Re-assessing the D800 sensor as a Canon alternative
/forum/topic/1157003/2

1       2      
3
       end

snapsy
Registered: Feb 24, 2008
Total Posts: 4538
Country: United States

mmurph wrote:
I don't want to enter into these wars, it is not worth my energy usually.

But here is some info published quite recently on why RAW files are not truly linear. Including input from Eric Chan, of Adobe:

http://theonlinephotographer.typepad.com/the_online_photographer/2012/10/raw-is-not-raw.html


That article only talks about how different cameras set the gray point differently and how the raw processor has to adjust accordingly.. It doesn't affect the raw data nor its linearity.



skibum5
Registered: Jan 21, 2005
Total Posts: 16180
Country: United States

Monito wrote:
Exposing to the Right doesn't change contrast. It takes advantage of the full dynamic range available to raise as many tones as possible up out of the noise floor.


+1


I don't think skibum5 owns a D800e, but I could be mistaken.


I don't.



skibum5
Registered: Jan 21, 2005
Total Posts: 16180
Country: United States

mmurph wrote:
I don't want to enter into these wars, it is not worth my energy usually.

But here is some info published quite recently on why RAW files are not truly linear. Including input from Eric Chan, of Adobe:

http://theonlinephotographer.typepad.com/the_online_photographer/2012/10/raw-is-not-raw.html


Nah, that just talks about how different programs decide to interpret RAW or how different makers suggest a different mid-point choice.

Note the comment:

"Mike Shimwell: "Good article Ctein, but I would take some issue with your title. :-) The issue you address is the interpretation of the RAW file (as you state throughout) not the 'rawness' of the data in the file."



Monito
Registered: Jan 28, 2005
Total Posts: 10079
Country: Canada

mttran wrote: Monitor, look at this sample here: http://www.fredmiranda.com/forum/topic/1155718/0#11029322 and tell us your masking (issue) processes still work in these cases.

There is no monitor in this thread.

However, if you understood EttR, which is what you were responding to, you would know there is no masking with EttR.

Further, underexposure is the opposite of EttR.

In the case of the example, better exposure would go a long way.




splathrop
Registered: Feb 27, 2006
Total Posts: 527
Country: United States

Maybe all you smart guys can help me with my reasoning. I'm fine with clipping some shadows some of the time. Of course it depends on the image. But no clipped shadows ever is, to my eye, unnatural.

And I thought I understood this about contrast: if you have recorded it, and want to make it more apparent, that means you are going to have to spread some values in the regions where you want to see more. In turn, that means one of three things. 1) You might spread the values in one part of the continuum and close them in another, resulting in lost detail there, or 2) You might spread the values where you want to see more, and move the others away from there, towards the shadows or highlights, potentially moving the ones at the end of the line into clipped shadows or clipped highlights, thus losing them altogether, but not discarding other detail along the way, or 3) Your image might have some empty spots along the way that the moved values could fill into, and that might let you increase your contrast where you want to, and not lose anything else as a result. That third option is okay, but seems to presume a sub-optimal original exposure.

With that for context, I will leave it to your smartness to grasp, presumably instantly, why that context might have implications for how you use the D800/D800e, and why there might be reasons, besides minimizing noise, why you might want to expose to the right.



mttran
Registered: Nov 03, 2005
Total Posts: 6645
Country: United States

Monito wrote:
mttran wrote: Monitor, look at this sample here: http://www.fredmiranda.com/forum/topic/1155718/0#11029322 and tell us your masking (issue) processes still work in these cases.

There is no monitor in this thread.

However, if you understood EttR, which is what you were responding to, you would know there is no masking with EttR.

Further, underexposure is the opposite of EttR.

In the case of the example, better exposure would go a long way.




Who gives a rat about anythingelse. When you need more room in DR, you do need better DR cam: http://www.fredmiranda.com/forum/topic/1116053/16#11028037



Monito
Registered: Jan 28, 2005
Total Posts: 10079
Country: Canada

mttran wrote: Who gives a rat about anythingelse. When you need more room in DR, you do need better DR cam: http://www.fredmiranda.com/forum/topic/1116053/16#11028037

The applicable phrase, that I wrote, is "given whatever camera you have in your hand". That does not mean the camera on the shelf at home or the shelf at the store or the shelf at the shipper or the shelf at the repair shop.

