My SD card benchmarks.

[CuriousMike]

In the past couple of months, I've run into "FULL" 100% more than I have in my past 15 years of digital photography.

Typically, I use dual 16GB SD cards; the Nikon D610 has 2-card slots that you can use in a variety of ways.

[ As an aside: I tend to use slot 1 as RAW and slot 2 as high quality JPG - it is not written in parallel, rather sequentially. This means there is a penalty for burst shooting if using both slots. ]

I decided to purchase a couple of 32GB cards, and was looking for a set of value oriented cards. It's obvious that the Sandisk Extreme's are the fastest, but the cases where I've gone full haven't been when I've been particularly concerned with the best overall write speed - I just need good write speed.

Reviews on a Transcend branded card seemed "good enough", so I received the cards today.
Curiosity got the better of me, so I performed a Crystal Disk benchmark on a handful of cards I own.

I'm presenting the results below, along with the current USD price for each card from the 'zon.

I have a pair of these that were my backups to my Sandisks'.

The new guys

My go to budget card - these grace my other cameras and any other devices needing SD

My latest 16GB Sandisk Extreme Pro, which has a slightly different design than my original - purchased 6 months apart

My original 16 GB Sandisk Exterme Pro

 

[Syborg1211]

Gotta love the Extreme Pros!

On a side note, do you take your photos off the card and format them after ever shoot, or do you let them accumulate over multiple outings? I haven't filled up a card on a single outing in quite some time, but I've been consciously making an effort to reduce total shot count and "make each count" more.

 

[EOM]

MAAAAN..... i wish you'd have posted this yesterday! I just ordered http://www.amazon.com/gp/product/B00MBFPT44?psc=1&redirect=true&ref_=oh_aui_detailpage_o00_s00 this guy yesterday... Sandisk extreme (not pro) so it's 60mb read and 40mb write..... looks like i should have done with either the transcend or the Extreme Pro!

 

[CuriousMike]

Gotta love the Extreme Pros!

On a side note, do you take your photos off the card and format them after ever shoot,

Excepting the Samsungs, the Extreme Pro's are over twice as fast as writing as every other card I have.

I think it also goes to show the "90MB/sec" emblanzened on all the cards is for read speeds only - except the Extreme Pro's.

99% of the time, I empty my cards after each shoot. It's very rare that anything accumulates.

 

[CuriousMike]

MAAAAN..... i wish you'd have posted this yesterday! I just ordered http://www.amazon.com/gp/product/B00MBFPT44?psc=1&redirect=true&ref_=oh_aui_detailpage_o00_s00 this guy yesterday... Sandisk extreme (not pro) so it's 60mb read and 40mb write..... looks like i should have done with either the transcend or the Extreme Pro!

For a value card, the Transcend is OK.

I'm not exactly sure how often TWICE AS FAST WRITES will help me out, but TWICE AS FAST WRITES.

 

[EOM]

Can I inquire as to your reasoning to use Raw in Slot 1 and JPG in Slot 2 as opposed to shooting Raw and using the slots as a mirror? I assume you are going to bring them all home and import into lightroom anyway right?


I have a D7000 that I just got today as an upgrade to my D5000 and plan to shoot Slot 1 raw and Slot 2 as normal/medium JPG. Only reason for that is Slot 1 is a 64gb card and Slot 2 is an 8gb eye-fi card that backs up automatically to my server when i'm out shooting and transferring jpg is MUCH faster than RAW's over 4g. (i'll be making a post about that soon!) ...

 

[CuriousMike]

Can I inquire as to your reasoning to use Raw in Slot 1 and JPG in Slot 2 as opposed to shooting Raw and using the slots as a mirror?

My reasoning up until recently has been two-fold:
1. Slot 2 is my emergency backup. Since the writes aren't in parallel and there is a small penalty for the duplicate write, I want to make sure that write time is as small as possible, and 5-7MB JPG's are faster to write than 25-30MB RAW's.
2. On my home PC, I have always had some form of flickr or GooglePhotos auto-backup running in the background (and they only support JPG). Again, I use these services as emergency. The RAW's get backed-up onto external USB hard disks as semi-irregular intervals.

With that said, I've become more tempted lately to do the mirror for a few reasons.
1. The backup becomes a full-quality backup. Duh.
2. When copying files off the cards, I only need to copy from one card. A small thing, but it skips the step of my copying both cards onto the computer.
3. For my best shots, I end up editing them in LightRoom and exporting them at high quality onto flickr/smugmug -- so, in essence, I end up with emergency backups. At least for those photos I deemed good.

 

[CuriousMike]

I have a D7000 that I just got today as an upgrade to my D5000 and plan to shoot Slot 1 raw and Slot 2 as normal/medium JPG.

