Histograms and jpeg: ETTR?

[richarddd]

Is the general advice to expose to the right when shooting jpeg, the same as when shooting raw?    Others had said so, but I just read John Shaw as recommending centering the histogram when shooting jpg.  http://www.johnshawphoto.com/tips/workflow_outline.html

[digitaldog]

Is the general advice to expose to the right when shooting jpeg, the same as when shooting raw?    Others had said so, but I just read John Shaw as recommending centering the histogram when shooting jpg.  http://www.johnshawphoto.com/tips/workflow_outline.html
[a href=\"index.php?act=findpost&pid=124342\"][{POST_SNAPBACK}][/a]

No, not really. Yes, you'd like all the data as close to the right end but you're out of the loop in the rendering here (its all in camera) so you'd want to make the histogram which represents the JPEG conversion NOT clip data you want to retain.

[John Sheehy]

Is the general advice to expose to the right when shooting jpeg, the same as when shooting raw?    Others had said so, but I just read John Shaw as recommending centering the histogram when shooting jpg.  http://www.johnshawphoto.com/tips/workflow_outline.html
[a href=\"index.php?act=findpost&pid=124342\"][{POST_SNAPBACK}][/a]

I expose to the right using my cameras' histograms at minimum camera (JPEG)contrast, and don't have any problems bringing them back down with about 0.8 gamma.  All the pale embedded JPEGs turn beautifully saturated when I do so.  ETTR applies, IMO, to all file formats.  The sacrifice with JPEGs is a little bit of posterization of the brightest highlights, but how sensitive are we there, anyway?

[richarddd]

I agree that the key is to avoid clipping and I don't see why raw v. jpeg should make a difference in this context (other than you may have to be more careful to avoid clipping the jpeg).

On the other hand, John Shaw is a very talented and knowledgeably photographer and I'm wondering why he would have a different view.

[AjantaKVS]

Hi,

Even with so much of discussion, I am not clear.

When shooting JPEGs

            Should we expose such that histogram centered with no clipping as advised in above referred article
                                       
        OR

        Should we expose such that histogram towards right with no clipping  contrary to above referred article and in line with RAWs method.

 Please be little bit clear, sorry I am poor at understanding

Particularly digital dogs answer here  put me in a little  confusion. though I have read lot of his CMS articles and digested it with no confusions and to be frank i have become a color  expert on reading his almost all CMS articles on the net

And also I want to Thank him for writing such great articles on CMS.

Thanks

AjantaKVS

[Jonathan Wienke]

If the histogram has room on both sides, it is always advantageous to shift it to the right so that there is only a small amount of space on the right. Always expose as much as you can without clipping important highlights where detail is still needed.

[Lisa Nikodym]

To answer, I'd have to ask why you are shooting jpg instead of raw?  If it's because you want a quick photo with a minimum of processing and don't want to spend the time doing the raw conversion, then I'd try to get the exposure exactly the way I want it in the camera and forget about ETTR, since ETTR would require an additional post-processing step to adjust the brightness.  Perhaps this is the reasoning that John Shaw is using.  If the time doing post-processing isn't the issue, however, then why do you want to use jpg instead of raw?  Smaller file size, or a camera that doesn't produce raw, or...?

Lisa

[AjantaKVS]

If the histogram has room on both sides, it is always advantageous to shift it to the right so that there is only a small amount of space on the right. Always expose as much as you can without clipping important highlights where detail is still needed.
[a href=\"index.php?act=findpost&pid=144030\"][{POST_SNAPBACK}][/a]

Does this statement  is true irrespective of Jpeg or raw shooting is the real confusion.

KVS

[AjantaKVS]

To answer, I'd have to ask why you are shooting jpg instead of raw?  If it's because you want a quick photo with a minimum of processing and don't want to spend the time doing the raw conversion, then I'd try to get the exposure exactly the way I want it in the camera and forget about ETTR, since ETTR would require an additional post-processing step to adjust the brightness.  Perhaps this is the reasoning that John Shaw is using.  If the time doing post-processing isn't the issue, however, then why do you want to use jpg instead of raw?  Smaller file size, or a camera that doesn't produce raw, or...?

Lisa
[a href=\"index.php?act=findpost&pid=144053\"][{POST_SNAPBACK}][/a]

I think things are getting clear.

Let me state my understanding from the above post, to make things  crystal clear.

 If I want immediate images with no post processing (whatever may be the reason for no post processing), shoot jpeg with no ETTR as advised by John Shaw.


If I shoot jpegs with ETTR, then post processing is a must and no escape from it
as in RAW, So instead of shooting jpeg with ETTR, you are advising me to shoot in RAW with ETTR to get better results. Is my understanding correct?.

