ETTL, Bar in the middle, or ETTR?

R

reheat module

Suspended / Banned
Messages
2,681
Name
Paul
Edit My Images
No
I have asked this question before, but technology has marched on...
Do you ETTL, use mid bar marker, or ETTR?
I'm aware of the earlier theories of ETTR due to the attributes of the sensor build, and light capture, but have never used ETTL.
I personally use mid bar marker and the RAW capability to recover, if needed + or - 2 stops.
Am I missing out on something?
 
ETTL isn't anywhere near predictable enough to make sweeping generalisations about overriding it.

I think the best tip is to set centre weighted metering for flash, and I'm more likely to set -FEC than +FEC, but I wouldn't suggest any hard and fast rules.
 
Hi guys, - apologies but it's obviously my bad phrasing of the question
I was referring to the use of the histogram in Exposing To The Left, or Exposing To The Right for the sensor.
I'm using the 5D Mk3 and wondering if the technology means keep the bar in the middle in manual mode, the sensor's good enough?
Hope that makes sense
 
I always found with Canon that I had to ETTR between 1/3 of a stop and a full stop otherwise right in the middle gave a slightly underexposed shot.
ETTR has always been good for shadow preservation.
 
I've always been a "proper exposure" guy... I'll even let it underexpose whites or over expose blacks (meter centered) in tough situations. I will use ETTR with really dark scenes if possible because recording a lack of light easily creates editing issues (banding/etc).
I have never heard of ETTL in this context, but I found a good article on it... looks interesting ;). https://petapixel.com/2017/04/01/expose-left-really-maximize-image-quality-night/
 
reheat module said:
Hi guys, - apologies but it's obviously my bad phrasing of the question
I was referring to the use of the histogram in Exposing To The Left, or Exposing To The Right for the sensor.
I'm using the 5D Mk3 and wondering if the technology means keep the bar in the middle in manual mode, the sensor's good enough?
Hope that makes sense
Click to expand...
Now I'm confused.
This post is in 'lighting and studio' and you mention ETTL (Canon auto flash mode). But it seems that was a red herring.

Am I to understand you're simply talking about exposing yo the left or right in Manual mode?

The answer to which is that there's no rule. The whole point of Manual mode is to have the option of overriding the metered setting, and you'd override it depending whether your subject is brighter or darker than 'average' so you go Right for bright subjects and Left for dark ones.
 
Orangecroc said:
Are you sure it's not the other way around to prevent blowing the lighter subject and to raise details in the shadows for darker ones?
Click to expand...
I'm sure.

If you have a white wall, a grey wall and a black wall, the meter reads them all as grey, if you want them rendered correctly, you use compensation the way I posted.
 
I’m a little confused as to what the OP is asking. Are you referring to whether to add +/- exposure compensation when using the auto modes (P/A/S) or about generally whether to under or over expose an image. If it is about exposure compensation then I’m with Stuart, on my 5D3 I generally used to have about +2/3 exposure compensation dialled in. Obviously once checking you’re first image you can adjust accordingly if needed.

If you are asking about how to expose in general then the true answer should be to expose in the middle, however due to the limited DR of sensors providing you don’t go too far the logic is as follows:

By increasing the exposure in PP (or raising the shadows) to bring out the low end detail you also introduce noise which degrades the image, lowering the exposure or reducing highlights doesn’t. Therefore the thinking was always to overexpose and lower the exposure to bring back the details.

I’m sure others will chip in with more technical info about the amount of detail that is recorded at each level, however my info is more based on what I see in Lightroom.
 
Phil V said:
I'm sure.
If you have a white wall, a grey wall and a black wall, the meter reads them all as grey, if you want them rendered correctly, you use compensation the way I posted.
Click to expand...
Orangecroc said:
That makes sense, I still don't totally understand how the metering works totally.
Click to expand...

It might be a little easier to understand it as: the meter tells you if something *will be recorded* as mid-tone ("grey" when centered), or brighter/darker. If something is brighter than mid-tone (i.e. white) and you want to record it brighter, then the meter should tell you that it will be recorded brighter (over exposed = above mid-tone).

I think things are being confused a bit... as I understand the question it was in regards to offsetting the exposure in order to optimize image quality/data (i.e. ETTR/ETTL), or not bothering ("mid bar").

