FAQ FAQ  Forum Search   Events   Register Register  Login Login

Practical comparison of fast and slow cards

Page  12>
Author
Bob J View Drop Down
Admin Group
Admin Group
Dyxum Administrator

Joined: 23 December 2005
Country: United Kingdom
Location: London
Status: Offline
Posts: 25573
Post Options Post Options   Quote Bob J Quote  Post ReplyReply Direct Link To This Post Topic: Practical comparison of fast and slow cards
    Posted: 04 June 2009 at 12:49
Some of you may be aware that since I got my a700 I’ve done various comparison timings of Compact Flash and MS cards in both the a700 and KM5D.

My timings had concentrated on trying to estimate what sort of MB/s the card/camera combinations were achieving – but of course, they only measured how fast the camera wrote to the card, and took no account of the effect of the camera’s built-in memory buffer. This led me to try to find a ‘real world’ test that could be applied.

It seemed to me that the biggest effect of write speed should be on the length of burst that a camera is capable of before it needs to slow down for the card. A faster card will clear the first data into the buffer quicker and will allow for more shots to be added before it hits the wall.

I decided to carry out some comparative tests shooting at the highest continuous speed available on the camera to see how many shots could be taken before the frame rate dropped for various card/quality settings.

To give a reasonable level playing field I set up the camera for aperture priority, lens wide open, pointed at a bright scene (so shutter speeds were 1/1000 or above) ISO 200 and manual focus (adjusted so the lens was in focus).

I then carried out the same tests on a fast CF card (Kingston Ultimate clocked at over 30Mb/s), a slower CF card (1st generation Extreme III, clocked at about 15MB/s) and a Sandisk MS card (clocked at just under 10MB/s).

The results I came up with were quite surprising (not to say puzzling) and enlightening.

The fast card was able to take 10 shots on RAW+ Jpg, before slowing, which was exactly the same number of shots that the slower CF card – AND the MS could manage.


[edit]At the suggestion of ArtUK, I tried this test again with DRO off - for the fast Kingston card, the a700 now took 12 shots before any pause, the Memory Stick however, still could only manage 10 shots[edit]

When shooting just RAW, there was more of a difference, with the fast CF card managing 15 shots, the slower CF doing 14 and the MS managing 13 – but still far less difference than I had been expecting.

The biggest surprise for me (and something I don’t fully understand) was the results for shooting extra-fine JPG, where all cards slowed after 4 shots(!?).


[edit]...all of which would seem to be down to using DRO. Without DRO the fast Kingston card rattled away with no sign of any pause and I game up after 150+ shots had been taken. The Memory Stick managed 47 Extra-fine JPGs before any pause.[edit]

I tried a similar test on the KM 5D for comparison, which showed the fast Kingston having no real-world speed advantage over a slower card, with each managing 3 RAW+, 5 RAW and 8 x-fine JPG before slowing. I did further tests using the slower CF card and the lower quality JPG settings and got 20 shots before slowing on ‘fine’. On standard, the KM 5D could shoot at top speed continually until the card was full.

The one other area I have found a practical difference between fast and slow cards is in image review – you don’t seem to get an image on screen after having taken a shot until the card stops writing, so you can wait some time for the last shot to show – there may also be a greater drain on the battery with a slow card as the camera is having to maintain the buffer and write to the card for longer, but I can’t think of a practical way to measure this.

There are still some unknowns here for me – I’m surprised that there is so little difference in RAW continuous shots (15 against 13) between my fastest card and one that I reckon is about 1/3 of the speed.

Doing this set of tests has changed my priorities about memory cards though; my priorities had been for quality supplier, followed by card speed and then capacity. I think now I would look for quality supplier, capacity then speed.

I don’t put these conclusions forward as definitive – I’d also be interested in other people’s real-world tests of any cards they have to see if they back-up or contradict what I found.

Bob


Edited by Bob J - 04 June 2009 at 15:11
 



Back to Top
artuk View Drop Down
Senior Member
Senior Member

Joined: 06 July 2007
Country: United Kingdom
Location: United Kingdom
Status: Offline
Posts: 3751
Post Options Post Options   Quote artuk Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 13:40
Very interesting Bob.

I don't know if it still applies, but Kingston cards never gave good performance in KM bodies ad didn't achieve anything like their claimed speed, but it seemed to be a camera/card interaction as other cards did better in the bodies, and the cards did better in other brand cameras.