If so, does the amount of time required to  post process "Jpeg with ETTR" and "RAW with ETTR" are same. Also the image processing skills requried for the both work flows same or does it differ a lot.

Any feed back and advice welcome.

Thanks
KVS

[Tom Maher]

I think things are getting clear.

Let me state my understanding from the above post, to make things  crystal clear.

 If I want immediate images with no post processing (whatever may be the reason for no post processing), shoot jpeg with no ETTR as advised by John Shaw.
If I shoot jpegs with ETTR, then post processing is a must and no escape from it
as in RAW, So instead of shooting jpeg with ETTR, you are advising me to shoot in RAW with ETTR to get better results. Is my understanding correct?.

If so, does the amount of time required to  post process "Jpeg with ETTR" and "RAW with ETTR" are same. Also the image processing skills requried for the both work flows same or does it differ a lot.

Any feed back and advice welcome.

Thanks
KVS
[a href=\"index.php?act=findpost&pid=144204\"][{POST_SNAPBACK}][/a]

From my understanding, you got it right.

[Lisa Nikodym]

If I shoot jpegs with ETTR, then post processing is a must and no escape from it
as in RAW, So instead of shooting jpeg with ETTR, you are advising me to shoot in RAW with ETTR to get better results. Is my understanding correct?.

Correct.

If so, does the amount of time required to post process "Jpeg with ETTR" and "RAW with ETTR" are same. Also the image processing skills requried for the both work flows same or does it differ a lot.

There is no easy answer to those questions.  It will be very different for different images.  Sometimes a JPG image starts out just right, and will require no adjustment.  Sometimes a RAW images starts out just right with the RAW converter's default settings, and will require no adjustment.  Doing just a brightness adjustment, which is what you would often need if you ETTR, is quick and easy.  But sometimes a difficult RAW image needs considerable adjustment to get it looking good, and doing a good job of it requires time and practice; but sometimes that same image, taken as JPG, cannot be rescued at all.  That is the advantage of RAW; with JPGs, you can sometimes have clipped highlights or severe color balance problems or other major problems that you could have fixed pretty easily if you were using RAW.  Most of the time, a JPG will require less post-processing; but sometimes (often!) an image with problems or difficulties (such as a large dynamic range) is much easier to fix in RAW than in JPG.

Lisa

[AjantaKVS]

Correct.
There is no easy answer to those questions.  It will be very different for different images.  Sometimes a JPG image starts out just right, and will require no adjustment.  Sometimes a RAW images starts out just right with the RAW converter's default settings, and will require no adjustment.  Doing just a brightness adjustment, which is what you would often need if you ETTR, is quick and easy.  But sometimes a difficult RAW image needs considerable adjustment to get it looking good, and doing a good job of it requires time and practice; but sometimes that same image, taken as JPG, cannot be rescued at all.  That is the advantage of RAW; with JPGs, you can sometimes have clipped highlights or severe color balance problems or other major problems that you could have fixed pretty easily if you were using RAW.  Most of the time, a JPG will require less post-processing; but sometimes (often!) an image with problems or difficulties (such as a large dynamic range) is much easier to fix in RAW than in JPG.

Lisa
[a href=\"index.php?act=findpost&pid=144227\"][{POST_SNAPBACK}][/a]

Hello,

Thanks a lot for clarification and people are really helpful here. What a nice way to learn  digital photography from all the stalwarts.

Thanks again

KVS

[Tim Gray]

Lisa's post is probably sufficient, but it helps to go back to first principles.

ETTR means expose as far right as possible, without clipping.  Why is that a "good thing"?

It's important to note that the amount of data available to each stop is one half of the available space of the next highest stop.  So for jpg, assuming 8 stops of dynamic range, the topmost stop gets 128 levels to work with and the bottom stop only  gets 2 levels.  So if you have to do any adjusting to the deep shadows, you have a basically impossible task.

For 12 bits, the top stop gets 4096 levels, and the bottom now gets 32 and for 14 bits, the numbers are 16384 and 128 respectively.  Note that at 14 bits, the number of levels for the bottom stop equals the number available for the top stop in a jpg, which now begs the question as to how important is ETTR in a 14 bit world....  In my view, it remains a "best practice" but the consequences of missing that last bit of headroom diminishes as the available bits go up.

I've certainly noticed that it's easier to dig high quality detail out of the shadows with a 1d3 than a 1d2.

[jjj]

John Shaw also says - If camera has auto-rotate feature to display vertical images on camera LCD as a vertical when camera is held horizontally, turn it off.  You should be working on a tripod so the camera will be in a vertical orientation already.

