Could use some advice -- very frustrated!

Ysarex

Been spending a lot of time on here!
Joined
Nov 27, 2011
Messages
7,139
Reaction score
3,693
Location
St. Louis
Can others edit my Photos
Photos OK to edit
Hi All,

As many of you are aware I've taught photo at the college level for years. I've suffered with this problem for years now but it's getting worse as my students keep getting younger (yep, I know I'm the one not getting younger).;) Right now I have a group of first semester students in a beginning digital photo class. A little less than half of them have entry level DSLRs and the rest have various Coolpixs and Cybershots and Powershots and other assorted Crapshots.

I'm supposed to teach them how to use their cameras to best advantage, some design and composition basics, a little photo history, and some processing basics. The college supplies Photoshop in the lab and many of the students have software at home like Elements, iPhoto etc..

They've been taking landscape and cityscape photos over the past couple of weeks. My problem has to do with the processing basics part and their cameras. You'd think that one of the first things you'd teach a beginning photo student to do with their photos is to examine the histogram and evaluate it. Look for problems like clipped highlights for example and then learn to use that information to guide future control of the camera as well as suggest approaches for editing the photo at hand.

All of my students cameras SUCK! In fact my camera (5DmkII) sucks too in the very same way. Outside on blue-sky sunny days they all massively clip the blue channel. So one of my students, let's call her Emily, pays attention and goes out on a nice sunny day and takes a landscape photo front-lit with a clear blue sky. She gets a good exposure. She's been reading the book and knows enough now to look at a histogram in Photoshop and recognize what clipped highlights look like. She checks the histogram for her landscape and sees that her highlights are clipped to high heaven. (Pained look on her face) Emily asks, "what did I do wrong?"

What do I tell her? I look at her photo and see that the red and green channels are fine. I check the luminosity channel and it's compressed; if anything this photo is pretty flat and could use a contrast boost. And then there's that blue channel smashed up against the histogram right wall. The RGB composite channel takes the blue channel into account of course and so indicates badly clipped highlights.

I can show Emily the different color channels, but what's the explanation? And more importantly what do I tell her to do now? What to do next time she takes photos? What to do to process this photo?

An Example from my 5DmkII:

$screw_up.jpg

The inset histogram is the RGB composite and indicates clipped highlights. This is the JPEG the camera would have created. I used DPP to recreate it from the raw file. All the camera settings were at neutral when the photo was taken and I set a -.3 exp. comp. NOTE: The blue channel is massively clipped with a -.3 exp. comp. applied! Here's how badly the blue channel is clipped:

$blue_clipped.jpg

I did a threshold test of the blue channel. There's really nothing I could have done with the settings on the camera to avoid this at the time I took the photo. I reduced the saturation setting to the maximum and reprocessed the photo -- blue channel was better but still clipped and the rest of the photo looked sickly. So you can't tell me that appropriate settings in the camera will take care of this. Camera settings that would avoid clipping the blue channel would destroy the rest of the image.

Now this doesn't bother me personally because I have for a long time ignored camera processed photos and processing this photo from the raw file and not clipping the blue channel is quite easy to do. Went ahead and did that just to make sure this photo wasn't some strange special case:

$not_hard.jpg

So that wasn't hard, something I do almost daily. So then WHY CAN"T THE BLEEPIN' BLEEP CAMERAS DO IT?!!!

I can't tell Emily to just save raw files and learn to process them herself because her Coolpix, Cybershot, Powershot won't save raw files. Brittney has a new D3200 and I tried to say raw file to her -- that didn't go well, and because half the class has crapshot cameras I really have to deal with their camera processed photos.

So, the question again: What do I tell Emily?

I have dozens and dozens of student photos right now of blue-sky sunny day landscapes and cityscapes from my class and they have blown out blue channels. They're not bad exposures -- red and green channels are fine. A fair number of my students have learned to use the white balance presets and the color is pretty good, but the blue channels are clipped. A couple of my students have even gotten grey cards and set a custom white balance and the blue channels are clipped. Some of my students have shot brackets and even though they've set negative exp. comp factors the blue channels are clipped. These photos are not backlit, they're not even sidelit.