I believe ETTL is the same thing as ETTR, the only difference is semantics in relation to whether you are concerned about saving highlights by underexposing the scene (ETTL), or getting the most from shadows by overexposing (ETTR)... but in both cases the technique is to set the exposure/histogram to avoid highlight clipping.
 
One thing I would also suggest is, get to know your sensor. By this I mean understand how far you can go before you lose information. I did this recently with my new camera. I used the built in histogram to push the image as to the right as I could without clipping the highlights. I then increased the exposure by 1, 2 and 3 stops taking shots along the way. I then looked at whether when I pulled the highlights back in LR whether they were still blown or whether it had recorded the detail. I was quite impressed to discover that I could go over by about 2 stops before I lost the details. The reason for doing this is it helps to understand the dynamic range so that you can pull the most out of the shadows without introducing noise.
 
ETTR is a method for optimising dynamic range (not maximising it) and minimising noise. The upside is enhanced shadow detail, but at the expense of blown extreme highlights (that are not regarded as too important). Technically, it's just controlled over-exposure.

ETTL is a new technique only possible with a few of the latest cameras with 'ISO-less' or 'ISO invariant' sensors - some Sony and Nikon cameras, with Canon getting close to it on one or two models, eg 5D4. Basically they allow an extraordinary amount of shadow detail to be pulled up in post processing with no noise penalty - or rather, no more noise than you would have got by raising ISO to an equivalent level. This allows you to under-expose, so that all highlights are retained, and then pull up the shadows in post. This maximises dynamic range, with minimal noise penalty. It can work amazingly well, but is restricted to a small number of high-end cameras and is a long way from becoming mainstream.
 
sk66 said:
I think things are being confused a bit... as I understand the question it was in regards to offsetting the exposure in order to optimize image quality/data (i.e. ETTR/ETTL), or not bothering ("mid bar").

I believe ETTL is the same thing as ETTR, the only difference is semantics in relation to whether you are concerned about saving highlights by underexposing the scene (ETTL), or getting the most from shadows by overexposing (ETTR)... but in both cases the technique is to set the exposure/histogram to avoid highlight clipping.
Click to expand...

You have nailed my question perfectly!
Apologies again, I see how ETTL may have incorrectly steered towards flash.
I'm asking with today's technology, is exposing to the left or right for compensation (shadow/highlight), rather than mid-bar worth it?
 
reheat module said:
You have nailed my question perfectly!
Apologies again, I see how ETTL may have incorrectly steered towards flash.
I'm asking with today's technology, is exposing to the left or right for compensation (shadow/highlight), rather than mid-bar worth it?
Click to expand...
Especially as you posted in the flash section o_O
 
The forum thread is actually entitled 'Lighting and Studio'
To me, this indicates a forum thread for any aspect of lighting which anyone may have an issue with.
The 'and Studio' to me, refers a separate topic of Studio work and not 'lighting within a Studio' only.
It appeared the most prudent place to place my question.
It has nothing to do with flash - although I did apologise twice for any poor phrasing.
 
Last edited:
reheat module said:
The forum thread is actually entitled 'Lighting and Studio'
To me, this indicates a forum thread for any aspect of lighting which anyone may have an issue with.
The 'and Studio' to me, refers a separate topic of Studio work and not 'lighting within a Studio' only.
It appeared the most prudent place to place my question.
It has nothing to do with flash - although I did apologise twice for any poor phrasing.
Click to expand...
Your question isn't about 'lighting' it's about 'exposure'. :)
Not picking on you btw, but if you look at the other threads in the section, you'll get an idea of what 'lighting and studio' is about. The title isn't as ambiguous as you suspected:)

Traditionally 'lighting' issues arose in a studio, but nowadays with portable light sources etc, we take studio techniques anywhere.
So this is the place we discuss all flash, backdrops, meters, gels and other modifiers.
 
Then I stand corrected.
Thank you all for your input.
 
reheat module said:
I'm asking with today's technology, is exposing to the left or right for compensation (shadow/highlight), rather than mid-bar worth it?
Click to expand...
IMO it was never "worth it," except in extreme situations.... and in those cases I believe combining multiple exposures to be a better answer if that is a possibility.
 
sk66 said:
IMO it was never "worth it," except in extreme situations.... and in those cases I believe combining multiple exposures to be a better answer if that is a possibility.
Click to expand...

