Ad
Ad
Ad
Pages: [1] 2 3 »   Bottom of Page
Print
Author Topic: Canon Highlight Priority  (Read 72157 times)
Gary Ferguson
Sr. Member
****
Offline Offline

Posts: 522


WWW
« on: November 25, 2007, 07:29:43 AM »
ReplyReply

I'd be interested if anyone can help explain the method and practical advantages of Canon's new "Highlight Priority" setting, available on the 40D, 1D MkIII, and the forthcoming 1Ds MkIII. Bob Atkins gave a good analysis here,

http://www.bobatkins.com/photography/digit...D_review_6.html

but he concludes that there's still some unknowns in Canon's method. Does anyone know of further analysis, and has anyone any comments based on practical experience?

Thanks.
Logged
John Sheehy
Sr. Member
****
Offline Offline

Posts: 838


« Reply #1 on: November 25, 2007, 08:13:30 AM »
ReplyReply

Quote
I'd be interested if anyone can help explain the method and practical advantages of Canon's new "Highlight Priority" setting, available on the 40D, 1D MkIII, and the forthcoming 1Ds MkIII. Bob Atkins gave a good analysis here,

http://www.bobatkins.com/photography/digit...D_review_6.html

but he concludes that there's still some unknowns in Canon's method. Does anyone know of further analysis, and has anyone any comments based on practical experience?

Thanks.
[a href=\"index.php?act=findpost&pid=155741\"][{POST_SNAPBACK}][/a]

The only unknowns are how each converter deals with it.

The RAW data are just taken at double the ISO, with half the exposure.  The conversions should bring the grey point up 1 stop, but roll in the extra highights.  Allegedly, some converters, like ACR, bring up the grey point but don't roll in the highlights automatically.

There is definitely no intelligent assessment of highlights, if that is what you are hoping for; just a fixed, dumb exposure compensation at a higher ISO, which mandates noisier shadows.

EDIT - that should have been at a "lower" ISO, IOW, the higher ISO exposure is used with the gain and digitization of the lower ISO.
« Last Edit: November 25, 2007, 08:41:11 PM by John Sheehy » Logged
Gary Ferguson
Sr. Member
****
Offline Offline

Posts: 522


WWW
« Reply #2 on: November 25, 2007, 08:53:19 AM »
ReplyReply

John, are you saying it's the same as doubling the ISO then dialling in one stop's worth of underexposure?
Logged
Slobodan Blagojevic
Sr. Member
****
Online Online

Posts: 5530



WWW
« Reply #3 on: November 25, 2007, 01:07:06 PM »
ReplyReply

This has been discussed in this thread:

http://luminous-landscape.com/forum/index....50&#entry153950

A lengthy discussion on the same subject at DP Review site: http://forums.dpreview.com/forums/read.asp...essage=25080916

Here is my take: It is useful if you are shooting JPEG: it adds a different tone curve to open up the mid-tones and shadows. Not that you could not come up with the same (or better) curve yourself though. In RAW it does nothing, except underexposing one stop (by shooting as if at the next full-stop lower ISO).

I think Bot Atkins got it wrong by claiming that HTP impacts RAW too (he is admitting several times in the article he is not sure what is going on with HTP). His update (to the initial review) and examples prove nothing, IMHO. What he did is the following:

"... RAW images shot at ISO 200, "developed" with a -2 exposure compensation..."

He used the same method for both non- and HTP files, which appears to be logical, but this is what really happened in my opinion:

Since HTP exposures are already underexposed by one stop, developing them with the same -2 compensation resulted in non-HTP files being underexposed by 2 stops, and HTP files being underexposed by 3 stops. No wonder the 3 stops underexposure retains more detail in the highlights.
Logged

Slobodan

Flickr
500px
Gary Ferguson
Sr. Member
****
Offline Offline

Posts: 522


WWW
« Reply #4 on: November 25, 2007, 02:11:36 PM »
ReplyReply

That's clear, thanks.
Logged
John Sheehy
Sr. Member
****
Offline Offline

Posts: 838


« Reply #5 on: November 25, 2007, 08:45:30 PM »
ReplyReply

Quote
John, are you saying it's the same as doubling the ISO then dialling in one stop's worth of underexposure?
[a href=\"index.php?act=findpost&pid=155752\"][{POST_SNAPBACK}][/a]

I said it partly wrong, and I've edited the statement.

