[Howard2k]

I just picked up a new SD card for my K10D and did some testing. I thought I would share the results in case there was interest, and because hopefully someone who knows more about this stuff than I do can provide some great feedback.

My primary card is a 4GB Sandisk Extreme III, which I should point out is SPECTACULARLY better than the regular Sandisk cards, which have given me atrocious reliability problems over the years. The Extreme III is Class 6. I just picked up a 4GB Class 10 Panasonic card.

I did a 20 second burst, plenty to fill the buffer and measure the card speed while the buffer is full and needs to be flushed.

Both cards managed around 28 images. (1 was 27).

Is the Extreme III Class 6 performing well in excess of the Class 6 specs? Or is the K10D just not capable of driving the Class 10 any faster than it can drive the Class 6?

Anyone shed some light?

[JohnBee]

Were those PEF or DNG files?
I found the K20D would give me aprox. 3 frames(depending on the wind conditions) with DNG. While it wouldn't make any real difference with PEF.
I suspect that has to do with the size of the files and the fact that the camera can write enough data during continuous shooting to make a difference.

[wlachan]

First of all, the class rating of SDHC is not to be taken seriously. Their actual read/write speeds can vary greatly between brands, models, and even batches. Also, different file sizes yield different write speed. The same card may also perform differently on different cameras. To get an idea how fast you cards are, you may copy a huge file from your computer via a faster card reader.

[Russell-Evans]

The K10D can write jpegs until the card was full. It is only when shooting RAW that the buffer fills. You should only get 14 or so RAW files before the buffer fills. It is after that that the speed of the card becomes visible. Simply shoot enough to capture 10 RAW files after the buffer fills and then compare the difference between the time stamps in the EXIF data between the two cards.

Thank you
Russell

[Howard2k]

Thanks guys,

wlachan - the Class system has some flaws but it is based on sustained speeds, as opposed to the "100x" type nomenclature which is much looser. So while the Class system is still far from perfect it's significantly better than the alternatives. Great idea about doing the copy on the computer. I suspect that the Class 10 card IS faster, it's just that there is something in the K10d preventing it from utilizing the card to it's full potential. The alternative is that the Extreme III was labelled Class 6 becuase Class 6 was the highest standard when the card was produced and it just happens that although it is labelled Class 6, the performance far exceeds the Class 6 standard. In other words, it's possible that the Class 6 Extreme III is faster than the Class 10 card, and the Class system still works in this case. Strange as that sounds.



Russell - Yep, I was shooting RAW. So during the 20 second burst there was a higher speed burst and then once the buffer filled the speed of the images was reduced. Not hugely compared to some of my other cards, but there is a noticable difference for sure. One of my other cards is supposedly a 66x card and it's very signicantly slower than the Class 6 and Class 10 cards. Though 66x would at best still be less than Class 4.

John Bee - PEF.

[Howard2k]

Wish I had found this article earlier!

Panasonic SD cards hit Class 10. Should you care? | Crave - CNET

[glanglois]

I just picked up a new SD card for my K10D and did some testing. I thought I would share the results in case there was interest, and because hopefully someone who knows more about this stuff than I do can provide some great feedback.

<snip>

Is the Extreme III Class 6 performing well in excess of the Class 6 specs? Or is the K10D just not capable of driving the Class 10 any faster than it can drive the Class 6?

Anyone shed some light?

Yes, the original Extreme III is class six only because that was the highest class at the time that the cards were introduced. It clearly performs well in excess of its class rating. I have 8-10 of those cards in 4 GB and 8 GB sizes. Couldn't be happier.

Visiting the Sandisk site for performance data on the new models, I left very confused. Last I looked, they wrote of maximum (up to) rate but did not specify read and write performance.

Many of us, as in the example above, are more concerned about write speed and that's usually slower than read speed. That's why marketers tend to talk about speed in a general fashion, without specifying read versus write: they want us to think that we're getting a really fast card even if the write is 10 MB/s and the read is 30 MB/s (or whatever).

