Ad
Ad
Ad
Pages: « 1 2 3 [4] 5 »   Bottom of Page
Print
Author Topic: Underexposure or Overexposure?  (Read 19807 times)
Panopeeper
Sr. Member
****
Offline Offline

Posts: 1805


« Reply #60 on: January 08, 2008, 06:58:58 PM »
ReplyReply

Quote
There is an easy solution to solving the problem of excessive levels in the highlights and  the paucity of tones in the shadows and that is to change the encoding of the data

That's right, however this thread is about exposing to the right, and that is incompatible with the proposal of hightlight-compressed encoding.

Certainly, the encoding could save some of the levels, but not to the degree, as this can be done on an image directly before finalizing it.

Already the fact, that one of the channels (usually the red) may have to be multiplied by two or more in order to achieve white balance, should be enough to think twice before reducing the data.

Quote
after obtaining your ETTR image, did you bring the tones back down to their proper values with the exposure controls? Since the sensor is linear, the properly processed ETTR and non ETTR images should be quite similar, except the ETTR image may have better tonality and less noise

I remember to have seen post of you proving just the opposite.
Logged

Gabor
The View
Sr. Member
****
Offline Offline

Posts: 1001


« Reply #61 on: January 08, 2008, 07:01:13 PM »
ReplyReply

Quote
"Character" is not in the metadata of raw images :-)

You need to define in technical terms, what kind of changes you mean, or provide comparative examples to point out what you mean.
[a href=\"index.php?act=findpost&pid=166000\"][{POST_SNAPBACK}][/a]

I guess we are in the stage of developing that. After all, RAW hasn't been around for that many years.

Sharp, descriptive terms you can use when you have complete overview about what you want to say.

At the moment, I have the impression, we are still pretty much at the hunch stage. A sign of this are the amount of opposing theories circulating.

So, at this time, metaphors may be better, as they do not try to give sharp descriptions where we don't have that much knowledge to describe sharply.

Wherever aesthetics and technical terms meet...
« Last Edit: January 08, 2008, 07:05:39 PM by The View » Logged

Deserts, Cities, Woods, Faces - View of the World.
Panopeeper
Sr. Member
****
Offline Offline

Posts: 1805


« Reply #62 on: January 08, 2008, 07:04:07 PM »
ReplyReply

Quote
I think a non-linear interpretation across the zones introduced because of the zone-shift through ETTR might be a quantitative interpretation of "character change" ...

Exposing to the right is not meant to keep everything exposed higher than necessary.
Logged

Gabor
Panopeeper
Sr. Member
****
Offline Offline

Posts: 1805


« Reply #63 on: January 08, 2008, 07:09:29 PM »
ReplyReply

Quote
At the moment, I have the impression, we are still pretty much at the hunch stage

Fair enough regarding a sharp description, but you should be able to demonstrate what you mean. I guess when you mentioned this, you were thinking of specific cases you had observed. You should try to create such cases in controlled tests, i.e. shooting a serie with 1/3 stops apart, post-processing them, and pointing out the differences you find detrimental.
Logged

Gabor
The View
Sr. Member
****
Offline Offline

Posts: 1001


« Reply #64 on: January 08, 2008, 07:09:34 PM »
ReplyReply

Quote
Exposing to the right is not meant to keep everything exposed higher than necessary.
[a href=\"index.php?act=findpost&pid=166010\"][{POST_SNAPBACK}][/a]

The word "necessary" is key here.

How much exposure is necessary?

Isn't ETTR often voluntary over-exposure (to get important tones into the more brightness-levels area), and then reducing the exposure in the RAW software?

This raising of exposure in the camera and reducing the exposure in the software could deliver different results than exposing "correctly" right away...

Maybe ETTR is a great technical idea, which doesn't work out aesthetically?
Logged

Deserts, Cities, Woods, Faces - View of the World.
The View
Sr. Member
****
Offline Offline

Posts: 1001


« Reply #65 on: January 08, 2008, 07:13:23 PM »
ReplyReply

Quote
Fair enough regarding a sharp description, but you should be able to demonstrate what you mean. I guess when you mentioned this, you were thinking of specific cases you had observed. You should try to create such cases in controlled tests, i.e. shooting a serie with 1/3 stops apart, post-processing them, and pointing out the differences you find detrimental.
[a href=\"index.php?act=findpost&pid=166011\"][{POST_SNAPBACK}][/a]

I haven't tested it out so far.