The concept is: that "whatever camera you have in your hand", it will have a given dynamic range and the photograph you want to make may or may not exceed the dynamic range of that camera and that you have to adjust your technique appropriately.

It doesn't matter whether the camera in hand is the SuperDooperFlex 2019 model or a Sony Mavica. It doesn't matter if the camera on the "shelf" is a Nikon or Canon or Holga or Speed Graphic, it won't help you. You have to know how the film/sensor in your hand behaves and how to read or meter the scene and how to evaluate it against your pre-visualized photo.

Now if any part of this is not clear, point it out and I will help you understand it.

Monito wrote: First thing to understand is that if you wish to get the most noise free images, given whatever camera you have in your hand at the time, there are at least seven ways to obtain that.



mttran
Registered: Nov 03, 2005
Total Posts: 6645
Country: United States

Ettr is a process to work around camera DR issues. It always works against your final IQ. It's your last resource if you don't have anything else better on your hands at that time.



Monito
Registered: Jan 28, 2005
Total Posts: 10079
Country: Canada

mttran wrote: Ettr is a process to work around camera DR issues. It always works against your final IQ.

EttR is one of many processes to use because all cameras have DR issues sometimes for some scenes. EttR always improves your final IQ whenever it is applicable.



akclimber
Registered: Aug 01, 2002
Total Posts: 2700
Country: United States

splathrop wrote:
akclimber, another question if I may. Have you made any systematized attempt to compare results in your D800e files with different exposure approaches? Given your last post, I suppose maybe not. But I would be intrigued to know what the overall effect on color and contrast is if you use a D800e more or less as if it had less dynamic range than it does, and try hard to expose to the right. I'm guessing that doing that would take some of the flatness out of the images, or at least facilitates post processing to accomplish that.


Oh, I still expose to the right, just like my other cams. The difference is that with the D800 I can recover much more shadow information if needed. The practical difference between the sensors is that with the d800 sensor, at ISO 100 or 200 I really don't have to sweat underexposing shadows in high DR images - I know I'll be able to recover lots of clean detail, unlike my Canon sensors. The only different exposure approaches I've tested are the ISO tests vs DR and noise (just RAW conversion eyeball type tests).

Cheers!
Joe



snapsy
Registered: Feb 24, 2008
Total Posts: 4538
Country: United States

splathrop wrote:
With that for context, I will leave it to your smartness to grasp, presumably instantly, why that context might have implications for how you use the D800/D800e, and why there might be reasons, besides minimizing noise, why you might want to expose to the right.


I ETTR specifically to provide the maximum flexibility in post, including for global contrast adjustments (tone mapping) you described in your post. Provided the original scene's DR allows it, I like to have all my capture tones compressed in the upper half or even upper 1/3 of the histogram when possible, which allows me to "fill the gaps" as you described in point #3 when performing global contrast adjustments via the black level. Of course there is also local contrast adjustments, which I achieve via large-radius USM.



mttran
Registered: Nov 03, 2005
Total Posts: 6645
Country: United States

akclimber wrote:
The difference is that with the D800 I can recover much more shadow information if needed. The practical difference between the sensors is that with the d800 sensor, at ISO 100 or 200 I really don't have to sweat underexposing shadows in high DR images - I know I'll be able to recover lots of clean detail, unlike my Canon sensors. Joe


+1,

Sorry Monito, Ettr can't save your Queen...which happen i have two of them in my hands don't get me wrong, i like your Queen but just can't stand people who tries hidding it's weakness



Monito
Registered: Jan 28, 2005
Total Posts: 10079
Country: Canada

mttran wrote: Sorry Monito, Ettr can't save your Queen...which happen i have two of them in my hands

There is more to life and photography than nonsensical one-liners. Nonsensical one-liners do not erase or negative or supersede informative posts.

But don't let that stop you from revealing yourself.



mttran
Registered: Nov 03, 2005
Total Posts: 6645
Country: United States

This is an informative forum, people loves the real stuffs here...so be real would you.



skibum5
Registered: Jan 21, 2005
Total Posts: 16180
Country: United States

splathrop wrote:
Maybe all you smart guys can help me with my reasoning. I'm fine with clipping some shadows some of the time. Of course it depends on the image. But no clipped shadows ever is, to my eye, unnatural.

