Who has 40D HTP (Hightlight Tone Priority) on 40D experience?

And not just a few snaps, yall!

I tried a low key shoot on sunday with HTP hoping it would live up to it's name. I noticed some differing exposure between pics, where they were not different before, all in M mode.

(no I can't post these pics, but just understand that some are exposed differently with more than enough flash recycle time).

Anyone experience this with HTP that has ACTUALLY shot HTP a lot?

edit: oops, messed up the title a bit, can't edit it, oh well!

You used it with low key? High key would be more appropriate I think; yeah I don't think that's the best application for it. Canon states in its documentation that it is intended to help when you have significant portions of your image that have highlights like wedding dresses and clouds.

Tests by other photographers suggest that you get a boost in highlight dynamic range of a stop at the expense of noise in the shadow.

This was high key. Maybe it is functioning as intended, but thought I would try it. Exposure corrections from raw have been spot on. And it's not a major difference, maybe ev +/- 1/3. Enough for me to notice. I guess I should have specified that.

syntrix wrote in post #6089327
And not just a few snaps, yall!

I tried a low key shoot on sunday with HTP hoping it would live up to it's name. I noticed some differing exposure between pics, where they were not different before, all in M mode.

(no I can't post these pics, but just understand that some are exposed differently with more than enough flash recycle time).

Anyone experience this with HTP that has ACTUALLY shot HTP a lot?

edit: oops, messed up the title a bit, can't edit it, oh well!

You said it not me.

syntrix wrote in post #6089402
This was high key. Maybe it is functioning as intended, but thought I would try it. Exposure corrections from raw have been spot on. And it's not a major difference, maybe ev +/- 1/3. Enough for me to notice. I guess I should have specified that.

I don't understand what this means.

Oh my bad. I said low key as in "not critical". Not as in lighting.

3 light setup, metered, etc...

I can correct the images, but that wasn't why I was trying this HTP thing. I think I'll just stay away from HTP.

Which raw converter were you using? DPP is the only one that I know for sure supports HTP. I don't know about ACR or any of the others.

I've found that HTP shots that are imported via a RAW processor that doesn't understand HTP will make the shots look a bit underexposed, and the proper correction curve to apply isn't a simple one.

Good point KC! I did strictly either Lightroom 2.0 or CS3 export tests. I totally forgot that it was supported under DPP!

That's probably why!!!!!!

Lightroom 2.0 doesn't appear to support HTP? That's rather disappointing. I've been contemplating buying it since I've otherwise been pretty impressed with Adobe's ACR and Lightroom is reportedly a very good raw batch processor.

Guess I'll have to do some searching. I'm sure someone has figured out by now how to properly import an HTP shot into Lightroom and/or CS3...

Well, I didn’t “actually shoot HTP a lot”, but I understand how it works and I’ve tried it out, just to confirm my understanding that it would be faster and more convenient for me (as a RAW only shooter) to get the exposure right in the first place, or at least dial in negative exposure compensation, in order to avoid blown highlights (which is something that I would really hate) and tweak the tone curves in post processing if I want a smoother transition in the highlight range.

Basically, HTP underexposes by about 2/3 to 1 f-stop and alters the tone curve a bit to make for a smooth transition in the highlights.

If you shoot RAW and use any 3rd party RAW converter, you will probably only notice the underexposure and nothing else. In DPP however you can see the difference. Look at the two screenshots below.

I have taken shots of my computer screen (the POTN website…)

The first one shows an exposure of 1/13 sec. at f/4 with HTP turned off. The capture was overexposed and the white parts were clearly clipping. I had to pull back by half an f-stop to correct for this.

The second example shows the very same situation with HTP turned on. No exposure compensation required, no clipping. But if you look closely, you can see that the first shot has much more contrast.

Now especially look at the RAW gradation curves on the right of the images. You will notice two things:
1. The whole dynamic range has been shifted to the right by about 1 f-stop.
2. The gradation curve is much smoother in the highlights, thus lowering the contrast there.

DPP recognizes the HTP setting on the camera, indicates the intentional underexposure by shifting the dynamic range to the right and tweaks the tone curve for a softer roll-off in the highlights. This way the mid-level contrast isn’t affected by much, but the highlight range gains some room to “breath”.

