Pages: [1]   Go Down

Author Topic: ETTR feature on Olympus E-PL1!  (Read 16300 times)

feppe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2906
  • Oh this shows up in here!
    • Harri Jahkola Photography
ETTR feature on Olympus E-PL1!
« on: September 03, 2010, 12:32:32 pm »

I was doing some shots with my Olympus E-PL1 Micro Four Thirds camera for another thread, and noticed it has three different spot metering modes: normal, hi and low. They are poorly documented, so I tried them out.

The hi mode is essentially an expose-to-the-right mode! When you're in that mode and point at the brightest spot in the scene, the camera will expose that spot for higlights, instead of middle grey. I did numerous test shots around my apartment, and almost all were very nicely ETTRd with no channel blowing out. Histograms verified in Lightroom. There were two or three which had blown highlights and one which was underexposed, but I suspect this has something to do with the limited accuracy of a 2% spot meter.

It's not fully automatic, of course. You have to determine which is the brightest spot in the scene and ETTR that. Everything brighter will be blown out. Technically I can see a fully automatic ETTR mode to be possible, but desirability is another matter. How to deal with specular highlights and how much image area is allowed to blow out should be left to the operator - those who care about ETTR are probably proficient enough to make that call themselves.

It is unclear how the camera determines the exposure in this mode. As it's doing it on the fly and you can see the results live on the screen, it must be using RAW data instead of in-camera JPEG - whether it uses all three color channels is another matter. In any case, it's a very exciting feature, something which many, including me, have been asking for on digital cameras. I'll do some further tests on its utility and reliability in real-life shooting outside this weekend.

There is also low mode which exposes the spot for shadows, utility of which I have not yet fully figured out. I guess you could drown out anything darker than the spot you're measuring with it, which could be useful for those who don't post-process their photos.

Attached demonstration with three frames, red spots mark where I put the highlight spot.

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20614
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: ETTR feature on Olympus E-PL1!
« Reply #1 on: September 03, 2010, 02:50:47 pm »

The hi mode is essentially an expose-to-the-right mode! When you're in that mode and point at the brightest spot in the scene, the camera will expose that spot for higlights, instead of middle grey.

Cool, sounds like a real useful feature. Kudo’s to Olympus for that.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Frodo

  • Full Member
  • ***
  • Offline Offline
  • Posts: 152
Re: ETTR feature on Olympus E-PL1!
« Reply #2 on: September 03, 2010, 03:29:26 pm »

Yes, very useful.  Better than being stuck with spot-metering middle gray zones.  Not surprising that it came from Olympus, given the use of multi-spot metering on their OM-4.

A work around for other cameras would be to set the exposure compensation to over-expose the spot-metered highlights, but this would involve some guesswork.
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: ETTR feature on Olympus E-PL1!
« Reply #3 on: September 03, 2010, 03:39:35 pm »

A work around for other cameras would be to set the exposure compensation to over-expose the spot-metered highlights, but this would involve some guesswork.

Not really guesswork, it's just that the spot meter may not be narrow enough.
I can use spotmetering on highlights with my 1Ds3 and adjust +3EV for ETTR without clipping the Raw data.
Of course one still need to use common sense, e.g. in tungsten lighting conditions, or with small specuar highlights.
Exposure meters are not an excuse to shut the brains down.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

MrSmith

  • Guest
Re: ETTR feature on Olympus E-PL1!
« Reply #4 on: September 03, 2010, 04:01:00 pm »

that's a great feature. Olympus seem to add lots of silly modes that people are unlikely to use yet something like this isn't promoted.
(your windows need a clean BTW)
Logged

feppe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2906
  • Oh this shows up in here!
    • Harri Jahkola Photography
Re: ETTR feature on Olympus E-PL1!
« Reply #5 on: September 03, 2010, 06:45:05 pm »

Exposure meters are not an excuse to shut the brains down.