What do I tell Emily?

1. Don't worry you're doing OK it's just that your camera sucks.
2. Oh yeah I forgot to tell you, you can't include the sky when you take landscapes.
3. No you didn't do anything wrong, there's an exception to the clipped highlights rule -- blue skies.
4. Ashton was lying in that commercial you saw.
5. Camera image processors don't work very well, so yeah your camera sucks.
6. No you didn't do anything wrong and there's nothing you can change on the camera to make it work, you're just screwed.
7. Did you know that the earliest b&w films were excessively blue light sensitive and photographers like Henry Jackson had to take two exposures for landscapes; one for the sky and another for the foreground? Welcome to the past!

I could just try and ignore it and smile and say that's a nice picture, but Emily read the part of the book about the histogram and I've got to deal with her cognitive dissonance. It's getting harder and harder to explain to Emily that the software in her camera sucks and there's nothing much she can do about it (I'm not nearly as cute as Ashton). It's very frustrating and I feel like I'm less and less capable of giving them satisfactory answers. This younger generation doesn't seem to cope as well as earlier generations; they don't like peeking behind the curtain or being told their technology emperor isn't fully clothed.

Thanks for listening,
Joe
 
Last edited:
1, 4, 5 & 6

Seriously, you are teaching them the theory of processing raw files for maximum information when (1) they don't have raw files, and (2) one can get a pretty good image with one of the channels clipped, up to a point. Probably, you should tell them that one should try to avoid clipping channels, but in-camera processing sometimes does it and still gives you a pretty good image, and if they get better cameras they'll have better control over this and other things.

I never taught photography, but taught something else (technical) and found that it was very difficult to make any absolute statement that I didn't end up qualifying later. I frequently had to tell them things their textbooks said were incorrect - sometimes people actually get part-way through college without having heard that from anyone, which I always found kind of amazing.
 
Joe, you tell them that histograms don't help much with P&S cameras because so much of the process is out of their hands and handled by the "automatic" software in the camera. You start regaining some of that control when you go to DSLR's. And some more control when you shoot RAW. But if you have a P&S, it's just theory, as is the concept of bokeh, or "shallow depth of field", or focus placement, even the "exposure triangle". Granted, there is a wide range of capabilities in P&S cameras so the previous statement does not apply to ALL P&S, but the point is that using many of the tools we consider basic is difficult with P&S. I suppose the same applies to cameras in smart phones, iPads and other touch-pads, etc. Can't help much except commiserate.
 
Hi Joe, I may be being a bit daft but as far as I knew the histogram is a gude. The real question is do the blues in your and your students camera look off?

In the pic you posted looks fine to me. And that is the ultimate goal. As photographers we should know the technical detail to produce the shot that's in our head..... but I dont see the problem. If it looks good that is what you want

Show your students the digital rev video pro tog cheap camera
 
When you shoot high key images, are your photos clipped?

Are high key images wrong, because they've got clipped highlights?

What about low key images? Ever shoot so dark you have clipping in the shadows?

Are low key images wrong, because they're got clipped shadows?

If I had them, what I would explain to my students is that a histogram is helpful -- but it's not infallible. As weepete said, it's just a guide...

I think the problem is you're explaining the histogram as a "rule" for good photos...
 
Let me also say that you and your students should be thankful you have that sort of sun in the middle of winter.

It just rains here. Cold, dark, grey, cold rain all day. :(:(:(:(

So tell your students: be thankful you have clipping in Feb!
 
1, 4, 5 & 6

Seriously, you are teaching them the theory of processing raw files for maximum information when (1) they don't have raw files, and (2) one can get a pretty good image with one of the channels clipped, up to a point. Probably, you should tell them that one should try to avoid clipping channels, but in-camera processing sometimes does it and still gives you a pretty good image, and if they get better cameras they'll have better control over this and other things.

I never taught photography, but taught something else (technical) and found that it was very difficult to make any absolute statement that I didn't end up qualifying later. I frequently had to tell them things their textbooks said were incorrect - sometimes people actually get part-way through college without having heard that from anyone, which I always found kind of amazing.

Thanks. Yeah, I don't mind calling a textbook wrong -- happens more and more these days, but in this case I'm agreeing with the textbook that clipped diffuse highlights are a bad thing except in special and rare cases. In this case that has me ranting it is only a single channel that's clipped but a blue sky is not a special and rare case -- happens a lot in fact. I think part of the problem is that I understand it doesn't have to be that way and that the sensor capture didn't clip anything.

So I tell Emily you didn't overexpose the photo it's a software error, but you can't fix it or prevent it in the future? Don't worry, be happy. That's not too helpful. Yeah, I've said things like, "there's only so much you can expect from automated algorithms and unless you get a camera that can save the sensor data and learn how to hand process your photos.....but that's a commitment with a big learning curve not to mention maybe a new camera. (Emily actually has a Canon Rebel and she could save raw files, but half the class can't). I think part of my problem is that I feel responsible in a college class to actually expect them to learn something and it's hard for me to switch back and forth tolerating from them what I don't tolerate myself. I'm left frustrated and uncomfortable telling them, "yep that's wrong but there's not much we can do about it so let's move on." My! That's a pretty blue sky!

Joe
 
Joe, you tell them that histograms don't help much with P&S cameras because so much of the process is out of their hands and handled by the "automatic" software in the camera. You start regaining some of that control when you go to DSLR's. And some more control when you shoot RAW. But if you have a P&S, it's just theory, as is the concept of bokeh, or "shallow depth of field", or focus placement, even the "exposure triangle". Granted, there is a wide range of capabilities in P&S cameras so the previous statement does not apply to ALL P&S, but the point is that using many of the tools we consider basic is difficult with P&S. I suppose the same applies to cameras in smart phones, iPads and other touch-pads, etc. Can't help much except commiserate.

I appreciate the commiseration, thanks. I keep telling myself it's legit in a college class to include folks with a less serious interest and make the class positive for them while still maintaining some reasonable standards. Just been saying that a lot lately. Kinda like my son who teaches physics for physics majors and then this other class they jokingly call "physics appreciation" for non majors.

Joe
 
Hi Joe, I may be being a bit daft but as far as I knew the histogram is a gude. The real question is do the blues in your and your students camera look off?

In the pic you posted looks fine to me. And that is the ultimate goal. As photographers we should know the technical detail to produce the shot that's in our head..... but I dont see the problem. If it looks good that is what you want

Show your students the digital rev video pro tog cheap camera

Yes the blue skies in my student's photos look off and the pic I posted does not look fine to me. When the blue channel is clipped it shows. Here's another example where the camera software badly clipped the blue channel. It looks off to me. If it looks good to you then we have different standards. Thanks.

Joe

$blue_clip.jpg
 
You could have all the students buy K1000's (or buy a bunch and loan them out to the students) and shoot negative film, which has a higher dynamic range (15 stops or so) than those digital cameras (probably much less than 12 stops). Then you won't have to worry about clipping highlights in scenes with moderate contrast like that landscape scene.

You can still teach digital post-processing techniques, of course, by scanning the film.

Or use a stronger negative exposure compensation. -0.3 is not very much compensation. I'm thinking of a full stop or two to prevent the sky from clipping. Then curves or dodging in post can bring the overall exposure back up to give an acceptable image. There doesn't appear to be much of anything buried in shadow in the scenes you posted, so there shouldn't be a noticeable increase in noise in the darker regions with this technique.
 
When you shoot high key images, are your photos clipped?

Nope. And I love shooting high key photos like a white-painted wall in the sun.

$yellow_green.jpg

Are high key images wrong, because they've got clipped highlights?

Mine don't have clipped highlights. There are always exceptions but clipped highlights are most often wrong in any photo. I think they're especially wrong if they were otherwise avoidable and they don't enhance the image.

What about low key images? Ever shoot so dark you have clipping in the shadows?

Are low key images wrong, because they're got clipped shadows?