Until now, I have just trusted ETTR, and done it, but doubts had started to form like clouds, and it may rain counter-arguments.

So, no, I haven't decided what to believe yet. And "to know" is even farther away. Which I, creatively, find exciting.

That's the most interesting part of discussion: when there's no firm belief, but exploration.

The tests, comparisons, explorations...
« Last Edit: January 08, 2008, 07:14:12 PM by The View » Logged

Deserts, Cities, Woods, Faces - View of the World.
djgarcia
Sr. Member
****
Offline Offline

Posts: 343



WWW
« Reply #66 on: January 08, 2008, 07:41:34 PM »
ReplyReply

Quote
Exposing to the right is not meant to keep everything exposed higher than necessary.
[a href=\"index.php?act=findpost&pid=166010\"][{POST_SNAPBACK}][/a]
Panopeeper, not sure about what your comment means relative to my statement, but I'm kind of dense tonight .

What I was trying to say is that if ETTR were to change the relative values across the brightness zones when development brings it back left to its intended position then that might be interpreted as a "change in character", meaning the tonalities have changed relative to each other thus giving the image a somewhat different look.

However this should hopefully not happen if the sensor and electronics are sufficiently linear and the development algorithms accurate enough, other than whatever might have been pushed out beyond the right margin is now hence blown.

The hoped for effect would be that the brightness values now in their proper place have finer steps differentiating them providing nicer gradients. At least the picture seems clear in my mind, though my verbal translation may have you befuddled and scratching your head   ...
Logged

Over-Equipped Snapshooter - EOS 1dsII & 1DsIII, Zeiss & Leica lenses
http://improbablystructuredlayers.net
Panopeeper
Sr. Member
****
Offline Offline

Posts: 1805


« Reply #67 on: January 08, 2008, 07:41:40 PM »
ReplyReply

Quote
The word "necessary" is key here.

How much exposure is necessary?

"Necessary" is the lesser issue. A real problem is, that you don't know, when you overexpose into clipping, and you won't know that even in the raw processing.

Consequently it can happen, that you think you exposed to the right, while the exposure is in fact over the right. This can lead to loss of contrast and to incorrect color.
Logged

Gabor
Panopeeper
Sr. Member
****
Offline Offline

Posts: 1805


« Reply #68 on: January 08, 2008, 07:49:39 PM »
ReplyReply

Quote
However this should hopefully not happen if the sensor and electronics are sufficiently linear and the development algorithms accurate enough, other than whatever might have been pushed out beyond the right margin is now hence blown.

I clicked on "add reply" before having seen this post of yours. You touched the sensitive point of ETTR.

I got into the habit of shooting exposure bracketing, or simply adjusting the exposure and shooting again, if I feel that the exposure is not the best. I admit to have made already many dozens of shots for a pano series of eight frames. Very unscientific, but the price of memory cards is constantly going down. (The "correct" exposure is particularly a problem with wide panos, which encompass a multitude of sceneries.)
Logged

Gabor
Jonathan Wienke
Sr. Member
****
Offline Offline

Posts: 5759



WWW
« Reply #69 on: January 09, 2008, 07:56:05 AM »
ReplyReply

My experience with ETTR (which I'm defining as exposing so that non-specular highlights are just short of clipping) is that when using ACR, it works 100% of the time with zero negative effects. The people who claim otherwise seem to be using RAW converters such as Capture One and some of the MFDB manufacturer converters that all seem to apply a strong contrast curve by default which one cannot always change. In addition to applying the default strong contrast curve, I suspect that these converters do not handle exposure adjustments as well as ACR.
Logged

eronald
Sr. Member
****
Online Online

Posts: 4126



« Reply #70 on: January 09, 2008, 08:29:37 AM »
ReplyReply

Quote
My experience with ETTR (which I'm defining as exposing so that non-specular highlights are just short of clipping) is that when using ACR, it works 100% of the time with zero negative effects. The people who claim otherwise seem to be using RAW converters such as Capture One and some of the MFDB manufacturer converters that all seem to apply a strong contrast curve by default which one cannot always change. In addition to applying the default strong contrast curve, I suspect that these converters do not handle exposure adjustments as well as ACR.
[a href=\"index.php?act=findpost&pid=166113\"][{POST_SNAPBACK}][/a]

Jonathan,

If you want the hi-levels of your image interpolated from rebuilt color channels, that's your right - but don't call this policy ETTR, call it ETFR (charitably read expose to the Far right).

