(Leaked?) 5D mkII Test Shots at ISO 25,600

Discussion in 'Canon EOS' started by dcheung, Sep 18, 2008.

  1. Found this on flickr:

    http://www.flickr.com/photos/erikfive/2868119820/

    Judge for yourself.

    The EXIF says Canon EOS 5D Mark II captured today at 1:33pm

    I hope this is real cuz I'm impressed.
     
  2. Just found another site that has 5D sample shots at high ISO:

    http://www.akihabaranews.com/en/news-16717-EOS+Canon+5D+Mark+II+HANDS+ON+with+Video+&+Pictures+Sample%21%21%21.html

    I found this link on fredmiranda forums.... just downloading the 300mb file so not sure if it's real or not but the website looks legit.
     
  3. It IS cleaner than a 20D at 3200, and (being an old GAF 500 film user) I'm even happy with that.

    If real estate taxes weren't due in a couple of weeks, I'd buy this sucker right now.

    Despite my own common warning (the Microsoft Law): "Never buy a point.zero release of anything."
     
  4. "It IS cleaner than a 20D at 3200"

    haha, I'm not sure about that but I wish it's true!
    It seems to apply a stronger noise reduction (esp chorma noise) at ISO25600 and that's why it looks cleaner than
    3200 on 20D but if you look at the detail, there is also more loss of detail compared to 3200 on 20D.
     
  5. Very impressive. I hope it's real. I am seeing some significant horizontal banding in the darker out-of-focus areas, but at 25,600 I could live with that.
     
  6. Looks very impressive!
     
  7. >>>>>>The EXIF says Canon EOS 5D Mark II captured today at 1:33pm

    That's strange Weiyang
    I tried using Opanda Exif Reader 2.3and it shows no EXIF data available.

    How did you check the Exif data?
     
  8. if that is real, it's really impressive... banding noise... but common... 25,600!

    If you want 1/2000 indoors at f2.8 using natural light, I suppose you could do it.
     
  9. Did you guys check out the rest of the shots? Looks better at 6400 than my 40D does at 1600. Chrominance noise is usually fairly easy to get rid of in PP, and the luminance noise is really quite manageable I think...

    In the D700 vs 5Dii ISO 6400 comparisons, I thought the 5D was a little better, which is good considering it's a 13mp to 21mp comparison. If this is the kind of quality at high ISO's we can expect, then i think we're in for a treat with this thing.
     
  10. John >> I simply opened up the properties in windows and went into "summary", then "advanced".
     
  11. I believe the samples are real. Looks like Fotovideo in Oslo. Canon is going to sell their products( and have somebody from their own staff)
    in the shop. They started this week, and have 5DII on display. I´m going there tomorrow, and hopefully they´ll let me take a few shots onto
    my own card.

    The samples look impressive, even though it´s a very well lit shop. I´m more interested in pictures taken in darker conditions.
     
  12. I am primarily a studio photographer and I am let down that the fastest sync speed is only 1/200. My Canon 30D
    can do 1/250, is there some mistake did I read this wrong? I use the higher sync speeds often when I have too
    much ambient light in the room and I want to make my black backgrounds really dark.

    Opps, did I answer my own question? Could I compensate for this by going to ISO 50?
     
  13. The 5D has a full-frame sensor, and therefore a somewhat larger shutter and mirror than the 30D does. If the 5D shutter blades merely move as fast as the 30D's, the flash sync on the 5D will be 1.6 times slower than the 30D, or about 1/150. It looks like they move just a little faster than the 30D blades, but not quite fast enough to reach 1/250. I suppose this is also the reason for the modest 3.9 fps frame rate: The larger mirror is hard to move as fast as the 1.6 mirror can be moved.

    Incidentally, being in a math geeky mode, I calculated that the shutter has to be moving about 16 miles per hour to achieve that sync speed (assuming it travels the long way across the frame). Considering the space it has to accelerate and decelerate in, that ain't bad.
     
  14. It travels across the short side of the frame. That is why it is called an electronically-controlled vertical-travel focal plane shutter.
     
  15. Well then, I get 10.7 mph.
     
  16. Can I compensate for the slower shutter speed by lowering the ISO from 100 to 50? What else will this afect in my images. Will it also give me finer (less grain) photos as it would with film? Would it be anything like the discontinued Kodak Ektar 25 or Kadak Elite 50? Or is it essentially the same as ISO 100 grain but requireing more light?
     
  17. The sync speed was a dissapointment as there does not seem a reason IMO not to improve, the 5d does not sync well with most strobes at 1/200 and needs to be at max 1/125 in some cases, looks as tho will have to get nd filters for the strobes. I mainly shoot 1/30~1/90 studio. Using 50iso will lose some DR but some portraiture should be ok with lower contrast. still if 6400 is a big improvement I suspect 50iso likewise :)
     
  18. jhp

    jhp

    A little off ISO topic but about the only thing at first glance I'm just completely turned off from is the lack ofoss focus points and the tight configuration of the ones that are there. I see a whole lot of recomposition necessary for my liking. They improved the AF system over the old 5D yes, but limited us to one center point with cross focus.

    Not bad for those who shoot only on center point....
     