Disagree. Blending multiple exposures is way more difficult and limited (tripod, static subject).

ISO invariance is what we used to call exposure latitute with film, and it's one of the few things we've lost in the switch to digital. Colour neg film could take huge amounts of over-exposure and still produce very usable results, great for landscapes and shooting against the light. We've come to accept things like blitzed and blown-out skies as a fact of life with digital, and either have to find inconvenient workarounds or just put up with it. But if doesn't have to be like that, I'm all for it. In fact, I will not buy another camera without ISO invariance, or something close to it.
 
Does that mean you'll be switching systems or just waiting for Canon to catch up? ;) :P
 
Nod said:
Does that mean you'll be switching systems or just waiting for Canon to catch up? ;) :p
Click to expand...

I have thought about switching, but not seriously. What I really want is a lighter Canon D850, if you follow, or a 5D Mk4 in a 6D Mk2 body, or I dunno, Canon FF mirrorless?! I'm not in a rush :)

sk66 said:
I did say "if that's a possibility." I wouldn't say that exposure blending is at all difficult in most cases, it's built into LR now.
Click to expand...

Blending is easy, but being limited to static subjects and a tripod is a total non-starter. Since when was wide exposure latitude anything less than highly desirable?
 
HoppyUK said:
Blending is easy, but being limited to static subjects and a tripod is a total non-starter. Since when was wide exposure latitude anything less than highly desirable?
Click to expand...
You can often get away with relatively static subjects handheld using 2 stop auto bracketing steps and a high frame rate... And if it's the sky/BG that's the issue (more typical), comping in a second exposure is usually pretty easy.

I wouldn't say that being ISO invariant is really the same as exposure latitude/DR... at least not in the way it is with film. For one thing, a lot of that latitude/recoverability goes away quickly at ISOs above base. Another thing is that the problem with ISO noise isn't really the amplification, it is the lack of light recorded... underexposing an image isn't going to help that. And then there is the fact that all (?) of the ISO invariant cameras have small pixel/high resolution sensors (probably related somehow) and they all tend to show noise relatively early at 1:1 (from what I know/have seen anyway).

I often shoot things where exposure blending isn't an option and where large exposure latitude would be fantastic... for instance bald eagles in flight in bright light (not ideal, but I'm there). In these situations I'm not able to use base ISO as I'm stopped down (long lenses/TC's/etc) and I need higher SS's to freeze the action. My D810 is supposedly ISO invariant, but I find that if I expose for the bright white head feathers there's very little recoverability in the dark shadows... above ISO 800 (equiv) the image gets noisy very quickly. I can get away with more by oversampling/limiting output size but that kind of defeats the purpose of using that camera... if I have to recover an image hard (either end) it's not likely to be a keeper, regardless of the camera used. To a certain extent, blown out whites are not as much of a problem because that's the way they actually appear. The same is true of dark shadows... they're no worse than noisy shadows. Don't get me wrong, I'll do whatever I can... but I'll often accept some blown out highlights over having to do a lot of shadow recovery.

IME, there just are not that many situations where one can use base ISO and still have enough SS/DOF in order to maximize the ISO invariant characteristic... maybe if I did more work with wide open f/1.x primes I would feel differently, but I don't. I tend to work at the opposite extremes (studio/product/landscape or fast action) where either exposure blending is more often a better option, or ISO invariance is not enough to really make a difference.
 
There's no pleasing some people :D

I'll agree though that often over-extending the dynamic range just looks wrong - like over-cooked HDR. Sometimes subjects just are bright white or very dark, and that's how they should be recorded.
 
HoppyUK said:
There's no pleasing some people :D

I'll agree though that often over-extending the dynamic range just looks wrong - like over-cooked HDR. Sometimes subjects just are bright white or very dark, and that's how they should be recorded.
Click to expand...
Nah, I'm actually pretty happy just "centering the meter" and letting the DR/latitude provide me with a decent amount of protection against messing up in either direction ;). That's probably part of why I said "not worth it."
 
You must log in or register to reply here.
Back
Top