Highlight Weighted Metering - ETTR Without Technical Complications
Nov 20, 2021 19:55:36 #
Grahame wrote:
Perhaps your inflated ego actually makes you think you are teaching something?
Apparently that’s impossible.
Nov 21, 2021 09:21:12 #
a6k wrote:
I don't do video either. But according to at least... (show quote)
Thanks for the link. I'll take a look at the article.
I don't do much video either, but just thinking about the issue it seems to me that the rule of exposing to the right makes sense primarily if you intend to post-process so as to bring up the detail in the dark areas. ETTR allows you to take advantage of the full dynamic range your camera is capable of.
But with video, along with reduced DR there is a significantly reduced editing capability; perhaps not for professional movie studios but for my situation this is the case. In view of this, when shooting video I tend to look for a good compromise exposure somewhere in the middle - maybe I have to accept a bit of over exposure in places if that is what is the only way to make the overall image about right, right out of the camera. I know I will not be able to do much to correct problems in PP so I just try to get it right in the camera - something like the situation used to be with film cameras.
Nov 21, 2021 10:11:39 #
selmslie wrote:
The objective of ETTR is to make the most of the scene's DR without blowing the raw highlights.
For anyone still skeptical about the ability of Highlight Weighted Metering to produce the same ETTR results as zebra warnings, here are the Z7 results for the same two subjects.
Note that the EC setting is significant for a wide DR scene where ETTR might be useful.
ETTR does not benefit narrow DR.
Narrow DR subject, EC+0.7 (brightest pixel about 2.3 stops below the raw limit)
(Download)
Narrow DR JPEG SOOC
(Download)
Wide DR subject, EC+0.3 (only 769 pixels out of 45.7 MP are blown)
(Download)
Wide DR JPEG SOOC
(Download)
Nov 21, 2021 10:33:39 #
selmslie wrote:
.....Note that the EC setting is significant for a wide DR scene where ETTR might be useful.....
Do you need to reduce the EC as the dynamic range of the scene increases?
Nov 21, 2021 11:04:34 #
R.G. wrote:
Do you need to reduce the EC as the dynamic range of the scene increases?
No. You just need to get it right for a wide DR scene.
When the camera sees a narrow DR scene it simply reverts to matrix metering.
Nov 21, 2021 11:19:16 #
selmslie wrote:
No. You just need to get it right for a wide DR scene.....
Would that be +0.3 or +0.7?
Nov 21, 2021 12:39:53 #
Nov 21, 2021 13:55:30 #
R.G. wrote:
Would that be +0.3 or +0.7?
EC+0.7 resulted in 28k blown green raw pixels and 262 blown blue raw pixels and some blinkies.
EC+0.3 resulted in 749 blown green raw pixels and 20 blown blue raw pixels and fewer blinkies.
EC+0.0 resulted in 214 blown green raw pixels and 0 blown blue raw pixels and almost no blinkies.
That suggests that some of the blinkies are caused by specular highlights or that they are starting before the raw values reach the raw limit. This is consistent with what I have found in other tests.
So EC+0.3 is a little safer than +0.7 and that +0.0 may be excessively conservative.
Even if you only consider the blinkies in the captured image on your LCD you can easily see whether they are flashing over small areas that might be specular highlights and that you might not care if they end up as pure white in the final image.
Nov 21, 2021 13:56:35 #
R.G. wrote:
Would that be +0.3 or +0.7?
It's apparent now from the examples that it varies dependent upon how many blown pixels the user is going to deem acceptable even in images without any single point light sources.
Nov 21, 2021 13:58:30 #
BebuLamar wrote:
I have the Df and it doesn't do this. Too bad!
The Df is eight years old. It preceded the introduction of Highlight Weighted Metering.
But the blinkies behave the same as the newer Nikons.
Nov 21, 2021 14:10:55 #
Grahame wrote:
It's apparent now from the examples that it varies dependent upon how many blown pixels the user is going to deem acceptable even in images without any single point light sources.
That is exactly the point I was making in my earlier post: https://www.uglyhedgehog.com/t-719275-2.html#12718085 when I stated, "Now consider the number of pixels it takes to record a specular highlight. Here is a 20x18 snip of some of the fountain highlights in the second image.
That's only 380 pixels but it contains about a dozen specular highlights from water droplets."It's not just specular highlights that you might no be concerned about blowing. You might not be too concerned about blowing a tiny handful of pixels that don't really harm the image.
Nov 21, 2021 14:15:21 #
selmslie wrote:
.....So EC+0.3 is a little safer than +0.7 and that +0.0 may be excessively conservative.....
So returning to the suggestion of the thread's title, the most uncomplicated, most worry-free option might be to set EC to +0.3 and forget about it. An extra 1/3 of a stop doesn't seem worth risking potentially troublesome blown highlights.
Having to take a test shot then review it for blinkies isn't prohibitively bothersome but then again it isn't the least complicated option, and it may not be a practical option in some situations.
Perhaps through time one would learn to spot the situations that were potentially troublesome and then lower the EC accordingly while keeping it higher for the safer options. However, I know from experience that I don't always spot the situations that need EC, despite being familiar with the warning signs. I'm all for the safe, simple option
.Nov 21, 2021 14:23:25 #
R.G. wrote:
So returning to the suggestion of the thread's title, the most uncomplicated, most worry-free option might be to set EC to +0.3 and forget about it. An extra 1/3 of a stop doesn't seem worth risking potentially troublesome blown highlights.
That's where I leave mine set now. But I still watch for blinkies if I'm not sure sure.
There is no absolutely fool-proof way to use ETTR.
Yes, a 1/3 stop change is insignificant.
Nov 21, 2021 14:30:05 #
selmslie wrote:
That's where I leave mine set now. But I still watch for blinkies if I'm not sure sure.
There is no absolutely fool-proof way to use ETTR.
Yes, a 1/3 stop change is insignificant.
There is no absolutely fool-proof way to use ETTR.
Yes, a 1/3 stop change is insignificant.
I'll be giving it a try. It sounds like my kind of metering
.
If you want to reply, then register here. Registration is free and your account is created instantly, so you can post right away.