  19. ISO 25,600... It's only four-five years ago that I shot with ISO 3200 film for the first time and found that totally thrilling. A similar leap from here will take us to ISO 204,800. (Right?) My only concern here is that these high ISO numbers will make it less likely that there will be new versions of some of the dated amateur-priced primes. Pros will always have the cool f/1.2L etc. series for maxmum DOF control and speed. But what about the rest of us? Just a thought.
     
  20. Anders

    good point. actually, i think the improved sensors will kick-start the production of better short primes. i
    believe we're already seeing that with the sigmas and new zeiss primes. canon has the new 24 1.4.

    pro zooms are nice, but no match for the best primes. canon knows that -- they have to be considering the
    implications of having killer bodies but primes that don't meet the sensor's potential.

    i'll bet we're in for a good year for short primes, including conventionally fast ones like f1.8 to f2.8, etc,
    for both pro and amateur use

    ps -- iso 102,400 can't be that far off
     
  21. Anders Carlsson... Are you the one I know or do you just happen to also have the unusual spelling of Carlsson?
     
  22. on the issue of the sync speed, of all the full frame digitals, the fastest X-Sync speed is 1/250th and that belongs to the 1Ds-series. Even the 1D MkIII only has 1/250th. The much bigger mirror in the full frame cameras means more travel distance and thus more travel time (as an above user have already stated). The 5D's 1/200th isn't bad at all considering it's only 20% slower than the 1-series.

    If memory serves me correctly, for APS-C cameras, 20D-50D are all 1/250th; D30,D60 and 10D are 1/200th, and the only canon digital camera with a higher X-Sync belongs to the original 1D with a smaller than full frame sensor and 1/500th X-Sync.
     
  23. Alan you might be a maths Geek, but you clearly ain't a camera geek. Shutters have been doing this since the Nikon FM-2 and my years ago I had an old EOS 620 which also had 250th X-sync John Vanacore download the image and use something to see the EXIF there ... I get this:
    00Qtsn-71899584.jpg
     
  24. Calculation of shutter curtain speed: at 1/250th (= 4ms) X sync, say curtains fully open for 1.6ms to accommodate a full power flash. During curtain travel, on average half the frame is exposed, but there are two curtains to move - therefore each curtain takes 2.4ms to cross the frame of 24mm (now the choice of fully open duration is revealed!). That implies an average curtain travel speed of 10m/s, or about 22.5mph. To improve the X sync to 3ms (1/333rd) would require the curtains to make the journey in 1.4ms each (nearly 40mph!), or to sacrifice the fully open duration: you can also get there by having a sensor that is just 14mm wide (cf. APS of about 15mm). If the curtains can't be speeded at all, the limitation on sync speed should be clear. IIRC, Minolta was the first manufacturer to offer 1/350th with a film camera and a focal plane shutter (I suspect with a lower curtain open duration). Of course, if you use a diaphragm shutter (e.g. by adapting a LF lens which has an inbuilt shutter) you could achieve higher X sync.

    Perhaps an interesting question is that, given the 5D Mk II has an electronic shutter for the purposes of video and live view, can this be taken advantage of for X sync in the way that happens with the CCD in the Canon 1D? Maybe such a feature will appear on a future 1 series again.
     
  25. But is the problem the speed of how fast the shutter needs to move? Isn't during normal operation, you could set a shutter speed of 1/4000s ?

    So why is achieving 1/250s sync so hard? I'm confused.
     
  26. There are two shutters, generally called first curtain and second curtain. The first curtain opens to expose the sensor, the second curtain closes to cover it.
    The interesting part is, high shutter speeds can be achieved by overlapping these two operations. That is, the second curtain begins closing before the first curtain has finished opening. For fast exposures like 1/4000, the second curtain is only just right behind the first. The gap between them is very small. And every portion of the sensor recieves a 1/4000 second exposure, even though the entire exposure as a whole took much longer than 1/4000 to occur.
    The problem with a flash is, the whole frame has to be exposed when the flash fires -- you can't have either curtain covering the sensor. So the flash sync shutter speed is much slower than the fastest shutter speed achievable.
    When I figured my 10.7 mph, I was leaving no time for the flash to fire. Mark U suggests 1.6ms would be required for a full power flash -- that would imply that 1/625 is the fastest flash sync shutter speed that can be achieved at all, no matter how fast the shutters move. As you approach 1/625 the required speed of the shutter blades becomes very very fast. Passing 1/300 enters a world of diminishing returns -- doubling the shutter travel speed only slightly improves the flash sync speed.
     
  27. http://www.photozone.de/standard-flash-sync

    http://www.photozone.de/hi-speed-flash-sync (also shows how high shutter speeds work)

    http://www.eosdoc.com/manuals?q=Flash+Curves (how long do curtains need to be open for X sync)
     
  28. I love it when these things go off topic sometimes!

    Anyway, my 1D syncs at 1/500, how? Oh, I understand my chip is smaller than the full sized one being talked about but it is still pretty fast. It does it by useing a CCD rather than the "better" CMOS, the shutter is open longer than 1/500 but the sensor is only sensitive for that, kinda cool huh, so don't see why they couldn't sync CCD's much faster, well they probably do in some other imaging areas, but progress and the perceived advantages of CMOS chips over CCD's have left you guys all struggling. Silly thing is although I have a good flash set up I hardly use it!
     

Share This Page

1111