With the first shot, taken without HTP, I could have done a similar thing by underexposing on purpose and adjusting the tone curve accordingly. But in this example it was not even necessary, because I was still able to retain the highlight range by simply pulling back the exposure. Of course I would have been able to do the same with the HTP-shot (if it had been still clipped).

HTP gives you an increased safety margin against blown highlights and is valuable when shooting JPEG only, but with RAW, you can have a similar effect by simply underexposing and probably tweaking the tone curve in pp.

HOSTED PHOTO
please log in to view hosted photos in full size.

HOSTED PHOTO
please log in to view hosted photos in full size.

I don't know that but the initial pics I was going over was an export to web from LR 2.0, and it's just the trial. I'm trying it out first before I upgrade.

Good info performa01! It would be interesting to test a few shots with HTP over normal M mode sometime soon! I'll try to do that next session.

When I first started messing with it, I thought that HTP was a simple underexposure as well.

But if I take my HTP shots, import them using something other than DPP, and add 1 stop of exposure compensation, the results I get are very, very different from the HTP shot. In fact, when I do that I get lots of blown highlights.

That suggests to me that the actual raw data contains something different than merely an underexposed shot, because if it were just that then I could add +1 stop of exposure compensation and get a correct-looking result.

Unless I'm missing something fundamentally important...

kcbrown wrote in post #6089756
But if I take my HTP shots, import them using something other than DPP, and add 1 stop of exposure compensation, the results I get are very, very different from the HTP shot. In fact, when I do that I get lots of blown highlights.

Initially, I was tempted to say, well, actually it is not one full f-stop, more like 5/6 – but that would not explain a huge difference.

So I tried it out and imported my two test shots in ACR. Fasten your seat belts now:
The shot without HTP needed a –0.45 exposure correction to get rid of any clipping.
But the shot with HTP turned on actually required at least –0.75 !!!

ACR is a generic RAW converter, not calibrated for any specific sensor, so all you have is some dubious numbers, even in the exposure compensation department there are no actual calibrated exposure values. This is one of the reasons why I normally prefer DPP.

But what I just found out is simply weird. The shot that is clearly underexposed, which is also the reason why you don’t have ISO 100 available any more, and why Canon states that you might get more noise etc. needs more negative exposure compensation in ACR – I don’t get it.

In DPP, I have a clear separation between operations and displays regarding the RAW data (1st tab sheet) and the resulting RGB data (2nd tab sheet). In ACR there is only one single histogram, obviously showing luminance + RGB, which indicates that it shows the already processed RAW data (since RAW data does not know about R, G and B initially).

RAW processing is a pretty complex process and has to be calibrated differently to each sensor type, regarding (at least) sensitivity, thresholds and the color matrix. You can do it in a way that only ensures that the final result will look correct (especially colors), or you can calibrate all intermediate steps, which isn’t easy because there is a mutual influence. It makes a big difference if you apply exposure compensation to the RAW data as a first step, or do it after demosaicing and performing the color matrix operations. Only in the first case it is equivalent to a real-world exposure compensation, e.g. by closing the aperture; once the raw data have been transformed to any RGB color space it’s a completely different thing, since any shifting in magnitude does not translate linearly to the original pixel values any more.

As a conclusion, ACR simply does not reflect the values that a photographer is accommodated to, it rather uses internal parameters.

Actually, now that I look at it, it seems ACR (version 4.3.1, in conjunction with PE6) does properly interpret the HTP raw image. At least, it seems to produce an image that looks the same as the non-HTP variant without any exposure compensation (this is with all the adjustments dialed to zero on both).

Bibble Pro, on the other hand, interprets the HTP shot as being underexposed. If I bring the exposure up by one stop, it generally looks the same as the non-HTP shot but some of the highlights are stronger afterwards than in the non-HTP shot. It's weird.

I guess the only way for me to really know what's being recorded is to use a RAW converter like dcraw, where you can control everything and thus equalize everything.

Without stirring up a debate I'll link to some threads from Naturescapes.net

Noise and Highlight Tone Priority

Highlight tone priority

HTP and the Canon 1Dm3

What HTP does