1. Reminder: Please use our affiliate links for holiday shopping!

ETTR still valuable on the E-M5

Discussion in 'Olympus Cameras' started by dhazeghi, Jul 20, 2012.

  1. dhazeghi

    dhazeghi Mu-43 Hall of Famer

    Aug 6, 2010
    San Jose, CA
    Dara
    Finally had some nice weather here in NY today - 68 degrees, heavy overcast, and occasional rain. Hey, it's better than the usual heat!

    Anyhow, I made the mistake of leaving the E-M5 on ISO 400. Truth be told, I forgot I had it on 400 from yesterday. That coupled with typically conservative ESP metering (I could probably have exposed by another stop without blowing highlights in many cases) resulted in a lot of surprisingly noisy files. The level of detail I got was rather disappointing compared to my shots a few days ago (all ISO 200, many slightly overexposed).

    So it looks like for clean results, I'm going to have to return to what I did with the E-PM1 - scrutinize the histogram closely, be liberal with positive exposure compensation, and when in doubt overexpose. Oh and stick to base ISO where possible!

    Cheers,

    DH

    i-3qcLg9P-L.

    i-CCMNCKq-L.
     
    • Like Like x 1
  2. I usually run both the E-M5 and E-PM1 at +0.7 EV using ESP metering.
     
    • Like Like x 2
  3. dcassat

    dcassat Mu-43 Veteran

    272
    Nov 16, 2011
    Cloverdale,CA
    I can directly relate as I have been shooting in overcast conditions. If I don't ETTR, wow it's easy to lose the exposures to noise. (not an E-M5 ... yet). But when I get it right and the clouds have some shape to them, I love the shots.

    It is my plan when I get my E-M5 to continue to shoot as I do now (ETTR, manual) as I believe my results are most consistent that way. I understand that the E-M5 does a better job but I have seen what happens on the beast when you pull from the shadows and I really don't like the mud that lies there.
     
    • Like Like x 2
  4. dhazeghi

    dhazeghi Mu-43 Hall of Famer

    Aug 6, 2010
    San Jose, CA
    Dara
    That's starting to sound like a good rule of thumb.

    Certainly. And pulling down the exposure in post, as long highlights weren't majorly blown, doesn't really have any negative ramifications.

    Agreed. I will say I was hoping for a little more latitude with respect to noise and exposure on the E-M5, but oh well... Shooting full-frame for 3 years made me pretty lazy in that regard. Still, with the 12-60, I can certainly keep the ISO lower next time around, just by opening up the aperture another stop.

    DH
     
  5. krugorg

    krugorg Mu-43 Hall of Famer

    Jul 18, 2011
    Minnesota USA
    So, are you guys settling in with the histogram, rather than using the red/blue blinkies? Maybe I need to go back to the histogram.

    I've done this a few times!

    I wish the auto-ISO had some customization options (min. shutter speed, IBIS on vs. off).

    Huh, I am going to have to start adjusting exposure shot to shot. I am using ESP and usually in high contrast/dayligh stuff, I keep it at -0.3, thinking that I am just going to pull shadows and be totally safe on highlights. I think I need to play with this a bit to dial in each shot a bit better. Not too troubled with noise, but always nice to start with less before PP.
     
  6. dhazeghi

    dhazeghi Mu-43 Hall of Famer

    Aug 6, 2010
    San Jose, CA
    Dara
    The problem with the blinkies is that they let you know if you've blown something, but not how your exposure is otherwise. If you're in a high contrast environment, that's probably enough, but in a low contrast one like I had for part of today, I could have exposed +/-1.3 EV easily without blowing/clipping anything.

    DH
     
    • Like Like x 1
  7. David A

    David A Mu-43 All-Pro

    Sep 30, 2011
    Brisbane, Australia
    The histogram tells me if something is blown but not what is blown. If I've got the sun in the frame, or a specular reflection from something like a car window, I'm not going to be able to avoid having that blow out no matter how much I pull the exposure to the left. It's handy to know that something's blown but you really need to know what is blown also if you're going to decide whether or not to try and do something about it. Blowing highlights that you want detail in is bad, but underexposing everything because you're trying to avoid blowing something that you can't avoid blowing is just as bad.

    The blinkies tell me what's blown and I can decide whether I want to do something about avoiding blowing that area or leave it. That's much more useful to me.

    And if there are no blinkies in evidence and I decide I want to ETTR, I can simply add positive exposure compensation until I start to get the blinkies. I've got the highlight warning set to 245 on both my bodies and I know from running a couple of informal tests that if I ETTR until I get the first blinkies showing in the viewfinder/on the screen, I have no problem with that highlight blowing. It either comes in without blowing or a minor touch on the highlight recovery rescues it with no adverse affects.

    If you want to run the sort of test I did, and I stole the idea from Pekka Pottka, it's easy to do. Set your highlight warning level to 245 (default on Olympus bodies is 255). Get a piece of white paper and place it in sunlight, and set the camera to spot meter mode. Start dialling in positive exposure compensation until you start to get blinkies on the paper. Take some shots with the exposure one or two exposure compensation steps (1/3 stop steps) before the blinkies show, when they show, and a step or two above that. Import into your processing software of choice and check which shot comes in without blown highlights on the paper, and how far further you can go before you can't recover the blown highlight on the paper.

    Once you know that, then you know what to do when exposing to the right, or whenever you get the blinkies showing. You'll know just how far above or below the point where the blinkies first show you can go and know that you'll be able to get a photo without blowing the highlight area.
     
    • Like Like x 3
  8. I find both of them to be distracting, so I just rely on what I am seeing on the screen and evaluate the exposure based on that. The reservation that I have over using the histogram or blinkies is that you can pay too much attention to them and underexpose by attempting to avoid blown highlights. Some highlights are okay to blow.
     
