DPP - subtle change

Hi Folks

I'm using Canon DPP4 to process CR2 files from my 6D, and CR3 files from a 90D.

I've noticed a strange thing when I open up an image thumbnail to look at it at "fit to screen" size for application of adjustments.

A second or two after it opens at the larger size, I notice a subtle change in the sharpness of the image take place. It's almost as if a subtle noise reduction setting is being added. I've checked in the tools menu though and there are no adjustments set in any of the tabs. Here's a pic showing the difference... the left hand side of the image shows the initial state, and then the right shows the slight loss of detail.

Does anyone know what's happening here? (you may need to view the pic at full size as it's quite subtle!)

I'm only guessing, but maybe the program is opening the embedded JPEG followed by the actual RAW image? Depending on how much the JPEG is compressed, you may see a difference in the image after the RAW is opened.

DPP never shows the raw data, it is always applying the filters on the screen, and initially they are populated with the values from the camera (picture style settings, high ISO NR values, etc).

My theory is that you initially see the image with all the JPG settings applied, but then it takes just a bit of time for the NR values to run through the algorithm to finally render the JPG (or rerender). The slower the processing of your PC/laptop, the more noticeable this "blip" is. It almost like the anomaly Neo would see in the matrix from time to time.

Also, we know that even if you "disable" the high ISO NR in camera or in DPP, it still applies some form of NR on its own. I am pretty sure we have seen in the past where a person took the same raw, and set disabled for high ISO NR and compared to output from another raw processor that truly didn't apply NR, and there was a difference.

TeamSpeed wrote in post #19441008
DPP never shows the raw data, it is always applying the filters on the screen, and initially they are populated with the values from the camera (picture style settings, high ISO NR values, etc).

My theory is that you initially see the image with all the JPG settings applied, but then it takes just a bit of time for the NR values to run through the algorithm to finally render the JPG (or rerender). The slower the processing of your PC/laptop, the more noticeable this "blip" is. It almost like the anomaly Neo would see in the matrix from time to time.

Also, we know that even if you "disable" the high ISO NR in camera or in DPP, it still applies some form of NR on its own. I am pretty sure we have seen in the past where a person took the same raw, and set disabled for high ISO NR and compared to output from another raw processor that truly didn't apply NR, and there was a difference.

We are probably splitting hairs, but DPP is a RAW viewer (see https://www.usa.canon.​com …hoto-professional-express ).
I could agree that one isn't seeing the data as stored on the camera card, but seeing the RAW file on the computer after application of the filters as you describe, and these take time to render.

You're looking at 2 different renderings of your raw data.

A raw file is just data, a ledger containing a set of numbers for each pixel in the frame, assigned when the shutter closed. Software is required to interpret those numbers and render an image that can be displayed on a screen. Let's call that rendering a jpg.

There is software in your camera that writes a jpg file as soon as the data is collected. It uses that file to display the image on the back of your camera, and it also embeds that jpg in the raw file. You can use camera settings to manipulate the way the camera interprets the data, but once written, it's only one interpretation.

The embedded jpg is tiny and fast and standardized so it can be read by a wide variety of software.

DPP is different software that renders raw data into jpgs. When you first open an image, it displays the embedded jpg to get you started viewing quickly. But if you start any action where DPP really gets to work, it's going to render its own jpg. That's that little delay you get when switching to the big size.

So you're literally looking at 2 different jpg renderings. I don't have any insight on the specific differences you mentioned other than to say, there will always be differences.

I don't use the embedded jpg for anything beyond a rough first cull. And even then, DPP will render a card full of raw files in a few seconds. I just let DPP do it in a batch. Once it's done, it's done. The embedded jpg is now relegated to Windows Explorer.

drsilver wrote in post #19441016
You're looking at 2 different renderings of your raw data.

A raw file is just data, a ledger containing a set of numbers for each pixel in the frame, assigned when the shutter closed. Software is required to interpret those numbers and render an image that can be displayed on a screen. Let's call that rendering a jpg.

There is software in your camera that writes a jpg file as soon as the data is collected. It uses that file to display the image on the back of your camera, and it also embeds that jpg in the raw file. You can use camera settings to manipulate the way the camera interprets the data, but once written, it's only one interpretation.

The embedded jpg is tiny and fast and standardized so it can be read by a wide variety of software.

DPP is different software that renders raw data into jpgs. When you first open an image, it displays the embedded jpg to get you started viewing quickly. But if you start any action where DPP really gets to work, it's going to render its own jpg. That's that little delay you get when switching to the big size.

