Oak Technologies OTI 911 Bug
The problem was that an increasing number of people were complaining that CDDA and CDDA32 were failing with 'Jitter correction errors'. We finally went out and bought one of these 'bad' drives. What we found was that the drive randomly screws up samples in the audio. There are enough of the 'bad' samples, that the jitter correction routines are unable to function. The OTI 911 is the main chip inside a CDROM drive that controls the drive and the IDE interface. Oak Technologies does not manufacture CDROM drives, they just make the chips that most of the rest of the CDROM drive manufacturers use. This particular chip is used by companies like Goldstar, NEC, Hitachi, and others. The OTI910 chip also appears to have the same problems, but the OTI912 and later chips don't seem to have the problem.
Here is a picture of a section of audio that has the problem. You can see
at one point the waveform level goes to 0 for a couple of samples then
picks up as if nothing had happened. This is a BAD thing for a drive to
be doing if you want to get clean audio.We have written a program called CDROMCHK which runs under Win95 and will attempt to check to see if your drive is one of the drives that has the problem. We also have written a lot of code in CDDA32 to try and get around the problem. We cannot 'fix' the bad samples, but can improve things such that the jitter correction errors are much less frequent.
We have been in contact with Oak Technologies, given them a copy of our test program, and after that we lost contact. They now don't return emails or phone calls. We can only assume that we have found a real problem and they are not prepared to deal with it. Maybe by letting the world know about the problem, they will address it.
Note: A few weeks after this page first appeared on the web, I was contacted again by Oak, who wanted to discuss the problem. Their feeling was that the bug might not be in the chipset, but in the firmware. Again, after the initial flurry of activity with them for the first couple of days, they have fallen back into their mode of not returning phone calls and not returning email messages. We are not impressed with the support Oak is supplying.
It may be that there is some explanation other than a bug in the chipset for this problem. We will be the first to update this page if we can be convinced that this is the case. In the mean time we will assume that there is a bug in the chip.
Check the download page for the test program chkcdrb1.zip