All those photos I've ruined over the years, by not using a tripod!  
Being inflexibly absolute like that line about tripods, undermines any good things he does have to say. I use Auto rotate as it saves me a lot of time rotating images in software later and even though I have tripods and monopods they rarely get used as they would be more likely to prevent me getting the shots I tend to take and not aid me at all.

As for the histogram thing, it kinda makes sense. Expose your JPEG correctly and your RAW files to the right, as you can tweak the RAW back to 'correct' exposure at leisure, whereas if you are using JPEGs you will tend to want them quicker or not bother so much with PS, LR..etc, so getting the exposure right in camera is a good idea. Though as the histogram is based on a JPEG of the RAW file [so I believe], even if you don't shoot JPEG, then the histogram is not too accurate with regard to the RAW file info anyway.

[John Sheehy]

All those photos I've ruined over the years, by not using a tripod!  
Being inflexibly absolute like that line about tripods, undermines any good things he does have to say. I use Auto rotate as it saves me a lot of time rotating images in software later and even though I have tripods and monopods they rarely get used as they would be more likely to prevent me getting the shots I tend to take and not aid me at all.[a href=\"index.php?act=findpost&pid=144926\"][{POST_SNAPBACK}][/a]

I shoot birds hand-held, often besides people using tripods.  I find that if you are shooting a bird that is stationary, and want to take a lot of shots, or wait for a certain activity, it can get a little tiring holding the camera and telephoto lens up for a long time.  That is when I think I might rather have a tripod at the moment.  However, when birds are moving about in trees, the tripod folks start to have their disadvantage, because of all the time investment in moving the tripod.  Windows of opportunity for shooting birds are often a 1/2 second or less.  Hand-held, I can move myself so that obstructions like leaves and branches can be avoided, very quickly.

[AjantaKVS]

Lisa's post is probably sufficient, but it helps to go back to first principles.

ETTR means expose as far right as possible, without clipping.  Why is that a "good thing"?

It's important to note that the amount of data available to each stop is one half of the available space of the next highest stop.  So for jpg, assuming 8 stops of dynamic range, the topmost stop gets 128 levels to work with and the bottom stop only  gets 2 levels.  So if you have to do any adjusting to the deep shadows, you have a basically impossible task.

For 12 bits, the top stop gets 4096 levels, and the bottom now gets 32 and for 14 bits, the numbers are 16384 and 128 respectively.  Note that at 14 bits, the number of levels for the bottom stop equals the number available for the top stop in a jpg, which now begs the question as to how important is ETTR in a 14 bit world....  In my view, it remains a "best practice" but the consequences of missing that last bit of headroom diminishes as the available bits go up.

I've certainly noticed that it's easier to dig high quality detail out of the shadows with a 1d3 than a 1d2.
[a href=\"index.php?act=findpost&pid=144868\"][{POST_SNAPBACK}][/a]

 Hello,

I think here is some maths mistake, if it is 12 bits the highest stop will have exactly half of total available levels, so it must be 2048 (not 4096 )levels and the last stop will have 16 levels not 32 as mentioned.

Similarly i think you have erred in calculations for other bit depths if my understanding is correct. Please correct me if i am wrong.

AjantaKVS

[Tim Gray]

Hello,

I think here is some maths mistake, if it is 12 bits the highest stop will have exactly half of total available levels, so it must be 2048 (not 4096 )levels and the last stop will have 16 levels not 32 as mentioned.

Similarly i think you have erred in calculations for other bit depths if my understanding is correct. Please correct me if i am wrong.

AjantaKVS
[a href=\"index.php?act=findpost&pid=145514\"][{POST_SNAPBACK}][/a]

Yes, I think you're right... xl formula error

[digitaldog]

Just noticed this is up on the web:

http://www.digitalphotopro.com/tech/exposing-for-raw.html

[John Sheehy]

Hello,

I think here is some maths mistake, if it is 12 bits the highest stop will have exactly half of total available levels, so it must be 2048 (not 4096 )levels and the last stop will have 16 levels not 32 as mentioned.

Similarly i think you have erred in calculations for other bit depths if my understanding is correct. Please correct me if i am wrong.
[a href=\"index.php?act=findpost&pid=145514\"][{POST_SNAPBACK}][/a]

In reality, the number of levels is irrelevant.  There is far too much noise in digital cameras for the number of levels to matter in a 12-bit RAW.  Quantization is not an issue.  Signal-to-noise ratio is the issue, and the farther to the right your signal is, the higher the signal-to-noise ratio.