The JPEG issue... was DRO on? This slows down continuous shooting alot since the DRO is applied to the image whilst it is in the buffer. It must be this, or some other A700 setting, that is only giving you a 4 frame continuous JPEG burst depth.

If you want JPEG files and want fast shooting in Sony bodies, always turn DRO off.

Edited by artuk - 04 June 2009 at 13:43
Art
Back to Top
vabijou View Drop Down
Groupie
Groupie

Joined: 08 August 2008
Location: United States
Status: Offline
Posts: 96
Post Options Post Options   Quote vabijou Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 13:47
I wonder if the cameras use the memory cards as temporary RAM during the jpeg processing? If so, maybe jpeg file processing isn't a one-way street and other differences become more apparent?
Back to Top
dilettante View Drop Down
Senior Member
Senior Member

Joined: 02 April 2007
Country: United Kingdom
Location: United Kingdom
Status: Offline
Posts: 2973
Post Options Post Options   Quote dilettante Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 14:02

The fast card was able to take 10 shots on RAW+ Jpg, before slowing, which was exactly the same number of shots that the slower CF card – AND the MS could manage.
When shooting just RAW, there was more of a difference, with the fast CF card managing 15 shots, the slower CF doing 14 and the MS managing 13 – but still far less difference than I had been expecting.


This is what you'd expect isn't it? The camera slows down when the buffer's full, and the number of shots before the buffer fills is independent of card. Yes a faster card will drain the buffer quicker than a slower one, but it still takes a significant time per frame even on a fast card. So you shouldn't expect to see a huge difference in the 'time to slow down' measurement. The difference between fast and slow cards would be in how much it slows down.

The biggest surprise for me (and something I don’t fully understand) was the results for shooting extra-fine JPG, where all cards slowed after 4 shots(!?).


Perhaps the extra fine JPEG processing uses a lot of buffer RAM for working space, i.e. the buffer gets used up more quickly. Again, the fact that it's the same for all cards shouldn't be a surprise. The 'time before slowdown' doesn't tell you much about card speeds, as the card's not being used before slowdown.
Back to Top
tpetpe View Drop Down
Senior Member
Senior Member

Joined: 30 December 2006
Location: Denmark
Status: Offline
Posts: 1472
Post Options Post Options   Quote tpetpe Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 14:39
Wouldnt writing jpgs take a lot of processor power? In which case the time taken to process the compression for a .jpg would be the limiting factor. Perhaps another thing would be to see if there is any difference after the buffer has been used, for example take 20 shots on just raw with each card and see if there is a difference then. It could be that after the 10 shots filling the buffer have been taken that the speed of the card becomes more apparent, so the advantage in having a fater card would be you dont have to wait so long before you can get that 11th shot?

tim
http://www.scientificillustration.net
Back to Top
Bob J View Drop Down
Admin Group
Admin Group
Dyxum Administrator

Joined: 23 December 2005
Country: United Kingdom
Location: London
Status: Offline
Posts: 25573
Post Options Post Options   Quote Bob J Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 14:47
Originally posted by artuk artuk wrote:

The JPEG issue... was DRO on? This slows down continuous shooting alot since the DRO is applied to the image whilst it is in the buffer. It must be this, or some other A700 setting, that is only giving you a 4 frame continuous JPEG burst depth.


Spot on! When I retested with DRO off, the a700 clatters on and shows no sign of ever stopping - so that deals with that puzzlement!

 



Back to Top
Blair7 View Drop Down
Senior Member
Senior Member

Joined: 11 October 2008
Location: Canada
Status: Offline
Posts: 547
Post Options Post Options   Quote Blair7 Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 14:52
I have been using what I thought were the slower cards in the 7D and my newer faster cards in the A700. Interesting test will keep me thinking.

Thanks for posting.
Sony A700, A6000. 16-50, 55-210 + numerous Minolta MD lenses. Left A mount for Nikon DSLR after more than 30 years with Minolta/Sony.
Back to Top
pubetter View Drop Down
Senior Member
Senior Member

Joined: 21 February 2008
Country: United Kingdom
Location: Denmark
Status: Offline
Posts: 230
Post Options Post Options   Quote pubetter Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 15:03
I remember how I counted aproximately 25 shots in RAW before my Sandisk Ducati slowed down... and it was a bit faster than my Kingston 266x (20 shot or so), but I'll test it for you if interested?

I think that you should go "pure raw"...