Good advice!

that's a great feature. Olympus seem to add lots of silly modes that people are unlikely to use yet something like this isn't promoted.
(your windows need a clean BTW)

Indeed. I did a brief google search and didn't find anyone discussing this particular feature. I have a feeling not too many outside of hardcore photography geeks know about and especially appreciate the value of ETTR. Too bad the feature is poorly documented and I figured out its true nature by trial and error. Now that adjustable auto-ISO seems to be a standard feature, I sure wish this would become the next widely adopted feature in digital cameras. Much more valuable than all those "creative" modes or "art" filters.
(what's the point of cleaner windows when the view looks like that :P )

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: ETTR feature on Olympus E-PL1!
« Reply #6 on: September 04, 2010, 04:05:07 pm »

I was doing some shots with my Olympus E-PL1 Micro Four Thirds camera for another thread, and noticed it has three different spot metering modes: normal, hi and low. They are poorly documented, so I tried them out.

The hi mode is essentially an expose-to-the-right mode!
[Sorry if my reply appears twice; my first effort seems to have vanished.]
Yes, that is a nice if poorly advertised and explained feature; Olympus has had it in its 4/3 and m4/3 cameras for some time, but under several different names.

My aging E-510 has it under "Gradation>High Key". Since blown highlights are my most common frustration with digital cameras, this is a nice metering option to have.
« Last Edit: September 05, 2010, 11:03:34 pm by BJL »
Logged

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re: ETTR feature on Olympus E-PL1!
« Reply #7 on: September 05, 2010, 05:56:01 pm »

Nice to hear that. It seems the E-P1 also has this feature, I'll try it next weekend on my camera. To find out how accurate this HI-spot meter is, you could HI-spot meter on a white paper and analyse the resulting RAW channels with DCRAW. BTW: your red spots could have been a bit smaller or transparent, so that we could see the highlights behind ;D

It would be even nicer that the camera could analyse the entire scene frame without having to manually spot meter the highlights. I don't think it would be too difficult to design an algorithm to distinguish specular reflections or small size very bright light sources in order to discard them from the ETTR analyse.

My little one:



Regards
« Last Edit: September 05, 2010, 06:14:08 pm by Guillermo Luijk »
Logged

feppe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2906
  • Oh this shows up in here!
    • Harri Jahkola Photography
Re: ETTR feature on Olympus E-PL1!
« Reply #8 on: September 05, 2010, 06:36:59 pm »

It would be even nicer that the camera could analyse the entire scene frame without having to manually spot meter the highlights. I don't think it would be too difficult to design an algorithm to distinguish specular reflections or small size very bright light sources in order to discard them from the ETTR analyse.

That's what I was skeptical about in my post. I'm sure it's possible to write such an algorithm, though - can't imagine it being any more difficult than the thousands of scenes Canon claims to have written into the exposure algorithms they use for center weighted averaging. When you have a desk lamp taking a good portion of the frame, or you just want to blow out the sky, you could still use manual spot metering.

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re: ETTR feature on Olympus E-PL1!
« Reply #9 on: September 05, 2010, 07:00:56 pm »

That's what I was skeptical about in my post. I'm sure it's possible to write such an algorithm, though

I think such an algorithm should be based on detecting large differences in brightness in just a few pixels.

A bright sky would typically be the brightest area of a scene, but its brightness would be graduated (no big exposure gaps, but a smooth gradation instead). On the other side a light source (sun, reflection, lamp) would create a big exposure gap between its brightness and the scene's (or its surrounding area) average brightness, forming an empty hole in the histogram with nothing in it.

If the Live View manages to collect different images from the scene at very different shutter speeds in order to be able to calculate these brightness gaps, a smart classification algorithm for bright areas should not be too complicated.

douglasf13

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 547
Re: ETTR feature on Olympus E-PL1!
« Reply #10 on: September 06, 2010, 02:33:10 pm »

Remember that exposing too far to the right can often lead to color issues, especially in the mid-tones, so you may have to decide on whether sacrificing color for less noise works for you.
Logged

feppe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2906
  • Oh this shows up in here!
    • Harri Jahkola Photography
Re: ETTR feature on Olympus E-PL1!
« Reply #11 on: September 06, 2010, 03:02:35 pm »

Remember that exposing too far to the right can often lead to color issues, especially in the mid-tones, so you may have to decide on whether sacrificing color for less noise works for you.

