[Leica] Question about M8 exposure

[henningw at archiphoto.com (Henning Wulff)]

Hi Hoppy,

What I'm really saying is that with the M8, shooting in Auto mode 
(manual mode implies measuring relevant point to get your exposure 
correct, so exposure compensation is done in the brain and not in the 
camera) means that I get the most useable shots under average 
conditions by setting the compensation to -1/3 or -2/3.

This does not apply to snow scenes, 3/4 portraits of people against 
white backgrounds, coal mines or other distinctly 'non-average' 
settings.

The M8 meter is a center weighted meter with a wider angle of 
acceptance than that of the M6 or M7, but it does not additional 
computation such as most DSLR meters now do. That means that on full 
auto, you'll get a wider range of exposures that are 'off' than with 
the matrix metering boxes, but it also means that you can learn what 
it's doing and act accordingly instead of wondering what the camera 
computer wants to throw at you this time.

The M8 meter, left on Auto and shot all day with average scenes and 
set at -2/3 compensation, will often underexpose but there is less 
chance it will overexpose. As I mentioned, overexposure is mostly 
_BAAAD_ for digital files, and underexposure is generally something 
you can deal with readily, and will not be visible in most prints. 
Clipped highlights will often be visible and uncorrectable.

In the end, perfect exposure is perfect exposure. Set the camera to 
manual, meter off various elements in the scene and integrate the 
values into a correct exposure. Take a shot, check the histogram and 
then take the 'money' shot.

Or, set the camera on Auto with compensation that avoids highlight 
blowouts, and take a dozen pictures in the meanwhile.




At 1:09 PM +1000 1/19/09, Geoff Hopkinson wrote:
>Henning as always your input reflects your extensive experience.
>What also happens in these threads is that different issues become
>intertwined.
>Please keep in mind this is strictly M8 Raw exposure I am talking about. Not
>film and not any other dSLR. I have zero experience in trying to recover
>highlights or shooting with any other dSLR. Obviously film is a whole
>different plate of sushi too.
>This line of yours is very important to what I have been trying to say too.
>"If you see no clipping, and have information in the furthest right short of
>clipping, you have the best chance of having as much accurate information as
>possible."
>When you set -.33 or -.66 you may not be getting as much highlight
>information as possible. And in turn you are potentially capturing much less
>tonal information, which is redistributed by the gamma curve applied.
>Absolutely agree that may not matter for some situations or even many
>situations or may be preferable to risking clipping. I try to manage
>clipping by the way I meter rather than automatically shifting everything
>left.
>I never assume that the DNG will always be fine when there is clipping
>evident in the LCD preview, but it is not gospel either.
>I understand what you are saying about recovering highlight information if
>one or more channels are clipped too.
>I don't know if you looked at my posted examples of the 1,200 year old tree
>in the snow at Red Rock Canyon. That exposure was made with +.66 exposure
>initially. The LCD showed significant clipping, then as the camera processed
>the preview you see the white balance and clipping indication both shift, on
>the LCD in the first couple of seconds (with later firmware). The white
>balance shift alone will tell you that you are seeing a preview of jpg
>settings.
>Most importantly the histogram of the DNG in ACR shows zero clipping. In
>fact there is room in the last 10% to the right.
>Contrast, clarity etc were adjusted afterwards. The final exposure is in
>fact +.61 on what the meter wanted to set (and that based on metering the
>tree, not the snow). There is a small amount of recovery but from zero
>clipping in any channel which conforms to what you have explained too.
>
>I'm not trying to be argumentative here and I greatly respect your
>experience. I do think that it is a very on-topic discussion for us with so
>many M8 shooters that are active here. For me it also demonstrates the
>capability of the M8. For my style of photography I have not at all had the
>experience of great recovery from underexposed shadows. A number of people
>have said that they do. For me any underexposure of the shadows makes for
>poorer quality there. More noise for example.
>
>Cheers
>Geoff
>http://www.pbase.com/hoppyman/e
>http://gallery.leica-users.org/v/gh/
>Pick up your camera and make the best photo you can.
>
>-----Original Message-----
>Subject: RE: [Leica] Question about M8 exposure
>
>A suggestion has been made that the clipping shown on the histogram for the
>jpeg might not be accurate, so just give a bit more exposure and check the
>clipping in the RAW later.
>
>As long as _NO_ channel is clipped, things are fine. The problems arise if
>one or two channels are clipped, and the other(s) aren't.
>Then when you try to bring some of that highlight info back, you'll get the
>info from that one or two channels, but the other will have nothing. You
>then get unbalanced information which is erroneous, and distorted.
>
>It's true that the jpegs aren't accurate, but you can't therefore assume
>that the DNG's will be fine. That doesn't follow.
>
>If you see no clipping, and have information in the furthest right short of
>clipping, you have the best chance of having as much accurate information as
>possible.
>
>My Canons certainly do not give me that with a + exposure compensation on
>average subjects on matrix metering, and with my metering technique (which
>works perfectly on my M6's with the meter set to 400 for Tri-X developed in
>D-76 1:1 etc.) I generally leave my
>M6 set to -2/3.
>
>When I bring the DNG's into LR, I try to use the highlight recovery as
>little as possible, and the exposure slider doesn't get used much unless I'm
>shooting in very dark circumstances (I've explained my technique there
>previously). Mostly I use the curves tool to do contrast adjustments,
>although the LR curves tool tries to outsmart me and not allow me the fine
>control I'm used to in PS.
>
>If I do have to use the exposure slider, I'm much happier to increase the
>exposure than decrease it, as I know that while I might be increasing some
>noise in the shadow areas which is quite easily dealt with, I won't be
>introducing uncorrectable distortion in the highlights.
>
>--
>
>     *            Henning J. Wulff
>    /|\      Wulff Photography & Design
>   /###\   mailto:henningw@archiphoto.com
>   |[ ]|     http://www.archiphoto.com
>
>_______________________________________________
>Leica Users Group.
>See http://leica-users.org/mailman/listinfo/lug for more information
>
>
>
>_______________________________________________
>Leica Users Group.
>See http://leica-users.org/mailman/listinfo/lug for more information

-- 

    *            Henning J. Wulff
   /|\      Wulff Photography & Design
  /###\   mailto:henningw@archiphoto.com
  |[ ]|     http://www.archiphoto.com