As we look at system speed, we must remember that it will be no faster than the slowest sub-system. There certainly is a point at which a faster card will provide no speed benefit to the system as a whole. The K10D may well be the limiting factor in your case.

[UnknownVT]

First of all, the class rating of SDHC is not to be taken seriously. Their actual read/write speeds can vary greatly between brands, models, and even batches. Also, different file sizes yield different write speed. The same card may also perform differently on different cameras. To get an idea how fast you cards are, you may copy a huge file from your computer via a faster card reader.

Yes, the original Extreme III is class six only because that was the highest class at the time that the cards were introduced. It clearly performs well in excess of its class rating.

Visiting the Sandisk site for performance data on the new models, I left very confused. Last I looked, they wrote of maximum (up to) rate but did not specify read and write performance.

Many of us, as in the example above, are more concerned about write speed and that's usually slower than read speed. That's why marketers tend to talk about speed in a general fashion, without specifying read versus write: they want us to think that we're getting a really fast card

OK the SD Association Class rating is based on minimum sustained WRITE speed - which is the more important factor when it comes to performance in our cameras.

From: Secure Digital - Wikipedia, the free encyclopedia

" Speed Class Rating

The Speed Class Rating is the official unit of speed measurement for SD Cards, defined by the SD Association. It is equal to 1 MB/s (8 Mbit/s), and it measures the minimum sustained write speeds for a card in a fragmented state.

The following are the ratings of some currently available cards:

* Class 2: 8 Mbit/s = 1MBytes/s(?)
* Class 4: 15 Mbit/s = ~ 2MB/s(?)
* Class 6: 20 Mbit/s = 2.5MB/s(?)
* Class 10: 30 Mbit/s = 3.75MB/s(?)

Even though the class ratings are defined by a governing body, like × speed ratings, class speed ratings are quoted by the manufacturers but unverified by any independent evaluation process. In applications that require sustained write throughput, such as video recording, the device may not perform satisfactorily if the SD card's class rating falls below a particular speed. For example, a camcorder that is designed to record to class 6 media may suffer dropouts or corrupted video on slower media. "

[NOTE: I just noticed the change in the Speed Rating Class is given in Mbits/s - and when I did the calculation (in Red) the MBytes/s rating is much lower than before - I had always understood that the ratings were for Write speeds Class2=2MB/s; Class4=4MB/s; Class6=6MB/s; and Class10=10MB/s - these "new" figures are WAY lower than what I had understood previously and the SD Association were no help as their own page: Speed Class - SD Association no longer shows the actual write speeds!]

Now whether these speed class ratings are trustworthy when they are on an SDHC card label may still be open to question.

I have seen many user reviews that have shown Class 10 cards to test significantly below their rating and many class 6 beating them (and not all SanDisk Extreme III) - the problem with these reports/tests is not the integrity of the users - but the associated equipment used - eg: the USB reader and possibly the USB port itself - both of which can have a significant impact on the test results. Just to give an example I have a portable hard disk that has shown in several tests to get close to 30MB/s for both read and write - but on my PC and USB port I do get close to 30MB/s read, but only just over 20MB/s write - so either I have a faulty drive - or more likely my USB ports may be the limiting factor.

Now back to the question whether faster cards perform better in our cameras.

First it shouldn't do any harm to have a faster card.

However there is a maximum write speed on our cameras (older models may well be slower - see: Post #4 ) and a card that is as fast as or faster than that write speed is going to be fine.

For example if a dSLR's maximum write is as fast as say 6MB/s - then in theory any class 6 card ought to be adequate (based on my previous understanding of the speed classes - see above in []), and although a class 10 will do no harm - but it is overkill and one would not see any gain in performance.

So getting the fastest card on earth will not make our dSLR's go any faster under any circumstances compared to an SDHC of adequate speed.

But there may be other side benefits for example of upload speeds to the computer - a faster READ speed will upload to one's computer faster - pretty obvious - that is provided one's USB reader and USB ports don't become limiting factors.