More precisely, if you have the camera set to ISO 200 and HTP, then the RAW image is exactly the same as if you had shot at ISO 100 with -1 EC.

The ironic thing, which shows how screwed up these big corporations get, is that it is not available for ISO 3200 on the 40D, yet ISO 3200 IS  ISO 1600 under-exposed by one stop - they just double the numbers, tossing away a stop of highlight headroom.
« Last Edit: November 29, 2007, 05:58:02 PM by John Sheehy » Logged
Gregory
Full Member
***
Offline Offline

Posts: 191


WWW
« Reply #6 on: November 28, 2007, 10:07:19 PM »
ReplyReply

having read the linked discussions, I only have one question. with HTP turned on, does the down-stepping of the actual ISO only occur when the sensor senses that highlights would otherwise be clipped? or is the ISO stopped down regardless of what is being photographed?

ftr, I use HTP all the time and it works quite well although most of my images are dark and need exposure adjustments in Aperture after transfer from the camera. now I understand why.
Logged

Gregory's Blog: An Aussie in HK
Equipment: Canon EOS 1D Mark III, 17-40L, 24-105L, 70-300 DO
Slobodan Blagojevic
Sr. Member
****
Online Online

Posts: 5530



WWW
« Reply #7 on: November 29, 2007, 10:30:12 AM »
ReplyReply

Quote
having read the linked discussions, I only have one question. with HTP turned on, does the down-stepping of the actual ISO only occur when the sensor senses that highlights would otherwise be clipped? or is the ISO stopped down regardless of what is being photographed?

ftr, I use HTP all the time and it works quite well although most of my images are dark and need exposure adjustments in Aperture after transfer from the camera. now I understand why.
[a href=\"index.php?act=findpost&pid=156837\"][{POST_SNAPBACK}][/a]

No, the sensor is not that smart, so it underexposes regardless of what is being photographed.

To the best of my knowledge, Aperture is not HTP-friendly, i.e., it does not recognize the HTP tag the way Canon DPP and Adobe CR do. I am not sure about other software.

As for keeping HTP on all the time, I am not sure it is a good strategy. It makes sense if you are shooting scenes with a lot of important highlights at risk of clipping, but for general work the downside is that you are simply underexposing majority of your shots, risking more noise, for instance.
Logged

Slobodan

Flickr
500px
John Sheehy
Sr. Member
****
Offline Offline

Posts: 838


« Reply #8 on: November 29, 2007, 06:05:38 PM »
ReplyReply

Quote
The ironic thing, which shows how screwed up these big corporations get, is that it is not available for ISO 3200 on the 40D, yet ISO 3200 IS  ISO 1600 under-exposed by one stop - they just double the numbers, tossing away a stop of highlight headroom.
[a href=\"index.php?act=findpost&pid=155981\"][{POST_SNAPBACK}][/a]

I just read mypost again, and thought that perhaps I could be clearer.  Basically, ISO 3200 on the 40D, which does not offer HTP for that ISO, is actually the same thing as HTP at ISO 3200 would be (same noise levels), but with the extra highlights thrown away (when the RAW numbers are doubled and clipped at 4095).
Logged
203
Guest
« Reply #9 on: December 01, 2007, 09:03:19 PM »
ReplyReply

You all might get some real answers over at prophotohome, where Chuch Westfall seems to answer questions on a regular basis.
Logged
John Sheehy
Sr. Member
****
Offline Offline

Posts: 838


« Reply #10 on: December 02, 2007, 08:20:00 AM »
ReplyReply

Quote
You all might get some real answers over at prophotohome, where Chuch Westfall seems to answer questions on a regular basis.
[a href=\"index.php?act=findpost&pid=157592\"][{POST_SNAPBACK}][/a]

Chuck appears on other forums I've been on, and very tight-lipped about anything going on under the hood - you'll get much more useful answers by reverse-engineering the camera's output.
Logged
203
Guest
« Reply #11 on: December 03, 2007, 12:53:04 PM »
ReplyReply

Here's some info:

http://www.prophotohome.com/forum/canon-1-...light-prio.html
Logged
keith_cooper
Sr. Member
****
Offline Offline