    • Like Like x 1
  9. jnewell

    jnewell Mu-43 All-Pro

    Jun 23, 2011
    Boston, MA
    Excellent suggestions. I'd been wondering about changing the 255 setting. This is how I'll figure out whether to do so and how much to change it if I do. Thanks! :thumbup:
     
  10. dcassat

    dcassat Mu-43 Veteran

    272
    Nov 16, 2011
    Cloverdale,CA
    When I use the blinkies predominately I tend to underexpose, so I use the histogram. If I know I've got a lot of top on it I check the blinkies to make sure it's not going to be my subject that's blowing out. I don't like clouds to be blown if they're interesting so that's always in the back of my mind.
     
    • Like Like x 1
  11. arentol

    arentol Mu-43 Veteran

    269
    Jun 29, 2012
    My experience is that the E-M5 will often get exposure pretty close. Closer than any camera I have used other than the Fuji X10. But it also almost always seems to have a lot of ton of headroom to pull back overexposure. So I usually try to over expose a tad as recovered highlights look a lot better than recovered shadows.
     
  12. dre_tech

    dre_tech Mu-43 Veteran

    314
    Jan 31, 2012
    Have you adjusted "Exposure Shift" from Custom menu K, or do you adjust to +0.7 manually when you shoot?

    I usually find my exposures to be pretty accurate in ESP, but I've found ETTR to work well in tricky situations.
     
  13. Liamness

    Liamness Mu-43 Veteran

    375
    Apr 20, 2011
    I just googled what ETTR was and I don't get why you'd use it. If you're worried about noise in the shadows, wouldn't you prefer a lower ISO with a normal exposure to a higher one with an exposure that will need to be 'pulled' later? In overcast conditions, I'm usually much more worried about my GF1 blowing out the sky than noise or lack of detail in the shadows.
     
  14. No, I leave the exp comp set to give me reminder on the screen to be careful with exposure. I find this to be safe enough on the E- M5 with it's greater headroom, but the E-PM1 has to watched more closely. I shoot this way in the knowledge that I will be making adjustments to the raw file in PP to make the end product look right.
     
  15. dhazeghi

    dhazeghi Mu-43 Hall of Famer

    Aug 6, 2010
    San Jose, CA
    Dara
    I guess there's no 100% solution, but I can't say I've encountered the situation you mention more than once or twice.

    Definitely a good test to try, thanks for the idea. Funny thing is that I used to use the blinkies religiously on the E-P2 because the histogram wasn't very good. But it does make sense to get a calibration. In theory, setting the JPEGs to something plain and flat should make the whole thing more accurate, but in practice I'd rather the image in the camera be close to what I expect to see on the screen, even at the cost of less accurate metering.

    No you're right - the first thing to do is to drop the ISO. But even after that, there's a substantial difference between exposing to the right and not.

    I won't say blown highlights are a non-issue on the E-M5, but on the whole it does a pretty good job, and with Lightroom you can actually recover a decent amount.

    DH
     
  16. There are a lot of examples of ETTR which pull the exposure of an over-exposed, higher ISO image to demonstrate the theory, but that isn't how ETTR should be used in practice. The general theory is to over-expose where you have headroom available that won't blow colour data that you want to keep and don't need to raise ISO to do it. Pulling an over-exposed image back to an average exposure level should result in slightly less noise than a correctly exposed image, and also less noise than an under-exposed image that has been pushed (assuming that all were shot at the same ISO). ETTR is not about choosing a higher ISO than you absolutely need to.
     
  17. David A

    David A Mu-43 All-Pro

    Sep 30, 2011
    Brisbane, Australia
    Though I have seen one recommendation (can't remember where) that suggested that the best way to expose to the right if you were going to do it, was to increase your ISO setting by the amount you were going to increase the exposure while maintaining the same shutter speed and aperture on the basis that this gave better results.

    Example:

    You're shooting at ISO 200, the recommended exposure is 1/200 @ F/4 and you have room to overexpose by 1 stop, either by going to 1/100 or to F/2.8

    What this view was saying was that you should go to ISO 400 instead and keep the 1/200 @ F/4 rather than staying at ISO 200 and reducing your shutter speed or opening the aperture.

    I haven't tried it and I don't know whether it works any better than staying at the same ISO and changing aperture or shutter speed.

    In my experience I find that I usually tend not to have the leeway to expose to the right, the drawback of living in a location with way too much light if you're outdoors during the day. And that's why I've discovered that simply relying on the histogram doesn't work particularly well for me. Before I got the E-M5, my E-P3 would rarely indicate that I didn't have clipping somewhere if I was shooting outside. If I relied on the histogram and reduced the exposure to reduce or eliminate the clipping, I ended up underexposing the things I wanted to capture, basically because most of what was clipping was the sky, and often large parts of it. Then I'd end up having to correct the underexposure in processing and got the blown skies back so I gained nothing except noise in shadow areas that wouldn't have been as noisy had I simply ignored the clipping in the sky.

    That's why I much prefer to use the highlights and shadow indication and decide what I'm going to do on the basis of just what is clipping rather than relying on the histogram.
     
  18. My assumption was that ideally you wouldn't adjust ISO upwards since it is basically adding noise for the sake of reducing noise. However, sensor characteristics may be that going from ISO 200 to ISO 400 results in very little increase in noise, whereas at some point the sensor may fall off a cliff, maybe going from ISO 800 to ISO 1600 for example. I guess there must be some benefit to this since both Fuji and Canon use ISO manipulation to provide enhanced DR modes. I don't know if you can simulate this manually.