ACDSee is a photo management/editing tool. In its "Pro" version it also supports RAW file editing. Despite all its good features it has one (in my eyes) major problem: its Jpeg engine struggles with chroma subsampling.
I´ve been using ACDSee Pro for quite a few years now. I´ve always liked its speed and ease of use. I started tagging my photos with it long ago and now I have quite a lot of information in its database. This is one reason why I do want to continue using it - switching to another program and losing all this information would be a pain, since I invested many many hours in tagging the images. However, there are some reasons why I still think about switching. One of them is ACDSee´s problem with chroma subsampling. For those who don´t know what that is I try to explain it quickly (all the others can skip the following paragraph).
What is chroma subsampling?
Chroma subsampling is a technique that is used for image compression. It exploits the fact that the human visual system is more sensitive to changes in brightness than to changes in color. In other words, the human eye has greater resolving power for black and white than for colors. Well, how can this be used for image compression? First, an image must be converted from the RGB color space into one where brightness and color information is separated, e.g. YUV. This can be done without any loss of quality. Now the color channels can be subsampled, which (simplyfied) means that only every second sample value is kept and every other is ignored. This task is called chroma subsampling. This leads to a compression ratio of 3:2.
When displaying such an image, the dropped values need to be interpolated somehow (you need color information at every sample position). When done properly the difference between the original and the subsampled image can hardly be seen. This is why this technique is standard in the Jpeg compression format.
There is one word in the previous paragraph which has to be pronounced strongly: "When done properly the difference between the original and the subsampled image can hardly be seen". Unfortunately in real life chroma subsampling is not just skipping every other chroma value. Before skipping these values the image has to be filtered in a certain way. And of course when preparing such an image for display, it is not sufficient to duplicate every chroma value. In order to achive good image quality a good interpolation algorithm has to be used. And this is what ACDSee developers obviously refused to do. It seems that they implemented the easiest method for chroma subsampling that they could find. This might be good to reduce processing time but it is bad for image quality. It brings up artifacts which are most visible at the edges of red objects. As said, chroma subsampling is used in almost every Jpeg image so chances are good that ACDSee will display all your Jpegs with such artifacts.
See the difference and decide for yourself...
Samples
Below you will find some samples of artifacts that ACDSee produces. To create these samples I opened an out-of-camera Jpeg file in ACDSee Pro 7 (Build 137) and saved it in PNG file format (using 'save as...' in ACDSee). PNG only applies lossless compression to an image, it does not apply chroma subsampling. So in fact the PNG represents what you see on the screen when you view a Jpeg in ACDSee. The other samples were produced in the same way, only I didn´t use ACDSee but paint (you know, the program that comes with windows). I´ve tested some other programs as well (Windows Image Viewer, Photoshop, Zoner Photo Studio, Picasa, Internet Explorer, Matlab...) but ACDSee is the only program that shows these artifacts. I simply used Paint in this test because it was easy to convert the JPG into a PNG with it. Finally I cropped the same area out of both PNG files and saved them again as PNG. The images below are enlarged via HTML code. You should be able to download them and view them at 100%. For those who have doubts: I´ve tested this with Jpegs taken with a Canon 60D, a Canon 6D, with Jpegs from camera tests on dpreview.com and with quite a lot other Jpegs with unknown origin (meaning I don´t know how they were processed). All show the same problems. The only way to get rid of these artifacts is to capture raw files and make sure to save Jpegs without chroma subsampling (which not all programs allow you to do).See the difference and decide for yourself...
crop of a Jpeg image as displayed in ACDSee Pro 7 |
crop of a Jpeg image as displayed in Paint |
crop of a Jpeg image as displayed in ACDSee Pro 7 |
crop of a Jpeg image as displayed in ACDSee Pro 7 |
crop of a Jpeg image as displayed in Paint |
The above samples were taken by myself. I could have used other ones, e.g. from dpreview.com but I don´t want to risk a law suit. If you want to reproduce the issue on your own, just grab your camera, switch it to Jpeg and shoot something red.
Well, if you now think that this is something you can live with, as it is only an issue of displaying the image, you should possible spend a thought on image editing. ACDSee is not only an image viewer, but also an image editing program. And it uses the same Jpeg decoding algorithm not only in display mode but also in develop/editing mode. And, even worse, it introduces the same artifacts when saving Jpeg images. So, if you decide to edit a Jpeg image in ACDSee and save it after editing, you have in fact stored the artifacts into the file. If you now open the image in another program you will see the same artifacts and there is no way back. You will never get these out of your images again. You don´t edit your images? What about rotating? Once you rotate an image from landscape to portrait, it will be saved as a new Jpeg file with all the mess in it.
If you are an ACDSee user and want them to fix this issue, contact their support or reply to the thread in their forum.
If you want to find out more about chroma subsampling artifacts visit Glenn Chan´s page.
If you are an ACDSee user and want them to fix this issue, contact their support or reply to the thread in their forum.
If you want to find out more about chroma subsampling artifacts visit Glenn Chan´s page.
No comments:
Post a Comment