Posts: 438



WWW
« Reply #12 on: December 05, 2007, 09:11:21 AM »
ReplyReply

Quote
Here's some info:

http://www.prophotohome.com/forum/canon-1-...light-prio.html
[a href=\"index.php?act=findpost&pid=157966\"][{POST_SNAPBACK}][/a]
Not much help as far I could see (without a subscription) - am I missing something, or could someone who has seen it précis the answer?
Logged

jerryrock
Sr. Member
****
Offline Offline

Posts: 558



WWW
« Reply #13 on: December 05, 2007, 09:58:47 AM »
ReplyReply

Quote
Not much help as far I could see (without a subscription) - am I missing something, or could someone who has seen it précis the answer?
[a href=\"index.php?act=findpost&pid=158401\"][{POST_SNAPBACK}][/a]

This is the response from Chuck Westfall:

"We haven't made any claims about increasing dynamic range with our current image sensors. Instead, we've said that dynamic range remains about the same or similar to previous generations despite higher resolution and smaller photosites. (Check 1Ds Mark III vs. 1Ds Mark II for example.)

Highlight Tone Priority mode has no effect on the actual dynamic range of the image sensor. It's just an alternative method of image processing that preserves more highlight detail than our standard processing, without significantly altering midtones or shadows. The effect of HTP is enhanced by our 14-bit A/D converter, which provides finer tonal gradations than the previous 12-bit system."
__________________
Chuck Westfall
Director/Media & Customer Relationship
Camera Division/Canon U.S.A., Inc.
Logged

Gerald J Skrocki
skrockidesign.com
keith_cooper
Sr. Member
****
Offline Offline

Posts: 438



WWW
« Reply #14 on: December 05, 2007, 10:54:25 AM »
ReplyReply

Quote
This is the response from Chuck Westfall:
...
Thanks very much for that.

I've only done a few (very) quick comparisons with my 1Ds3 since I'd suspected that I could get much the same results with appropriate exposure and subsequent attention to raw processing.

It does make a difference with in camera jpegs (as expected) but I don't think I'll be using it in normal day to day raw shooting
Logged

203
Guest
« Reply #15 on: December 05, 2007, 11:14:01 AM »
ReplyReply

The HTP is something which I will try and probably use a lot. From tests I've seen, it works well to preserve highlights AND also it seems to render much more shadow detail. The less time I spend in post production futzing around with these files trying to get highlights back, the better, IMO.
Logged
keith_cooper
Sr. Member
****
Offline Offline

Posts: 438



WWW
« Reply #16 on: December 05, 2007, 11:22:12 AM »
ReplyReply

Quote
... The less time I spend in post production futzing around with these files trying to get highlights back, the better, IMO.

I'd hardly call having a tone curve (or two) that offer levels of HTP much 'futzing around'... Still I suppose it depends on the type of work you do and your workflow.  It's just that in a series of images there may be as many I don't want it applied to, as to ones I -do- want it applied to. The tone curve approach gives a flexibility that I think would more suit my own workflow.

But who knows, I've barely had it for a week so far :-) :-)
Logged

espressogeek
Jr. Member
**
Offline Offline

Posts: 98


« Reply #17 on: December 05, 2007, 11:33:07 AM »
ReplyReply

If you are shooting raw would HTP be any better than applying a good curve to an ISO 100 shot?
Logged
djgarcia
Sr. Member
****
Offline Offline

Posts: 343



WWW
« Reply #18 on: December 05, 2007, 12:25:43 PM »
ReplyReply

Might not the two approaches be somewhat complementary? HTP does affect the raw data (I believe?) and curves is post-shoot. And there's the old "choose your exposure" thing too ...
Logged

Over-Equipped Snapshooter - EOS 1dsII & 1DsIII, Zeiss & Leica lenses
http://improbablystructuredlayers.net
203
Guest
« Reply #19 on: December 05, 2007, 12:58:21 PM »
ReplyReply

Quote
If you are shooting raw would HTP be any better than applying a good curve to an ISO 100 shot?
[a href=\"index.php?act=findpost&pid=158440\"][{POST_SNAPBACK}][/a]

I think the trick is finding a curve which does as good a job as the HTP does. It seems to be easier said than done. If this was just a curve issue, then Canon could have simply updated DPP and been done with it, I presume.

In any case, I have had pretty good luck with the Photoshop Highlight and Shadow tools...


-r
Logged
Pages: [1] 2 3 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad