SD card speed test - playtime ;)

[CheshireCat]

Advertisement (gone after registration)

Of all cameras that the poster mentioned? Note that he is talking about mirrorless in general, not your favorite flavour of firmware (Leica)

 

Not only Leica's of course

My harsh comments must always be taken with a grain (or grin ) of salt

Please see my previous post.

[pico]

[...] And if faster speed SD cards have the same performance of lower speed ones, then this means the SD card controller in the SL cannot take advantage of fast cards.

 

Given how well defined, and public the host controller code (the camera's part) is, then it seems likely that something other than the host software is to blame. Or am I being too optimistic?

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[CheshireCat]

Given how well defined, and public the host controller code (the camera's part) is, then it seems likely that something other than the host software is to blame. Or am I being too optimistic?

 

I don't own a SL, but assuming 48 MB for an uncompressed DNG, then 11 fps would require a throughput of about 500 MB/s. Honestly, I don't know if the newer cards can achieve that speed.

In any case, before blaming the controller hardware (which might be perfectly fine), we need to blame Leica for not implementing compressed DNG (which would require about 250 MB/s).

Edited January 17, 2016 by CheshireCat

[Jared]

Isn't the DNG truly raw binary anyway? When we read the data for post-processing it is the computer's software that interprets it and as such it does not require great speed.

 

 

Those speeds are advertised by the vendor, and are max READ speeds, not Write.

.

 

 

Actually, the 90MB/s speed I was quoting is the WRITE speed for my card, as advertised by the vendor.  The READ speed is 95MB/s.

 

And the DNG file is definitely not the raw output from the imaging chip itself.  It requires a fair amount of processing to generate a valid DNG with its annotations and embedded preview JPG.  Still a fair amount of processing work.

 

- Jared

[Jared]

 

 

I don't own a SL, but assuming 48 MB for an uncompressed DNG, then 11 fps would require a throughput of about 500 MB/s. Honestly, I don't know if the newer cards can achieve that speed.

In any case, before blaming the controller hardware (which might be perfectly fine), we need to blame Leica for not implementing compressed DNG (which would require about 250 MB/s).

 

 

 

No, I'm not aware of any cards that can handle 500 MB/s in write mode.  Based on the testing I could find on the internet, the Lexar is currently the fastest at something like 280MB/s.  I certainly wouldn't expect the camera to be able to empty the buffer as fast as it fills it (at least with raw images).  My understanding is that the faster write speeds quoted for UHS-II cards are in some sort of "burst" mode.  I'm not certain what sort of sustained throughput is possible.  In any event, my SL appears to be achieving about 30MB/s of sustained write speed.  Even assuming it is not able to "process" files and "write" files in parallel--that it needs to handle these operations sequentially--there's a pretty sizable gap between what the cards are supposed to be capable of and what the SL is achieving.  That being said, it's only about 5MB/s slower than a Nikon D800, and even the Nikon D4 only writes out at about 45MB/s, so it's in the ballpark of what at least Nikon is achieving.  

 

For me, this is purely an academic exercise.  This isn't a sports camera, and 30 or so exposures at 11 fps is more than enough for anything I might be doing.  Other photographers might have different needs.  I'm just curious why there is such a wide gap between what I would have expected for sustained write speeds and what the camera can actually achieve.  Why even bother with UHS-II slots if you can only write out at 30MB/s?

 

- Jared

[CheshireCat]

And the DNG file is definitely not the raw output from the imaging chip itself. It requires a fair amount of processing to generate a valid DNG with its annotations and embedded preview JPG.

 

Depends on your definition of "fair amount".

Anyways, a DNG with its embedded JPEG file requires less processing than a full JPEG file.

[Jared]

Advertisement (gone after registration)

Yes, I agree that is likely true.

[bilbrown]

SO Wow folks, you all are nailing this... IT IS THE SAME PROBLEM I HAVE BEEN HAVING... This makes sense now. IT's the buffer!!!

 

This isn't technical, but this is the feeling I get. I have been able to shoot much faster with my M240 (100 Years edition), and DNG compression ON, than with my SL even frame by frame. The write times are between the 45-60 seconds posters here have mentioned. I also noticed that my DNG files were very hight MP, even LARGER than my Pentax 635Z files in full resolution.... much larger. So, wow. I've changed cards from the Lexar 633x (I) to the Lexar 1000x (II), and it sped it up a slight bit, but still pretty much if I am shooting a subject I have to tell them to WAIT until my camera finished writing...