Are you referring to blowing out a color (which I referred to in my OP), or that there could be issues even if none of the channels are blown?

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re: ETTR feature on Olympus E-PL1!
« Reply #12 on: September 06, 2010, 04:08:31 pm »

Remember that exposing too far to the right can often lead to color issues, especially in the mid-tones, so you may have to decide on whether sacrificing color for less noise works for you.
I never had this problem with my Canon cameras. In fact I would say the mid-tones are best in all aspects the more you expose the RAW file. It is true that some pixels can start to behave non-linear if they are surrounded by saturated pixels, but this is not the case with proper ETTR.

This is a genuine response of a Canon 350D, highly linear:


Once the R and B channels get glipped, B seems to start losing some linearity (remember each B photodiode is surrounded by 8 non-B photodiodes).

douglasf13

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 547
Re: ETTR feature on Olympus E-PL1!
« Reply #13 on: September 10, 2010, 04:58:28 pm »

  Interesting.  I've seen examples to show the contrary.  To quote Iliah Borg, "I do not (ETTR) because the cameras are not optimized for ETTR. If one ever tried to profile his cameras he would know that cameras are optimized to produce good colour when the exposure is set by the exposure meter."

  I just happen to run across this link earlier:  I do not because the cameras are not optimized for ETTR. If one ever tried to profile his cameras he would know that cameras are optimized to produce good colour when the exposure is set by the exposure meter.
Logged

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re: ETTR feature on Olympus E-PL1!
« Reply #14 on: September 10, 2010, 05:27:41 pm »

I read the same from Illiah's daughter time ago in Dpreview's forums. I think they use Nikon, perhaps Nikons are less linear than Canons (maybe Bill Janes can give some info about this), but I am a bit skeptical.
A friend of mine, Manuel Llorens, analysed his Olympus camera response, and it seems to be quite less linear than Canons' once getting close to saturation in some channel however: Fixing sensor non linearity:

Olym­pus E-​​510 sensor response:


In this kind of sensor it seems true that the most linear response is achieved in the middle RAW exposure range.

__________

Going back to the thread's question, I did a quick test with my E-P1 and the spot meter HI mode to find out how much headroom the camera allows before clipping any RAW channel. Left are RAW histograms (first I calculated the precise camera's saturation point to set a correct 0EV reference), and right camera's JPEG from 3 different targets: white paper, blue folder and red plastic bag.



Only the red plastic bag got an optimum exposure from the RAW point of view. In the white paper and the blue folder, we could have exposed 1 entire extra stop. It seems this mode is not optimised with the resulting RAW in mind.

Regards
« Last Edit: September 10, 2010, 06:51:10 pm by Guillermo Luijk »
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Re: ETTR feature on Olympus E-PL1!
« Reply #15 on: September 11, 2010, 11:16:15 am »

I read the same from Illiah's daughter time ago in Dpreview's forums. I think they use Nikon, perhaps Nikons are less linear than Canons (maybe Bill Janes can give some info about this), but I am a bit skeptical.
Guillermo,
The Borgs are quite knowledgeable, and I hesitate to disagree with them, but I have found that the channels in my Nikon D3 are quite linear up to very near clipping as shown below. These values are derived from exposing a Stouffer wedge and only the brightest steps are analyzed here. I do not hesitate to use ETTR with this camera.

Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: ETTR feature on Olympus E-PL1!
« Reply #16 on: September 11, 2010, 11:06:10 pm »

If there's any doubt about possible degradation of upper tones due to an ETTR exposure, simply bracket every exposure and choose the best.

If the subject is moving, that might not always work. Even the expression on someone's face can change in a fraction of a second.

On the other hand, if that does occur, you have a choice between a slightly underexposed shot with the best facial expression, and a fully exposed ETTR shot with a slightly less interesting facial expression, but cleaner midtones and shadows.

When bracketing exposures, it's sensible to get the first exposure as the 'normal', metered exposure.
Logged
Pages: [1]   Go Up