I forgot to say congrats - you're going to have fun with that D7000. The sensor in the D5100/D7000 was what I considered a big jump. The extra megapixels are nice for cropping and the high-iso abilities are useful as well... but your D3200 is likely on par ! Yay for more direct controls !

 

[carlton_fritz]

I use Sandisk Extreme Pro or Lexar Pro. All 32Gig. I have a 1000x Lexar professional card that is waiting on a twin before I put it in my camera. I got a deal like this, but cheaper.

 

[CuriousMike]

Update.
I recently purchased a Lexar 2000X 32GB SD card - it uses the newer SD card standard where there are essentially twice as many pins on the card. It came with a USB 3 reader to support the format.
Note that the card is fully backward compatible with all my cameras... none of my cameras actually support the faster read/write speeds.
However, the real benefit comes when copying the card onto my computer - it's faaast.
The card comes with the reader for about $50 - so, there's a big premium cost associated with the speed. Keep your eyes open for them to drop in price.

Also, I just bought a pair of 64GB Patriots for cheap... these are going to be used where space matters and not speed - i.e., for doing timelapses. I was surprised how well they did considering a) 64GB and b) $20.

Also also note that I've gone from CrystalDiskMark 4.1 to 5.1 in these tests - it appears to still be an apple-to-apple comparison

 

[BrainEater]

Wish I'd seen this when you first posted it.

I use micro SD cards in my gopro4 black.I've gone through all the testing you talk about trying to film decent 4k/30 video.

The ONLY card that has ever worked properly is my Sandisk Extreme .

:thumbsup:

 

[AkumaX]

Update.
I recently purchased a Lexar 2000X 32GB SD card - it uses the newer SD card standard where there are essentially twice as many pins on the card. It came with a USB 3 reader to support the format.
Note that the card is fully backward compatible with all my cameras... none of my cameras actually support the faster read/write speeds.
However, the real benefit comes when copying the card onto my computer - it's faaast.
The card comes with the reader for about $50 - so, there's a big premium cost associated with the speed. Keep your eyes open for them to drop in price.

Also, I just bought a pair of 64GB Patriots for cheap... these are going to be used where space matters and not speed - i.e., for doing timelapses. I was surprised how well they did considering a) 64GB and b) $20.

Also also note that I've gone from CrystalDiskMark 4.1 to 5.1 in these tests - it appears to still be an apple-to-apple comparison

Haha why you gotta bring a UHS-2 card to this fight

You like how in the previous test, the Extreme Pro U1 and U3 card pretty much performed similarly? It's like they all they did was slap a U3 label on the U1 card lol.

Look out for the new SD card association standard -- the V-series!

All minimum sequential writes per second:

C = Class series. C2/C4/C6/C10 = 2MB/4MB/6MB/10MB/s
U = UHS Speed Class series. U1/U3 = 10MB/30MB/s
V = Video series. V6/V10/V30/V60/V90 = 6MB/10MB/30MB/60MB/90MB/s

 

[jhansman]

Don't want to throw in a monkey wrench, but I'm thinking the camera's firmware plays a role in how quickly it can write out a file, raw or JPEG. Keep in mind that the camera must convert the initial captured data into a JPEG, and almost all cameras now apply sharpening, color management, etc in in that write out process, all of which take time, albeit very quickly. If your camera can buffer, say, 10 shots before writing the data out to the card, I would assume (never a safe practice, I know) that it does so as quickly as can be engineered. Just a thought.

And yeah, the D7000 is a great APS-C camera. It took me three lesser models before I got mine, and I love it.

 

[AkumaX]

Don't want to throw in a monkey wrench, but I'm thinking the camera's firmware plays a role in how quickly it can write out a file, raw or JPEG. Keep in mind that the camera must convert the initial captured data into a JPEG, and almost all cameras now apply sharpening, color management, etc in in that write out process, all of which take time, albeit very quickly. If your camera can buffer, say, 10 shots before writing the data out to the card, I would assume (never a safe practice, I know) that it does so as quickly as can be engineered. Just a thought.

And yeah, the D7000 is a great APS-C camera. It took me three lesser models before I got mine, and I love it.

Wouldn't it start writing to the card the moment it's ready (if JPEG, then sharpening/color mgmt/NR/etc...)? So if you took 10 JPEG shots in 2 seconds, I would say at least 4-5 of the shots have been written to the card after the 2 seconds.

I like this video showing the difference in camera buffer writing vs different SD cards (uses Sony)

He uses a slow Sony C10, but you get the gist of it

https://www.youtube.com/watch?v=GkgNgXqYbqk

Also

https://www.youtube.com/watch?v=CdaRPxqlT7g