Back to Top
Bob J View Drop Down
Admin Group
Admin Group
Dyxum Administrator

Joined: 23 December 2005
Country: United Kingdom
Location: London
Status: Offline
Posts: 25573
Post Options Post Options   Quote Bob J Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 15:18
I did do just RAW shots, and that was where the biggest difference was visible, but the best I got out of my Kingston 266x was 15 shots on high speed; any comparisons would be useful as if there are differences we can try and work out why the differences exist.

Back to Top
kefkafloyd View Drop Down
Senior Member
Senior Member

Joined: 04 July 2006
Country: United States
Location: Massachusetts
Status: Offline
Posts: 2451
Post Options Post Options   Quote kefkafloyd Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 17:17
In my case, I have a 4GB 266X UDMA PNY card and a 8GB 133X Lexar Professional card.

The 266X UDMA card could go about 18 RAW frames before a slowdown. The slowdown was then more like 3 FPS. Still pretty usable. cRAW even more.

The 133X card, though, would go to about 15 frames, then slow to about 1 frame per second. Very, very slow, almost like my KM 5D when it hit the buffer wall.

The slower card was also not good for reviewing/deleting images after a burst. It took much longer to finish spooling to the card. I actually had to wait for it. I never have this trouble with the DMA card, I can review instantly and delete right after finishing a burst with no issues.
Back to Top
stiuskr View Drop Down
Moderator Group
Moderator Group

Joined: 01 September 2006
Country: United States
Location: West Virginia
Status: Offline
Posts: 11376
Post Options Post Options   Quote stiuskr Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 18:46
To my thinking with the bigger cards (4 gig +) I'm wanting a faster card as much for quick uploading to the PC in addition to max. number of burst shots.
Rob Suits Jr.
a99M2 a99 a77 a700 KM7D|Min24/2.8 Min35/2 So50/1.4 So50/2.8 Min85/1.4G Tam90/2.8 Tam180/3.5|Tam17-50 CZ24-70G2 KM28-75D So70-200G1 So70-300G So70-400G1| SonyF60 AD200R2
Back to Top
bharnois View Drop Down
Senior Member
Senior Member

Joined: 20 April 2006
Country: United States
Location: USA (RI)
Status: Offline
Posts: 4583
Post Options Post Options   Quote bharnois Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 19:03
Slower cards and bursting RAW in the A700 is not very pretty. cRAW helps but not as much as a faster card. All of the cards I use in the '700 are 266X or better. I only did this after using Sandisk II's which were horrible. I save those for the 5D. If I think I'll be taking very long bursts I switch to Xtra Fine JPG.
Just luvin' DYXUM.
Back to Top
Gabriel View Drop Down
Senior Member
Senior Member
Emeritus Member

Joined: 05 December 2006
Location: France
Status: Offline
Posts: 1931
Post Options Post Options   Quote Gabriel Quote  Post ReplyReply Direct Link To This Post Posted: 04 June 2009 at 19:28
Could card speed have an effect on camera startup time?
On my KM5D, I had a fast 2GB card and a cheap, slow 8GB card (this last one was used when traveling)

There was a very significant difference in startup time with both cards, but I never knew if this was because of the size difference or the speed difference. (and in both cases, I still don't understand how either card size or card speed can influence startup time)
Pic a day 2008 /2009/2010 galleries
Back to Top
TCav View Drop Down
Newbie
Newbie

Joined: 25 April 2007
Country: United States
Location: Washington, DC
Status: Offline
Posts: 29
Post Options Post Options   Quote TCav Quote  Post ReplyReply Direct Link To This Post Posted: 11 June 2009 at 23:51
Originally posted by Gabriel Gabriel wrote:

Could card speed have an effect on camera startup time?
On my KM5D, I had a fast 2GB card and a cheap, slow 8GB card (this last one was used when traveling)

There was a very significant difference in startup time with both cards, but I never knew if this was because of the size difference or the speed difference. (and in both cases, I still don't understand how either card size or card speed can influence startup time)


During startup, the camera does an inventory of the card. Larger cards have more space to inventory, and slower cards take longer to inventory. So the camera will take longer to start with a large slow card than a small fast card.
Back to Top
Dyxum main page >  Forum Home > Equipment forums > Other photographic topics Page  12>

Forum Jump Forum Permissions View Drop Down



This page was generated in 0.125 seconds.

Monitor calibration strip

Dyxum.com - Home of the alpha system photographer

In memory of Cameron Hill - brettania

Feel free to contact us if needed.