So you're literally looking at 2 different jpg renderings. I don't have any insight on the specific differences you mentioned other than to say, there will always be differences.

I don't use the embedded jpg for anything beyond a rough first cull. And even then, DPP will render a card full of raw files in a few seconds. I just let DPP do it in a batch. Once it's done, it's done. The embedded jpg is now relegated to Windows Explorer.

How do you know it is being rendered as a second JPEG? In Windows, one can write data to a bitmap object (different from a .BMP file!). This allows you to display something that isn't a bitmap, such as a RAW file. See: https://learn.microsof​t.com …itmap-class?view=msvc-170
DPP (and other software, such as Adobe) likely reads the RAW and renders it into a device independent bitmap class.

This is more detail than needed to answer the question, as we agree that it shows an image rendered quickly, followed by another image that needed more time to open or render due to applied filters.

Capn Jack wrote in post #19441020
How do you know it is being rendered as a second JPEG? In Windows, one can write data to a bitmap object (different from a .BMP file!). This allows you to display something that isn't a bitmap, such as a RAW file. See: https://learn.microsof​t.com …itmap-class?view=msvc-170
DPP (and other software, such as Adobe) likely reads the RAW and renders it into a device independent bitmap class.

This is more detail than needed to answer the question, as we agree that it shows an image rendered quickly, followed by another image that needed more time to open or render due to applied filters.

Yes, all true. Image data can exist in many forms before (or even if) it's ever codified in the jpg format as a file you can post on PTON. For the sake of simplicity, I chose to call the rendered data, in whatever state it's in during the process, a jpg.

clippo wrote in post #19440963
Hi Folks

I'm using Canon DPP4 to process CR2 files from my 6D, and CR3 files from a 90D.

I've noticed a strange thing when I open up an image thumbnail to look at it at "fit to screen" size for application of adjustments.

A second or two after it opens at the larger size, I notice a subtle change in the sharpness of the image take place. It's almost as if a subtle noise reduction setting is being added. I've checked in the tools menu though and there are no adjustments set in any of the tabs. Here's a pic showing the difference... the left hand side of the image shows the initial state, and then the right shows the slight loss of detail.

Does anyone know what's happening here? (you may need to view the pic at full size as it's quite subtle!)

Hosted photo: posted by clippo in
./showthread.php?p=194​40963&i=i184697481
forum: RAW, Post Processing & Printing

DPP has been behaving that way for a long time.

When I use DPP I see a quite blurry version of the image followed by a properly sharp image. That would be consistent with a low-res thumbnail being replaced by a new image freshly generated from the RAW.

However, the OP claims a LOSS of detail after the update. Sounds like something different to me.

AntonLargiader wrote in post #19442624
However, the OP claims a LOSS of detail after the update. Sounds like something different to me.

I been thinking about this. I bet the jpg rendering engine in the camera is the best that Canon makes. A lot of folks just shoot jpg and only write those files to the card. A jpg file is the final product and Canon's reputation rests on its looking as good as it can. I'm sure it adds noise reduction and sharpening and all those things. The embedded jpg in a raw file is produced by that same jpg engine, albeit in a very-low-resolution form.

With DPP, on the other hand, Canon assumes you came to that app with the idea of making adjustments to the captured image yourself. The initial jpg rendering is meant to show general exposure and color information on the file, as shot, based on the histogram. The rest it leaves up to the user.

Maybe.

drsilver wrote in post #19445618
I been thinking about this. I bet the jpg rendering engine in the camera is the best that Canon makes. A lot of folks just shoot jpg and only write those files to the card. A jpg file is the final product and Canon's reputation rests on its looking as good as it can. I'm sure it adds noise reduction and sharpening and all those things. The embedded jpg in a raw file is produced by that same jpg engine, albeit in a very-low-resolution form.

With DPP, on the other hand, Canon assumes you came to that app with the idea of making adjustments to the captured image yourself. The initial jpg rendering is meant to show general exposure and color information on the file, as shot, based on the histogram. The rest it leaves up to the user.

Maybe.

I'm thinking it has more to do with how your setting in camera are set. I have always also seen this but in reverse, the first image is of significantly poorer quality than in a second, it sharpens up. I'm speculating that this is due to my camera settings being set at the lowest quality jpeg. Then when it finishes reading in the raw and doing it's magic, I get a full size jpeg with most of the in camera settings but I'm thinking, not all. BTW, I only see change if I am viewing the image straight from the card. If I download the images first, it happens so fast that I cannot see when it switches from the embedded thumbnail to the pre-processed image.