Stacking for increasing exposure?

A forum to ask questions, post setups, and generally discuss anything having to do with photomacrography and photomicroscopy.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

specious_reasons
Posts: 107
Joined: Tue Feb 24, 2009 12:46 pm
Location: Woodridge, IL

Stacking for increasing exposure?

Post by specious_reasons »

Is that even the correct term? This came up in a RL discussion about darkfield photography. I was complaining that darkfield exposures are very long, and while I'm generally satisfied with the results from long exposure times, sometimes it produces completely blown highlights. Here's an example of a hydra causing me trouble:
Image
Center-weighted metering, ISO-800, 0.8 sec exposure.

It was suggested to me to do something like astronomy photography: Take a series of pictures with much shorter exposure times and blend them together.

As an experiment, I tried this, but results have been - dramatically awful. There are several sources for the problem. The first is that I have no clue as to what I'm doing. The second is that I suspect I'm using the wrong tools (in this case CombineZP).

Is this a worthwhile line of pursuit?

Charles Krebs
Posts: 5865
Joined: Tue Aug 01, 2006 8:02 pm
Location: Issaquah, WA USA
Contact:

Post by Charles Krebs »

I'll defer to others here with more experience in astrophotography, but my impression is that they use that technique to reduce the digital noise and obtain "cleaner" subject image data from faint signals.

I've never seen it utilized for darkfield photomicrography. The blown highlights are not a problem caused by long exposure times per se... it's simply overexposure. Darkfield can be a tricky subject for a camera meter. That's one thing that makes digital so useful. If you are using a digital camera you might find you are get the best results with the camera set to manual exposure. If you check the histogram after taking an exposure you can quickly make any needed exposure adjustment.

One "exposure" situation where multiple images combined with software might be helpful is to achieve a higher dynamic range. It's not uncommon for certain darkfield subjects to be extremely contrasty. So that when the exposure is set to record the highlights properly, the darker tones disappear into black. If you have a subject that does not move, you can take a series of exposures and combine them into a single image that shows detail in a wide range of tonal values. TuFuse is a program that is quite good at doing this.

specious_reasons
Posts: 107
Joined: Tue Feb 24, 2009 12:46 pm
Location: Woodridge, IL

Post by specious_reasons »

Charles Krebs wrote: I've never seen it utilized for darkfield photomicrography. The blown highlights are not a problem caused by long exposure times per se... it's simply overexposure. Darkfield can be a tricky subject for a camera meter. That's one thing that makes digital so useful. If you are using a digital camera you might find you are get the best results with the camera set to manual exposure. If you check the histogram after taking an exposure you can quickly make any needed exposure adjustment.
(emphasis mine)

That might be the fundamental source of the problem. I'm generally using Aperture priority mode on the camera. I suspect that it sometimes just doesn't choose exposure correctly.

Well, tonight I think I will experiment a bit with the idea.

mgoodm3
Posts: 273
Joined: Mon Sep 08, 2008 8:50 am
Location: Southern OR

Post by mgoodm3 »

Not sure what the pic of the hydra should really look like, but it is overexposed. The picture has a wide range from light to dark - HDR imaging may be useful. HDR takes several pics at various exposures to get info in the shadows and the brightest parts and combine the info into a 32-bit image. You can them manipulate the image in 32-bit and them convert back to 8 or 16 for the final product.

You can properly expose several images and average them to reduce image noise. Kinda like oversampling.

specious_reasons
Posts: 107
Joined: Tue Feb 24, 2009 12:46 pm
Location: Woodridge, IL

Post by specious_reasons »

For my own amusement last night, I tried a experiment. I took one picture with a manual exposure setting and then I took several pictures at a much faster exposure time. Then I used GIMP to make each exposure as an additive layer.

This is a sample of the results:

Image

The combined image is not bad, but I think it's more worthwhile just to get proper exposure times.

I've not run across the term HDR, but the concept is familiar. I had learned about the exposure bracketing feature of the camera. Oddly, I didn't think to apply it.

Thank you both for your advice! The lesson to me is: get better as using the tools I have.

BTW, this is the same hydra in a lightfield picture:
Image


It's mostly colorless, which is why in darkfield it was a big blob of off-white, and the source of my exposure problems.

DaveW
Posts: 1702
Joined: Fri Aug 04, 2006 4:29 am
Location: Nottingham, UK

Post by DaveW »

I would have thought if you were using centre weighted metering, in view of that big white blob in the centre you should have got under rather than overexposure since the camera would try and expose it as mid grey, not white.

I have never done microscopy but could it simply be that the brightness range is beyond what the sensor can record and you are trying to bring up more detail in the dark background and simply overexposing the highlights (blooming)? See:-

http://www.digicaminfo.btinternet.co.uk ... posure.htm

As I say I know nothing about microscopes, but possibly one of the practitioners here could advise you how to light the image better to bring out detail and possible reduce the contrast range, though obviously dark-field must mean a dark background.

