Write speeds of D800 (or any other camera with two slots)

FlowtographyBerlinFlowtographyBerlin Posts: 477Member
edited February 2014 in Nikon DSLR cameras
Hello,

So I just recently had the situation again that I was shooting RAW and fired away, ending up using the full buffer capacity and noticed that I had to wait quite a bit until it cleared fully. So, I checked this again and think I found out that, when you use the backup mode (write on both cards), the write times simply add up. As in: The camera doesn't write to both cards simultaneously. Is this correct?

I have a SanDisk Extreme Pro 95MB/s card in the SD slot, which has a real write speed of approx 64 MB/s (I also checked this with an app on my computer)
And a SanDisk Extreme Pro 160MB/s in the CF slot, it has a real write speed of approx 91 MB/s

When I use just the normal mode (overflow?), and it writes to one card only, it takes 10 seconds with the SD as the primary card and 7 seconds with the CF as the primary card for the buffer to completely empty again (the display shows 16 images capacity). When I use backup mode, it takes 17 seconds until the buffer is empty. I.e. the write times add up.

For those of you that have two memory card slots in their cameras, is this the same with your camera?

Thanks,
Flow

Comments

  • PB_PMPB_PM Posts: 4,494Member
    Yes, my D800 behaves the same way The camera doesn't write to both at the same time, so naturally it will take longer. The camera also writes data based on the speed of the slowest card, something to keep in mind.
    If I take a good photo it's not my camera's fault.
  • Thanks for your feedback, @PB_PM!
    The camera also writes data based on the speed of the slowest card, something to keep in mind.
    No, that's exactly what it doesn't do, because that's what I thought as well. It writes to both cards sequentially, and the times add up, the faster card PLUS the slower card, not just the time of the slower card.
  • PB_PMPB_PM Posts: 4,494Member
    edited February 2014
    No, you missed the point here. The camera writes to both cards, one at a time, at the speed speed of the slowest card.

    It's like using two different speed RAM chips in your computer, the computer reads and writes to the RAM at the speed of the slowest chip in the chain.
    Post edited by PB_PM on
    If I take a good photo it's not my camera's fault.
  • Ah ok that's what you mean. Well, also that's not the case with my cards, I just checked again. it's 7 seconds for the fast, 10 seconds for the slow card, and in Both-Cards mode, it's 17 seconds, not 20. At least with my camera. But that's bad enough :-)
  • PB_PMPB_PM Posts: 4,494Member
    I don't really notice a difference, since my CF and SD cards are both rated at the same speed.
    If I take a good photo it's not my camera's fault.
  • heartyfisherheartyfisher Posts: 3,186Member
    That makes sense @flotography thanks for sharing. What your experiment shows is consistent with what we have known for a wile but with no test results to prove it. This is good info. Thanks
    Moments of Light - D610 D7K S5pro 70-200f4 18-200 150f2.8 12-24 18-70 35-70f2.8 : C&C very welcome!
    Being a photographer is a lot like being a Christian: Some people look at you funny but do not see the amazing beauty all around them - heartyfisher.

  • TaoTeJaredTaoTeJared Posts: 1,306Member
    When I use just the normal mode (overflow?), and it writes to one card only, it takes 10 seconds with the SD as the primary card and 7 seconds with the CF as the primary card for the buffer to completely empty again (the display shows 16 images capacity). When I use backup mode, it takes 17 seconds until the buffer is empty. I.e. the write times add up.
    That seems rather slow to me. I'll shoot Raw and fill the buffer and can keep shooting in a couple of seconds at full speed again. I never timed it as I'm only worried about at what point I can fire away again, not until it fully clears.

    What file settings are you trying to store at? Are you trying to shoot Raw + Jpeg fine/basic/medium? Do you need too?

    This is one of the reasons I don't understand those who shoot both only so there is a Raw file "just in case needing to do super edits" on. With the D800's files being so large to me, it is just a waste of space. I'll shoot Raw mostly for work, but if the ISO will always be below 640, I'll just shoot Jpegs since the quality is so good. That is situation dependent though.
    D800, D300, D50(ir converted), FujiX100, Canon G11, Olympus TG2. Nikon lenses - 24mm 2.8, 35mm 1.8, (5 in all)50mm, 60mm, 85mm 1.8, 105vr, 105 f2.5, 180mm 2.8, 70-200vr1, 24-120vr f4. Tokina 12-24mm, 16-28mm, 28-70mm (angenieux design), 300mm f2.8. Sigma 15mm fisheye. Voigtlander R2 (olive) & R2a, Voigt 35mm 2.5, Zeiss 50mm f/2, Leica 90mm f/4. I know I missed something...
  • PapermanPaperman Posts: 469Member
    In short, is one better off with only one card if shooting high fps continuous ? That is a surprise to me ...
  • PB_PMPB_PM Posts: 4,494Member
    If you are shooting fast action, the backup setting will slow down the flushing of the buffer.
    If I take a good photo it's not my camera's fault.
  • AdeAde Posts: 1,071Member
    Very interesting Flow, I never knew this... I always just assumed the cards were written in parallel.

    iFixit has a teardown of a D600 and its mainboard shows only one storage controller for the camera's dual SD slots. So I guess physically it makes sense that the writes will be in sequential order, or at most sequentially pipelined.

