On Fri, 10 Jul 2009 09:40:04 -0700, John Navas
wrote:

On Fri, 10 Jul 2009 06:25:59 GMT, "David J Taylor"
wrote in
:

John Turco wrote:

I use a mixture of 2GB SD and 4GB SD-HC now. For my purposes I haven't
found card speed to be a limiting factor - the camera buffer is enough,
although the faster reading of higher-speed cards and an optimised USB 2.0
hi-speed setup is nice.

Depends on the camera, of course, and how you use it.

While there are lots of times when slower cards make no difference,
I saw enough differences between a middle-of-the-road Kingston card and
SanDisk Extreme III with my Panasonic DMC-FZ28, especially in critical
situations, that I've now switched entirely to the latter, and settled
on 8 GB as my own price performance sweet spot, in part because I can
also use these cards as super fast computer DVD replacements.

Flash card speeds tends to go down as cards get larger, making it more
important to have a faster card at higher capacities.

Lastly, if you're going to use a fast card like the SanDisk Extreme III
with a card reader(writer), be warned that not all SDHC readers are
created equal -- many will severely limit speeds, which is why I also
use and recommend the fast SanDisk MicroMate SDHC Reader.

Good points. There's lots of variables involved in which the card speed
might or might not be important. As well as the capabilities of any
card-reader if using that to access your files. Keep in mind too that
advertised speed rates of cards are their READ speeds which is always
faster, not their WRITE speeds. There is often a wide difference between
the two. You have to buy and test them to find out the true write speeds,
or find benchmark lists online by those who have tested cards for their
write speeds.

Speed of the camera's own firmware writing to card being one bottleneck.

[For some interesting information: See the difference in the very same SD
cards being used on various camera platforms and how their speeds change
from Digic II, III, and IV camera processors.
http://chdk.wikia.com/wiki/Benchmarks The interesting part of these tests
as these are done in-camera. Direct camera to card benchmarks. How the
camera will see and use them independent of any card-reader and computer
CPU speed tests. Also note the difference in speeds between the very same
cards being formatted in FAT16 or FAT32. FAT16 often affording much faster
access times, ~25% faster. That's like going from Class-4 to Class-6, for
free.]

If burst-modes are often needed.

Size of files being saved. JPG? RAW? etc.

If video mode is important. What audio-sampling rate you have chosen for
videos. If you can adjust that video quality bandwidth over a wide range as
in CHDK supported cameras.

There are just way too many variables for anyone to be able to give advice
to another on what speed of card they should get for their camera. Camera
platform, camera usage, shooting styles, no two people's needs will be
identical.


Some interesting findings:

Oddly, and going against all known information. I have found that if I am
needing a fast burst rate for some experiment and the card starts to
bog-down to 2/3rds to 1/2 its beginning write-speed after some 50 to 100
frames. If I take out that card and use Window's Defragmentation utility on
it (in a card reader), then all original access speed returns. Normal
3rd-party defraggers seem to have trouble recognizing removable media. I've
tried quite a few to see if others would work. You might have to perform
this twice, as it seems like the 1st-run defragmentation isn't the same as
subsequent ones. After 2 or 3 defragmentations Windows' defragger is no
longer able to change any of he file structure on the card and there is no
added benefit by doing this again.

This shouldn't be true, due to the random-access nature of memory cards.
And indeed you'll read this cut-in-stone information everywhere on the net.
That "Defragmentation on any memory card is useless and only detracts from
the available R/W life of the card." As I also read when I went looking for
an explanation to this oddity. But I've run into this problem so often in
the past and used defragmentation as a simple fix each time, that I'm now
convinced that there is a fragmentation bottle-neck on random-access memory
cards. Anecdotal evidence, but one of which I've now convinced myself
through real-world tests and uses, in spite of all the "learned" advice all
over the net. (Though this wouldn't be the first time, far from it, that
some common advice all over the internet posted by self-proclaimed experts
has been proven wrong by my own tests and findings.)

Ignore all the advice you read online and try Window's defragger the next
time your card is acting sluggish. See if it doesn't restore fast access
time again. It does on all of the ones I've tried this on. I don't need to
do this often, but after using an SD card for multiple uses (MP3 files,
portable-apps, etc.) if there has been a lot of file changes on the card
between photography sessions, then I'll snug-up all the files again after I
have spooled off all the images. Leaving the next new camera images for the
large clear chunk beyond all the files already retained on the card. I
would simply do an in-camera reformat but my cards are used for many
devices and multiple uses. The CHDK camera cards also keep the CHDK
operating system on them (with boot code, settings, special FAT16
formatting for extra speed, scripts, e-books, etc.). Reformat is not very
practical for my needs. Defrag SD to the rescue.