DaveW

DaveW
Posts: 1702
Joined: Fri Aug 04, 2006 4:29 am
Location: Nottingham, UK

Post by DaveW »

Getting the exposure for the flower shots in this link would be pertinent to your problem I would have thought, provided you do not overexpose the highlights in the first place.

http://www.digicaminfo.btinternet.co.uk/levels.htm

DaveW

Charles Krebs
Posts: 5865
Joined: Tue Aug 01, 2006 8:02 pm
Location: Issaquah, WA USA
Contact:

Post by Charles Krebs »

Dave...
When put on a microscope trinocular tube, it's not uncommon that auto camera metering will need a significant "bias" plugged in if you want to rely on the camera meter.

As you mentioned, it's not always easy to predict what a center-weighted reading will yield with a subject like the darkfield hydra example. A tiny hydra in a black field will likely be overexposed, whereas a shot where the bright hydra body "fills the frame" could very well provide underexposure.

s_r ...
I seriously doubt there's anything to be gained doing additive exposures, you are only making more work for yourself.

HDR, with varied exposures is a different matter, and is capable of remarkable results in extreme conditions. But it is impractical with subjects that have any movement at all between exposures. A well exposed raw file and a little computer work will generally do a good job for most darkfield shots. Occasionally you will find just one section or part of the subject that "blows out". If the subject is stationary, it's possible to take an overall "best" exposure, and then a second exposure at a faster shutter speed to record the problem area. Then the overexposed section in the "overall" can be replaced with the darker exposure. Sort of "HDR Lite". :wink:

Careful "reading" (and understanding) of the histogram makes this all infinitely easier.

DaveW
Posts: 1702
Joined: Fri Aug 04, 2006 4:29 am
Location: Nottingham, UK

Post by DaveW »

I was presuming it was not a radical crop Charles, and if not the white subject fills most of the centre of the frame which a centre weighted meter will be biased to read, so I would have expected underexposure not overexposure.

However it looks like a bad case of sensor blooming to me due to overexposure where the pixels concerned have "overflowed" into the surrounding ones burning out the highlights?

DaveW

rjlittlefield
Site Admin
Posts: 23564
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Underexposure is what I probably would have expected also.

But what one expects and what actually happens can be very different when the setup is not what the manufacturer designed.

For example, the manufacturer assumes that stopping down the lens by 2 stops will make the light on the sensor become 4X less bright.

But when the camera is pointed down a microscope eyepiece, that assumption is wrong.

At wide settings, the limiting aperture is in the microscope, not the camera. Stopping down the camera lens by 2 stops may have no effect at all on what gets to the sensor.

Metering is typically done with the lens wide open, predicting what the effect will be when the lens is closed. The manufacturer predicts 4X reduction at 2 stops down. If it doesn't change at all because of the added optics, then you end up with severe overexposure like we see here. Hence the need for correction as Charles mentioned.

What appears to be blooming is probably just OOF blur, augmented by some CA in the optics. Notice that some of the bright blobs are surrounded by blue fringes, others by yellow. This is a typical effect of longitudinal CA -- fringes in one color appear around bright spots in front of the focal plane, the other color around spots in back.

--Rik

specious_reasons
Posts: 107
Joined: Tue Feb 24, 2009 12:46 pm
Location: Woodridge, IL

Post by specious_reasons »

DaveW wrote:I was presuming it was not a radical crop Charles, and if not the white subject fills most of the centre of the frame which a centre weighted meter will be biased to read, so I would have expected underexposure not overexposure.

However it looks like a bad case of sensor blooming to me due to overexposure where the pixels concerned have "overflowed" into the surrounding ones burning out the highlights?
For the record, the picture was not cropped at all, just reduced in size. The hydra was free floating in a well slide., and it was much larger than the depth of focus at the magnification I was using. The thing that looks like a really bad bloom near the "head" is actually an out of focus tentacle.

I was also pretty shocked at the result from using center-weighted metering. I assume it's like what Rik suggested and the camera simply did not meter the image correctly. The next logical step should have been to make adjustments to the exposure, but the next logical step isn't always what comes to my mind....
Charles Krebs wrote:A well exposed raw file and a little computer work will generally do a good job for most darkfield shots.
I generally haven't been using raw mode, but I suspect this may change when I get difficult subjects such as this.

All of the information so far has been helpful, thank you!

DaveW
Posts: 1702
Joined: Fri Aug 04, 2006 4:29 am
Location: Nottingham, UK

Post by DaveW »

With digital it is always better to slightly underexpose than overexpose though since lost highlights cannot be recovered,

DaveW

Post Reply Previous topicNext topic