And I thought I understood this about contrast: if you have recorded it, and want to make it more apparent, that means you are going to have to spread some values in the regions where you want to see more. In turn, that means one of three things. 1) You might spread the values in one part of the continuum and close them in another, resulting in lost detail there, or 2) You might spread the values where you want to see more, and move the others away from there, towards the shadows or highlights, potentially moving the ones at the end of the line into clipped shadows or clipped highlights, thus losing them altogether, but not discarding other detail along the way, or 3) Your image might have some empty spots along the way that the moved values could fill into, and that might let you increase your contrast where you want to, and not lose anything else as a result. That third option is okay, but seems to presume a sub-optimal original exposure.

With that for context, I will leave it to your smartness to grasp, presumably instantly, why that context might have implications for how you use the D800/D800e, and why there might be reasons, besides minimizing noise, why you might want to expose to the right.



1. That doesn't make the D800 sensor worse since you could always just develop as you would with a 5D3 and if that is all you ever want then maybe its not worth it for you although....

2. If you do more careful tone mapping you can save shadows and retain better contrast, many people dont bother to flip the middle contrast around, which is quick and easy to do in just seconds, or to do trickier multi-part tone mapping where you spread out various areas and retain nicer contrast at many different levels. If you do that perhaps you would get something out of it AND like it. If you still don't like that then yeah the 5D3 sensor is plenty good enough and since it has better liveview and more FF fps and such it's probably the better cam for you (unless you need reach from the D800 MP count). Or if you shoot scenes that never need it or where it makes it look worse 99% of the time 5D3 all the way too.



skibum5
Registered: Jan 21, 2005
Total Posts: 16180
Country: United States

mttran wrote:
Ettr is a process to work around camera DR issues. It always works against your final IQ. It's your last resource if you don't have anything else better on your hands at that time.


How does it work against final IQ? The only tricky thing is that the default profiles in some converters are twisted so it can make it trickier to get natural colors perhaps, but with a custom profile that won't matter. And well yeah it can be trickier to tone curve with an extreme ETTR scene but if you take the time it won't hurt final IQ at all, it will make it better, on any camera.



mttran
Registered: Nov 03, 2005
Total Posts: 6645
Country: United States

skibum5 wrote:
mttran wrote:
Ettr is a process to work around camera DR issues. It always works against your final IQ. It's your last resource if you don't have anything else better on your hands at that time.


How does it work against final IQ? The only tricky thing is that the default profiles in some converters are twisted so it can make it trickier to get natural colors perhaps, but with a custom profile that won't matter. And well yeah it can be trickier to tone curve with an extreme ETTR scene but if you take the time it won't hurt final IQ at all, it will make it better, on any camera.


I found, a hard ETTRed one with pushing shadows does not have the silky/smoothness look compared to the normal/under exposed one. Data seem to be distorted with more system noise since the pushing is too close to the h/w limits? For sure, it will work better with EXMOR sensor.



skibum5
Registered: Jan 21, 2005
Total Posts: 16180
Country: United States

mttran wrote:
skibum5 wrote:
mttran wrote:
Ettr is a process to work around camera DR issues. It always works against your final IQ. It's your last resource if you don't have anything else better on your hands at that time.


How does it work against final IQ? The only tricky thing is that the default profiles in some converters are twisted so it can make it trickier to get natural colors perhaps, but with a custom profile that won't matter. And well yeah it can be trickier to tone curve with an extreme ETTR scene but if you take the time it won't hurt final IQ at all, it will make it better, on any camera.


I found, a hard ETTRed one with pushing shadows does not have the silky/smoothness look compared to the normal/under exposed one. Data seem to be distorted with more system noise since the pushing is too close to the h/w limits? For sure, it will work better with EXMOR sensor.


I don't think it it is ETTR doing anything bad or revealing anything more, all ETTR means is you expose longer than needed to place the scenes middle gray higher so long as you have enough highlight room relative to the scene relative gray to move to the right and you should get less system noise, that is why you do it.

Yeah Exmor delivers cleaner shadows at low ISO, not sure what that has to do with ETTR though. ETTR would be more critical on non-exmor since it helps you avoid the system noise.



1       2      
3
       end