ETTR = Expose to The True Right.
ETFR = Expose to the Far Right.

Please understand my position: I want software to give you your options (highlight recovery) and me my information (good indication of the real linearity domain of the hardware). If people start to consider that highlight recovery means "real" data, then we're going to head for another marketing war with manufacturers claiming added headroom DR that simply isn't there. I would have assumed that someone as smart as you would hesitate before sucking in the exhaust from the Adobe marketing engine.

Edmund
« Last Edit: January 09, 2008, 08:36:10 AM by eronald » Logged
Panopeeper
Sr. Member
****
Offline Offline

Posts: 1805


« Reply #71 on: January 09, 2008, 10:56:54 AM »
ReplyReply

Quote
If you want the hi-levels of your image interpolated from rebuilt color channels, that's your right - but don't call this policy ETTR, call it ETFR (charitably read expose to the Far right).

Jonathan wrote

which I'm defining as exposing so that non-specular highlights are just short of clipping

this does not include true recovery, i.e. guessing some pixels based on the others.

It is a different issue, what a particular raw processor calls "recovery". For example the "recovery" slider of ACR does not indicate true recovery but a different treatment of the very highlights.
Logged

Gabor
Jonathan Wienke
Sr. Member
****
Offline Offline

Posts: 5759



WWW
« Reply #72 on: January 09, 2008, 01:27:14 PM »
ReplyReply

Quote
Jonathan,

If you want the hi-levels of your image interpolated from rebuilt color channels, that's your right - but don't call this policy ETTR, call it ETFR (charitably read expose to the Far right).

I'm not talking about recovering blown highlights via software tricks. Where did you get that notion? I don't care if there are are a few scattered 1-2-pixel specular highlights that clip in an image; that is not going to kill perceived highlight detail even if you aren't using any highlight recovery. If you expose so that there are no clipped pixels at all, your main subject will usually be several stops underexposed and noise will be horrible.

If I'm shooting a bride in a white dress, I'll expose so that the brightest part of the dress is about 1/3 of a stop from clipping in the RAW, and apply a -1/3 to -1/2 stop exposure adjustment in ACR to get the overall luminance right. Same thing with clouds in a landscape shot, assuming they are the brightest objects in-frame.

I think channel reconstruction is a cool software trick that can sometimes save a shot from being unusable, but it's not a crutch I expect to support me all the time.
Logged

John Sheehy
Sr. Member
****
Offline Offline

Posts: 838


« Reply #73 on: January 09, 2008, 09:57:09 PM »
ReplyReply

Quote
I remember, in black-an-white film, when you exposed too much (and the negative got too "dense", you lost some skin details.
[a href=\"index.php?act=findpost&pid=165993\"][{POST_SNAPBACK}][/a]

With digital and RAW, the best "zone" is the one just short of clipping the RAW data.  It gets better, and better, as you increase exposure, and then, BOOM!; you clip.  Contrast this to print film, where you have maximum quality somewhere in a middle range, and contrast suffers and grain increases as you go towards the extremes and correct in the darkroom.
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8908


« Reply #74 on: January 10, 2008, 03:02:33 AM »
ReplyReply

It's 3 stops not 4 stops. However, my 20D doesn't have a spot meter mode so I can't confirm that.

Set the camera to manual mode, take a spot meter reading of the brightest part of the image, then increase exposure by 3 stops and all your worries are over. Perfect ETTR every time.  
Logged
jjj
Sr. Member
****
Offline Offline

Posts: 3541



WWW
« Reply #75 on: January 12, 2008, 05:57:48 PM »
ReplyReply

Maybe we should define different types of correct exposure if we adhere to the ETTR way of exposing. As a 'correctly' exposed JPEG will result in an 'underexposed' RAW file and conversely a 'correctly' exposed RAW file will result in an overexposed JPEG.
Obviously if you shoot both there is a bit of a problem!
Logged

Tradition is the Backbone of the Spineless.   Futt Futt Futt Photography
djgarcia
Sr. Member
****
Offline Offline

Posts: 343



WWW
« Reply #76 on: January 12, 2008, 06:15:50 PM »
ReplyReply

Who knows - sometimes correct looks boring and incorrect looks interesting.

