cody: I had two bad recordings today, both one hour long programs. When I tried to run the command you gave (substituting actual filenames) it gave me this error:
dd: invalid number `188*8*1024'
I'm not all that proficient in Linux so I typed "dd --help" and learned that the bs value sets a block size. So just for the heck of it I tried specifying "bs=10G" since I figured that should copy the whole file, but it didn't. Instead it said:
0+1 records in
0+1 records out
2147479552 bytes (2.1 GB) copied, 37.1445 s, 57.8 MB/s
So it only got a part of the file. It turned out it was a segment of just over 8 minutes but I could skip around in it with no problems. However it was clearly not from the start of the program.
Since you talked about cutting the first 1 MB from the start, I then tried using the options "bs=1M skip=1 count=999999999" with the goal of getting the entire program except for the first 1MB. This returned:
16309+1 records in
16309+1 records out
17101800224 bytes (17 GB) copied, 326.141 s, 52.4 MB/s
And did indeed produce a file 1 MB shorter than the original. The new file had the correct time and I was able to skip around in it with no problem whatsoever. I repeated that operation on the other bad recording, and the newly created file on that one also played like any normal recording, and I could skip around in it. If I am understanding the dd command options correctly, the ONLY difference was that I skipped the first 1 MB of the original file (and NOTHING at the end because of the outrageously high count value specified, which I'm assuming specified more blocks than the file contained).
You wrote,
cody Wrote:it's possible at beginning and the end due to timing problem some data got corrupted - TVH saves the data faster than the driver send them to the data buffers. if that's confirmed what can be done is we reduce the size of the buffers, that way reduce such latency that can cause such timing problem.
It does seem that skipping the first 1 MB eliminated the problem in at least two tests, so if that gives you enough information to go on I would very much appreciate it if this problem could be fixed in some way. If you need me to run any additional tests, please let me know, but please keep in mind that the dd option "bs=188*8*1024" doesn't seem to work in the version of dd on my system.