Nature Photographers Network
This is the preserved archive of the original Nature Photographers Network. The community continues at naturephotographers.network.

Digital and exposure!

Discussion

9 replies
Posted by Prashant Khapane on May 2, 2006 at 10:50 AM
How do you set your exposure on digital cameras? I recently made a trip to Harz with LF and digital and relied on histogram. I'm reading some books and understand that the histogram shown is for the embedded jpg file and not the raw file. How do you use hitogram to set the exposure then?
Please note - I understand the basics of exposure and own handheld meters.
Posted by Richard Wong on May 2, 2006 at 10:59 AM
Spot metering. If you know how to use a lightmeter, then it's the same concept just TTL. Histogram doesn't set exposure as it's not a live reading, it only tells you after the fact. All files have it, RAW, JPEG. Seems like an odd question considering your images are excellent.

Posted by Prashant Khapane on May 2, 2006 at 11:03 AM
Richard,
am new to digital, (use film too). From what I have read it tells me that the exposure (handheld or in camera) should be handled differently for film and digital. If I'm not wrong one should try to use that headroom on the highlight side.
Just thought if asking. My images came out fine this weekend with the incident meter I used to get my base EV (PilotII).
Posted by Juza on May 2, 2006 at 11:31 AM
Hi Prashant,
I always use "evaluative" metering, but what really "determines" the exposure is the histogram: I take a test shot, I give a look to the histogram, and if necessary I repeat the photo with the right settings.
The histogram is not 100% precise, but it is accurate enough to get the right exposure; you can also use it to overexpose slightly the image (to optimize the dynamic range...more info: Dynamic Range and exposure).
MR has a good article about exposing to the right, too: http://www.luminous-landscape.com/tutorials/expose-right.shtml

:wavey:
Posted by Prashant Khapane on May 2, 2006 at 1:11 PM
Juza,
this is what I was thinking about. One has to expose a bit to the right. I find the histogram quite ok tool for checking the exposures.
Posted by Josh Anderson on May 2, 2006 at 1:41 PM
Wow. This is news to me. How often can you overexpose an image and NOT get areas that are burnt out? I guess it mostly depends on what kind of light you shoot in. Go for the sweet light, right? :biggrin:
Posted by Juza on May 2, 2006 at 2:06 PM
Hi Josh,
if the subject if in shade, soft or overcast light, it is possibible to overexpose even by 1 or 2 stops without burning the highlights. That said, I don't use always this technique, for two reasons:

1) even if the histogram looks OK, it is possible that a single color channel is "burnt"...it happens often with bright yellow or red subjects. To be 100% sure that the image is properly exposed you would actually need a RGB histogram as the one of the 1 series cameras...with the less precise luminance histogram of the lower-end bodies, it is better to be a bit more conservative with overexposure to avoid burning a channel.

2) when I photograph active subjects (e.g. birds) I prefer to not overexpose (even if it would be possible without burning highlight), because if the subject moves against a dark BG the "slight overexposure" become "strong overexposure" and you lose the highlights.

:wavey:
Posted by Josh Anderson on May 2, 2006 at 2:21 PM
Thanks for the clarification Juza! :toast:
Posted by Jack Graham on May 3, 2006 at 9:49 AM
I am finding that +1 in shade works quite well most of the time. Juza is right use thr RGB histogram.
JG
Posted by Guy Tal on May 3, 2006 at 10:17 AM
The luminance histogram should be the same (or very very close) between the JPG and raw capture. It's the individual color channels that may be different depending on things like white balance or other processing that impacts color.

I would say the luminance histogram is your best bet in the vast majority of cases. Seeing the individual color channels is certainly helpful, as Juza suggests, but in many cases you should be able to anticipate which color channels are at risk just by studying the scene.

Guy