At 11:48 AM 3/25/2005 -0800, Trey Bunn wrote:
>Mike, your mention of write speeds got my attention.
>I was always told that the slower the write speed, the
>more stable the CD will be (that is, less prone to
>errors).  So whenever I burn anything, I always set it
>to the lowest speed possible (my current configuration
>usually only allows me to go down to 8x or 4x,
>depending on the media).  But is this not a good

Yes, it is not a good thing. The best speed to use is the one which
produces the fewest errors; that sounds simplistic, but it's also true in
my experience and that of others.

>I've noticed that my burning software has an option
>under write speed for "Best," but I've never trusted
>it because I have no idea what standards the person
>who wrote the program (Toast 6 Titanium) used when
>determining what's best.  Best for audio?  Best for
>data integrity?  Best for creating coasters that look
>lovely on your coffee table?  But am I wrong?  Should
>I trust this "Best" setting, or should I stick with my
>practice of burning as slowly as possible?

Neither. I'd recommend determining what speed the software picks by timing
the operation if that's all that you can do. But more to the point, I
highly recommend running your own speed tests and measuring error rate.
That will find the region that works well for your drive and your media.

I use Mitsui silver and T-Y blanks as a compromise between reputation and
cost. I have tested and found both to be error-free in my Plexwriter at
12x, its highest speed. At 8x, they are almost free of errors, far below
the threshold where there would be a problem with error correction. At that
speed, they are just as well written in my Pioneer 107 and 108 combination
units. My normal practice is to write a single disc in the Plexwriter at
12x, multiples in the three drives at 8x. At lower speeds, errors are much
more common and 4x is unacceptable in any of those drives with those blanks.

Of course, YMMV - your mileage may (in fact, will) vary. But none of the
above is meaningful (except 4x and below) without measurement. The
idiot-light approach of writing software is inadequate to determine quality
of the recording. If you want longevity or interoperability, you need to
know more than complete failure in an optimal drive.

[log in to unmask]