 

So annoying!!

 

Any other clues on what to do? I could shoot JPG only, I guess.

[ramarren]

- With fw2.1 in the SL, off to "Ready" with adapted lenses seems to be about the same as with the SL24-90 lens: about .5-.75 seconds.  

 

- My SL captures 35 11fps DNG only files before slowing down using Lexar 1000x (UHS II) 128G card in the #1 slot. Card was low-level formatted with SD Card Formatter.

 

- Per my tests in another thread (http://www.l-camera-forum.com/topic/252697-recommended-sd-cards/?p=3043017), best throughput of the SL with this card in JPEG+raw mode is about 23Mbytes per second, about 34% of the card's best write speed. 

 

If you need faster or longer sustained sequence performance than this, you need a different camera, or switch to video mode and capture frames out of that... That's just how it is. The SL wasn't designed to capture full-resolution still frames for sustained lengths of time. 

[pico]

- My SL captures 35 11fps DNG only files before slowing down using Lexar 1000x (UHS II) 128G card in the #1 slot. Card was low-level formatted with SD Card Formatter.

 

By low-level, do you mean overwrite the card? If true, then does it make a significant difference? My understanding is that it cannot make a difference and only adds unnecessary wear.  Am I wrong?

[ramarren]

By low-level, do you mean overwrite the card? If true, then does it make a significant difference? My understanding is that it cannot make a difference and only adds unnecessary wear.  Am I wrong?

When using SD Card Formatter, it is best to prepare cards for use just once by doing the deep format with overwrite. After that, the quick format is sufficient. I find it does make a performance difference, and it does not cost excessive card lifespan.

 

After all, think about it: Flash storage locations are capable of being written more than a million times. Any decent flash storage driver knows how to vary the locations it uses for file writes on volatile data to balance write distribution. The most heavily written section of the flash storage volume is the boot and directory tables themselves, and any formatting program worth calling such knows how to react to an error when writing there and remaps the volume/sector when needed such that the volume remains usable, albeit with a few hundred bytes less storage capacity.

 

In fifteen years and heavy use of digital storage, I've made perhaps three quarters of a million exposures and saved about a third of those. I've not yet worn out a card, and have suffered only two failures in all this time which were not related to sector damage so much as controller line failure.

[Guest satrycon]

i use the SD Card format too on my 5DSR cards once a month..really makes a difference.

[Guest satrycon]

this guide i found might apply to the SL as well http://alikgriffin.com/best-memory-cards-canon-5ds-r

[arvin]

hi all,

 

How did you guys perform the SD test? Using the camera? or software?

I'm interested to test my out also. 

 

Thanks

[pico]

For the Mac there is a utility called SDspeed. It goes through the whole card and takes FOREVER to complete. Worthless, IMHO.

[arvin]

For the Mac there is a utility called SDspeed. It goes through the whole card and takes FOREVER to complete. Worthless, IMHO.

Thanks, I did use a software to test both cards I have.

 

One is Sandisk Extreme Pro 128GB UHS-1

And the other is SanDisk Extreme Pro 64GB UHS-II

 

However, both returns the same speed, write :70-80MB/S read: 90-91MB/S

 

The above test is carried out in the SD slot on my iMac 27", assuming it should be USB3.

 

Is this speed normal? When I take photo, the red light flashes. How should it behalf? Thanks a lot.

 

Future test :

(Sandisk Extreme Pro 128Gb UHS-1)

Drive mode: Single, Slot -2

It flashes six times every time I took a photo.

 

(SanDisk Extreme Pro 64GB UHS-2)

Drive mode: Single, Slot - 1

It flashes five times.

 

Is this normal?

 

 

 

Sent from my iPhone using Tapatalk

Edited March 26, 2017 by arvin

[arvin]

Thanks, I did use a software to test both cards I have.

 

One is Sandisk Extreme Pro 128GB UHS-1

And the other is SanDisk Extreme Pro 64GB UHS-II

 

Both SD cards performs correctly on a PC. However, when I take photo, the red light on the camera flashes five times for the 64GB card and six times for the 128GB card.

 

Is this a normal behaviour for the SL?

 

Sent from my iPhone using Tapatalk

 

Sent from my iPhone using Tapatalk