    Can anyone with a D4 verify if the same results apply with the XQD + CF slots?
  • I don't really notice a difference, since my CF and SD cards are both rated at the same speed.
    Thanks, PB_PM. Nevertheless, the rating often doesn't correspond to reality. I don't have an SD and CF card with the same rating here, though, to compare. Which is also due to the fact that there are no SD cards that are as fast as the faster CF cards on the market.
    In short, is one better off with only one card if shooting high fps continuous ? That is a surprise to me
    Yes, that's exactly what it means. Ironically, it's exactly the continuous shooting situations when I want the backup mode, i.e. concerts or events.
    That seems rather slow to me. I'll shoot Raw and fill the buffer and can keep shooting in a couple of seconds at full speed again. I never timed it as I'm only worried about at what point I can fire away again, not until it fully clears.

    What file settings are you trying to store at? Are you trying to shoot Raw + Jpeg fine/basic/medium? Do you need too?

    This is one of the reasons I don't understand those who shoot both only so there is a Raw file "just in case needing to do super edits" on. With the D800's files being so large to me, it is just a waste of space. I'll shoot Raw mostly for work, but if the ISO will always be below 640, I'll just shoot Jpegs since the quality is so good. That is situation dependent though.
    Totally agree with your JPEG+RAW statement. In those situations where I shoot a lot of pictures, I usually do the same as you do, shoot JPEG. But well, sometimes I choose RAW.

    I don't think it's slow, real 90MB/s for the CF and 60MB/s for the SD is about as fast as it can get, actually. Well except for the generation My setting for RAW is 14bit lossless compression, ending up in ~40MB files. I never shoot RAW+JPEG, I either choose RAW or JPEG, and when I shoot RAW, I don'T see the point in either compromising by lossy compression nor bit depth (nor do I see the point of not applying lossless compression).
    Very interesting Flow, I never knew this... I always just assumed the cards were written in parallel.

    iFixit has a teardown of a D600 and its mainboard shows only one storage controller for the camera's dual SD slots. So I guess physically it makes sense that the writes will be in sequential order, or at most sequentially pipelined.
    Yeah, that's also what I thought. Good to know that you guys are surprised as well. No separate controller means it's just adding another drive, but not adding the capacity to handle it properly in the way the mode suggests, which is parallel, not sequential or sequentially pipelined.

    I'd be interested in the info from the D4 people as well. You can test this like this:
    • Since the D4 images are smaller and I'd assume the D4 buffer is at least bigger, you might want to set the largest file size, i.e. RAW 14bit uncompressed
    • Set continous mode (CH)
    • Set the card handling mode to overflow (i.e. only writing on one card)
    • Check out what the max image number of the buffer is (should be 69, according to the manual)
    • Choose the CF card as primary slot
    • Hold down the shutter and let it fire away its 1000fps or whatever that beast does :-) until the buffer is full and the display says "r00"
    • Release the shutter and measure the time until it reaches 69 capacity again
    • Now set the XQD slot as primary and repeat the procedure.
    • Now set the mode from Overflow to Backup and do the same thing again.
    Thanks!
  • IronheartIronheart Posts: 3,017Moderator
    Flow, I assume you know you don't have to wait for the entire buffer to flush before you can start taking pictures again. You are just using this to determine if they are written to in parallel or not.
  • WestEndBoyWestEndBoy Posts: 1,456Member
    edited February 2014
    I was under the impression that it wrote to both cards in parallel to. Currently I have the fastest SD and CF cards and shoot Lossless 14-bit raw to the CF with a backup to the SD.

    I will buy more cards soon. However, I was thinking of buying slightly slower CF cards as I figured that the SD card was slowing everything down so why bother with the fastest CF.

    Based on this thread showing that the writing is sequential, I am back to the fastest of both. The Lexar 64GB 1066x is twice as much as the 800x though. Oh well...........

    If the D800 buffer was two or three times bigger, I would pay an extra $500 for that. Then I would use fairly slow cards and save a ton of money. I am not a sports action or news photographer, but there are times where I fill up the buffer even with the fastest cards. Kids will do that........
    Post edited by WestEndBoy on
  • @jshickele: Would you mind to measure the real speed of those cards you have? And say which cards exactly you have? Just out of curiosity,,,
  • WestEndBoyWestEndBoy Posts: 1,456Member
    @jshickele: Would you mind to measure the real speed of those cards you have? And say which cards exactly you have? Just out of curiosity,,,
    OK, I will do it this weekend.
  • KillerbobKillerbob Posts: 732Member
    If memory serves me right, none of the current Nikon DSLRs have interfaces much faster than 105MB/s. Hence spending the money on much faster cards is just not worth it.

    According to tests the SanDisk Extreme Pro are amongst the fastest SD cards, even faster than the 1000x Lexars (test in 12/2013). The same is true for the CF cards. However, this is tested using card readers and USB3 interfaces on computers, not in the write speed on your camera, and here I believe it is about equal speed, as long as it is the top-end cards. The interface in your D800 simply is the bottleneck, not the cards...
Sign In or Register to comment.