Like the bad guy said, "Shoot them both ..."
Logged

Over-Equipped Snapshooter - EOS 1dsII & 1DsIII, Zeiss & Leica lenses
http://improbablystructuredlayers.net
jjj
Sr. Member
****
Offline Offline

Posts: 3541



WWW
« Reply #77 on: January 12, 2008, 06:25:52 PM »
ReplyReply

Quote
I'm arguing that there are times when it doesn't work best [Manual]; case in point - when you don't have time to deal with optimal exposure and must gamble, and both your subject and background will both have an opportunity to suddenly change by a few stops.  Reading Edmund in context, my immediate impression was one of a situation where anything can pop up instantly, and you don't have time to mess around with all parameters of shooting, so something must be simplified and sacrificed, and a -1 EC in an AE mode is the chosen sacrifice.
And if the subject is back lit, then the exposure will be even more off than if say using neutral AE.

Quote
When you replied that you shoot manual, in that context, it seemed that you referred to one fixed manual setting, which included enough headroom for whites in the brightest light.
It'll be set to 'correctly' expose what I think I'll want in the situation/scene and usually unless the subject is very contre jour/in shadow one way and well lit the opposite way, that works very well.


Quote
if you are going to insist that you won't ever miss any opportunities while metering, I think you're probably overrating yourself.  However good you are at manually adjusting exposure to meet a relatively immediate need, there is always a more immediate need that you won't be ready for, and I assume that was the context here.
I wasn't talking about metering to get exposure, I am talking about using a sensible/compromise manual exposure and then if I have to point the camera at something quickly, that is obviously different, I'll just spin shutter or aperture dial in the right direction by a guessed amount, safer than auto against a bark or bright background. It may not be perfect, but if you are shooting that fast, perfect exposure is really not the most important thing.

Quote
If you have to completely automate exposure, it may be safer to do so with -1 EC, which on cameras like the D3 and Canon mk3 cameras, is not a particularly big sacrifice at ISO 200s and 100, respectively.
-1EC is as I mention above, is useless if you are suddenly shooting contre jour, not unusual occurence either and less than optimum for recovery with RAW as well as it doesn't survive under exposure that well. Most people do not have a D3 or Mrk III camera either. Anyway I'd rather have a small camera thatn one of those monsters for street shooting. A GRD II with a D3 chip in it would be ideal.  

Quote
I believe the D3 allows manual AV and Tv with auto-ISO, which is even better than AE, as you can keep the manual setting in a range of lighting.[a href=\"index.php?act=findpost&pid=165799\"][{POST_SNAPBACK}][/a]
It actually makes no difference as the RAW/JPEG exposure is still altered even though the shutter and aperture are fixed.
Logged

Tradition is the Backbone of the Spineless.   Futt Futt Futt Photography
jjj
Sr. Member
****
Offline Offline

Posts: 3541



WWW
« Reply #78 on: January 12, 2008, 06:27:05 PM »
ReplyReply

Quote
Who knows - sometimes correct looks boring and incorrect looks interesting.

Like the bad guy said, "Shoot them both ..."
[a href=\"index.php?act=findpost&pid=166818\"][{POST_SNAPBACK}][/a]
With you 100% on that.
Logged

Tradition is the Backbone of the Spineless.   Futt Futt Futt Photography
John Sheehy
Sr. Member
****
Offline Offline

Posts: 838


« Reply #79 on: January 12, 2008, 08:40:43 PM »
ReplyReply

Quote
Who knows - sometimes correct looks boring and incorrect looks interesting.

Like the bad guy said, "Shoot them both ..."
[a href=\"index.php?act=findpost&pid=166818\"][{POST_SNAPBACK}][/a]

However, you can do anything you want to, "incorrectly", in post-processing, with a larger variety of options.
Logged
Pages: « 1 2 3 [4] 5 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad