able tell why you get "different" checksums when "comparing" files in your way. I can´t.

You can download md5summer here

/off topic/
On request by PM, I have been asked to post one md5 checksum, when 1 (one) value is changed in the big file I used as example above.

OK, I opened the 223MB big ISO-file in texteditor (UltraEdit) and changed as per request.

The attached checksums is what I got. Compare by yourself.
/end of off topic/

Checksumming is designed such that just a single bit being off will result in a radically different checksum. This is why they're so widely used in verifying data integrity.

I didn't use a checksum method or hashing method like md5 to check if the burn is correct.

After burning I used 2 methods to check the dvd:
- when I used Nero for burning I marked the check "Verify Data" and after ending the burning Nero verified the written data; after burning I verified the dvd using a byte-by-byte comparison professional program (Araxis Merge) with the files from hd.
- when I used another burning software (RecordNow, Padus), I only used Araxis Merge for checking.

The checking made by Nero is a byte by byte check for every burned file. The same for Araxis Merge.

Another check was the fact that some files burned was "rar" files. After burning I checked them, and they was corrupted.

So, the check is really correct without doubt.

My question is why this behaviour? The number of errors decreased with every burn from the 7 burns(8->4->1->0->0->0->0), with last the 4 burns without errors (One error is one file corrrupted) ? Is a burner problem, or it was a media problem?

Any thoughts?