An entirely different thing, solid black is often very acceptable and even desirable in a photo.

If I had them, what I would explain to my students is that a histogram is helpful -- but it's not infallible. As weepete said, it's just a guide...

I think the problem is you're explaining the histogram as a "rule" for good photos...

As I said, I do agree with the textbook in this case that highlight clipping is generally bad. I'm pretty happy with that as a rule. I'm happy to break the rules if there's a positive result. I've seen few photos that break that rule to positive effect, but I've taken some myself when it was warranted. In the case of a sunny day blue sky I'm happy to insist that, as a rule, any channel clipping is an error.

Joe
 
Last edited:
When you shoot high key images, are your photos clipped?

Nope. And I love shooting high key photos like a white-painted wall in the sun.

View attachment 37055

Are high key images wrong, because they've got clipped highlights?

Mine don't have clipped highlights. There are always exceptions but clipped highlights are most often wrong in any photo. I think they're especially wrong if they were otherwise avoidable and they don't enhance the image.

What about low key images? Ever shoot so dark you have clipping in the shadows?

Are low key images wrong, because they're got clipped shadows?

An entirely different thing, solid black is often very acceptable and even desirable in a photo.

If I had them, what I would explain to my students is that a histogram is helpful -- but it's not infallible. As weepete said, it's just a guide...

I think the problem is you're explaining the histogram as a "rule" for good photos...

As I said, I do agree with the textbook in this case that highlight clipping is generally bad. I'm pretty happy with that as a rule. I'm happy to break the rules if there's a positive result. I've seen few photos that break that rule to positive effect, but I've taken some myself when it was warranted. In the case of a sunny day blue sky I'm happy to insist that, as a rule, any channel clipping is an error.

Joe

Ahh... Try again with high key, sometimes you want clipping:

18 Beautiful Bright Examples of High Key Photography

And, with your example did you use your MarkII? How would that come out with your students point and shoot?
 
When you shoot high key images, are your photos clipped?

Nope. And I love shooting high key photos like a white-painted wall in the sun.

View attachment 37055



Mine don't have clipped highlights. There are always exceptions but clipped highlights are most often wrong in any photo. I think they're especially wrong if they were otherwise avoidable and they don't enhance the image.



An entirely different thing, solid black is often very acceptable and even desirable in a photo.

If I had them, what I would explain to my students is that a histogram is helpful -- but it's not infallible. As weepete said, it's just a guide...

I think the problem is you're explaining the histogram as a "rule" for good photos...

As I said, I do agree with the textbook in this case that highlight clipping is generally bad. I'm pretty happy with that as a rule. I'm happy to break the rules if there's a positive result. I've seen few photos that break that rule to positive effect, but I've taken some myself when it was warranted. In the case of a sunny day blue sky I'm happy to insist that, as a rule, any channel clipping is an error.

Joe

Ahh... Try again with high key, sometimes you want clipping:

18 Beautiful Bright Examples of High Key Photography

You can want anything; if it's intentional and you got what you wanted fine. I can do high key and not want to clip the highlights. My students aren't trying to do high key they're trying to take sunny day landscapes and the blue sky isn't white paint in the sun.

And, with your example did you use your MarkII? How would that come out with your students point and shoot?

Actually I used my little compact Samsung, but it does save raw files. At least you could try and defend the camera claiming glossy white paint in the sun is pretty touchy. Should blue sky on a sunny day be considered pretty touchy?

Joe
 
Ok, so if youre asking really why cameras just dont clip the highlights out of the camera on a jpeg, i guess the answer is simply there is no market for it.

There is a market for point and shoots that take a picture that looks good out of camera

When people start paying more for a point and shoot that doesnt clip highlights according to your histogram rule, camera makers will make such cameras.
 
also, the sun is BRIGHT -- the brightest thing youll ever get a chance to photograph

what time of day are you taking these clipped photos? is your back to the sun or youre pointing slightly in its direction

maybe your answer is simply the sun will require such a
fast exposure to be too far outside your cameras dynamic range
 

Most reactions

New Topics

Back
Top