Interpreting Raw Data From Studio Images
October 2011, Carl Garrard
Adobe Photoshop Elements 9 (Win/Mac) [OLD VERSION]
Some of you may be considering one of the new Sony cameras using the 24mp sensor, and some of you have either seen or processed some of the available raw images. While others may not know what to think of all the back and forth comparisons and "warm" topic threads out there on the internet.
I started this thread to help those who may be confused by all this data and debates out there. I have just a few quick and simple things to say, truths that hold water you can count on. Besides that, I have experience using the new sensor on 3 different cameras and have used them in non studio settings that produce not only a more extreme range of light, but very uneven light as well. I've used them at all ISO's and evaluated as such.
Bottom line, real life images give you a much more realistic interpretation of a cameras' raw image quality. Studio images are great for cross comparing cameras under next to ideal conditions. They will give you a basic idea of how the camera performs against another in a relatively even setting, but only part of the story is really being told. This is really important to remember when evaluating image quality.
Most cameras today, most DSLRS in this case- will appear to perform better in evenly lit studio scenes than they will in real life. This is a fact and why some readers not only prefer me to do real life samples, but ask for them. The more real life raw samples you have to evaluate exposed at different times of the day in varying lighting conditions- the better. This is fact.
Therefore I want to show you a sample of just what I'm talking about. Here we have a studio scene using my custom built test chart and a US 5 dollar bill. Plenty of details to show, varying shadow light and dark tones, and fine colors. This image is taken with a camera that is not known to be a very good high ISO performer- the Pentax K20D. But to prove my point about studio lighting and raw interpretation, I ramped up the K20D to its highest extended sensitivity level- ISO 6400.
Generally speaking the K20D performs very good till ISO 1600 then starts to show signs of stress in the images from there on. Most users don't claim the K20D will perform acceptably above ISO 2,5
But here we have an ACR converted ISO 6400 image. This is a full size image that I developed at about 17.5mp in ACR (one step higher in output from ACR than its native 14.6mp resolution). This means it will show more noise than its native resolution- but have a look here and tell me what you think.
Not bad, is it? So, what's going on here? Why does the K20D suddenly look like a good performer at an ISO it's known to not be so good for?
Three reasons- even lighting, proper exposure (making sure my histogram was near perfect for a test subject, difficult to do in real life circumstances), and a test subject with high contrast outlines (this one is key). There aren't many shadows to deal with in this scene at all, I made sure to optimize my exposure and process the file with the utmost care, and the hard edged contrast lines hold up well under heavy noise and noise reduction (in this case only chroma noise reduction was applied).
Now, to even things out, lets take a look at a low light real life scene shot at ISO 6400 and evaluate the extremes between that and studio lighting. Not so good, right? I even took great care to be sure my exposure was perfect for this shot (preserving as many highlights and shadows as possible).
So what gives? Well that's the nature of studio lighting. My warning to you is to not base a sole opinion on image quality from studio lighting alone- always request to see real life images with as many lighting conditions as possible, and ask for raw files so you can tinker and play with them yourself to get an even more accurate analysis of a camera's capabilities.
To some of you who already know everything I've said, obviously this is repetitive and rhetorical. Yet there are those out there that might find this quick reminder helpful and refreshing- hope it was worth your while.
Carl
October 2011, Carl Garrard
Adobe Photoshop Elements 9 (Win/Mac) [OLD VERSION]
Some of you may be considering one of the new Sony cameras using the 24mp sensor, and some of you have either seen or processed some of the available raw images. While others may not know what to think of all the back and forth comparisons and "warm" topic threads out there on the internet.
I started this thread to help those who may be confused by all this data and debates out there. I have just a few quick and simple things to say, truths that hold water you can count on. Besides that, I have experience using the new sensor on 3 different cameras and have used them in non studio settings that produce not only a more extreme range of light, but very uneven light as well. I've used them at all ISO's and evaluated as such.
Bottom line, real life images give you a much more realistic interpretation of a cameras' raw image quality. Studio images are great for cross comparing cameras under next to ideal conditions. They will give you a basic idea of how the camera performs against another in a relatively even setting, but only part of the story is really being told. This is really important to remember when evaluating image quality.
Most cameras today, most DSLRS in this case- will appear to perform better in evenly lit studio scenes than they will in real life. This is a fact and why some readers not only prefer me to do real life samples, but ask for them. The more real life raw samples you have to evaluate exposed at different times of the day in varying lighting conditions- the better. This is fact.
Therefore I want to show you a sample of just what I'm talking about. Here we have a studio scene using my custom built test chart and a US 5 dollar bill. Plenty of details to show, varying shadow light and dark tones, and fine colors. This image is taken with a camera that is not known to be a very good high ISO performer- the Pentax K20D. But to prove my point about studio lighting and raw interpretation, I ramped up the K20D to its highest extended sensitivity level- ISO 6400.
Generally speaking the K20D performs very good till ISO 1600 then starts to show signs of stress in the images from there on. Most users don't claim the K20D will perform acceptably above ISO 2,5
But here we have an ACR converted ISO 6400 image. This is a full size image that I developed at about 17.5mp in ACR (one step higher in output from ACR than its native 14.6mp resolution). This means it will show more noise than its native resolution- but have a look here and tell me what you think.
 |
| Click in the image for a larger size |
Not bad, is it? So, what's going on here? Why does the K20D suddenly look like a good performer at an ISO it's known to not be so good for?
Three reasons- even lighting, proper exposure (making sure my histogram was near perfect for a test subject, difficult to do in real life circumstances), and a test subject with high contrast outlines (this one is key). There aren't many shadows to deal with in this scene at all, I made sure to optimize my exposure and process the file with the utmost care, and the hard edged contrast lines hold up well under heavy noise and noise reduction (in this case only chroma noise reduction was applied).
Now, to even things out, lets take a look at a low light real life scene shot at ISO 6400 and evaluate the extremes between that and studio lighting. Not so good, right? I even took great care to be sure my exposure was perfect for this shot (preserving as many highlights and shadows as possible).
 |
| Click in the image for a larger size |
So what gives? Well that's the nature of studio lighting. My warning to you is to not base a sole opinion on image quality from studio lighting alone- always request to see real life images with as many lighting conditions as possible, and ask for raw files so you can tinker and play with them yourself to get an even more accurate analysis of a camera's capabilities.
To some of you who already know everything I've said, obviously this is repetitive and rhetorical. Yet there are those out there that might find this quick reminder helpful and refreshing- hope it was worth your while.
Carl
Nice read. Just goez to show you that even a camera not so good at hi ISO's can look good in a controlled lighting circumstance- thanks for the read- Jess
ReplyDeleteI agree with you, Carl. Just got my A77 and at higher ISO, the raw files processed in lightroom is ok. If I don't pixel peel, even iso 6400 looks good on my Dell 27".
ReplyDeletebig difference here in how light changes the look of the output, thanks so much for pointing this out in an article it often goes overlooked by reviewers
ReplyDeleteCarl, are you aware of the "K20 magenta border remover" tool made by GordonBGood?
ReplyDeleteIt will take care of the ugly magenta color cast you get on the bottom edge of your pics at such high ISOs...
https://skydrive.live.com/?cid=610bb9d47b109171&id=610BB9D47B109171%21122