Photoshop Gurus Forum
Welcome to Photoshop Gurus forum. Register a free account today to become a member! It's completely free. Once signed in, you'll enjoy an ad-free experience and be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!
PC/Mac different colors
- Thread starter chraqek
- Start date
Why?
MrToM
Guru
- Messages
- 3,595
- Likes
- 3,322
Something doesn't add up here...
kolor2.png is riddled with jpeg artefacts despite containing nearly 5000 colours and being 41KB in size.
kolor.png is much cleaner, contains only 518 colours and is only 8.8KB.
So are these examples screen grabs, or conversions of the actual image file?
The difference could be attributed to saving a JPEG compressed file at different 'quality' values which would induce the artefacts visible in kolor2.png.
Saving in the PNG format would never introduce such artefacts so I'm confused as to exactly what the example images are and where they came from.
If these really are genuine PNG files saved out from the original image then why would only one be riddled with artefacts? Its as if they are two completely different images.
What's the true story?
Regards.
MrToM.
kolor2.png is riddled with jpeg artefacts despite containing nearly 5000 colours and being 41KB in size.
kolor.png is much cleaner, contains only 518 colours and is only 8.8KB.
So are these examples screen grabs, or conversions of the actual image file?
The difference could be attributed to saving a JPEG compressed file at different 'quality' values which would induce the artefacts visible in kolor2.png.
Saving in the PNG format would never introduce such artefacts so I'm confused as to exactly what the example images are and where they came from.
If these really are genuine PNG files saved out from the original image then why would only one be riddled with artefacts? Its as if they are two completely different images.
What's the true story?
Regards.
MrToM.
thebestcpu
Guru
- Messages
- 3,117
- Likes
- 2,873
Four distinct possibilities
1)PS color management policies not the same when reading the image
2) PS advanced color settings turned and and not the same between the two
3 Default rendering intents are not the same in both
4 ) human error
1)PS color management policies not the same when reading the image
2) PS advanced color settings turned and and not the same between the two
3 Default rendering intents are not the same in both
4 ) human error
thebestcpu
Guru
- Messages
- 3,117
- Likes
- 2,873
OK, I narrowed down the issue to #4 above human error. (Same as what MrTom said with visuals added)
The image below has the pair of original images at the top
Next set down is the luminance values of the images
Next set down is the saturation values of the images
And the last pair down is the hue values of the images
The lower image of the pair has all the telltale signs of going through a lossy JPEG compression while the upper image of the pair has not gone through any lossy compression.
Culprit - The OP did not use the same steps for both images so human error.
The image below has the pair of original images at the top
Next set down is the luminance values of the images
Next set down is the saturation values of the images
And the last pair down is the hue values of the images
The lower image of the pair has all the telltale signs of going through a lossy JPEG compression while the upper image of the pair has not gone through any lossy compression.
Culprit - The OP did not use the same steps for both images so human error.
Last edited:
Sorry for delayed comment and thank you very much for attention.
Indeed the second (more saturated) .png file could be originally saved as .jpg on Mac. Maybe I saved it as .png after comparing it to the project file on my PC but...
Why then these two last samples below, coming from Mac as .jpg and .png, aren't so noticeably changed in saturation like the first one from Mac?
original PC png
first Mac jpg -> png
second Mac jpg (10-quality)
seconf Mac png
Indeed the second (more saturated) .png file could be originally saved as .jpg on Mac. Maybe I saved it as .png after comparing it to the project file on my PC but...
Why then these two last samples below, coming from Mac as .jpg and .png, aren't so noticeably changed in saturation like the first one from Mac?
original PC png
first Mac jpg -> png
second Mac jpg (10-quality)
seconf Mac png
MrToM
Guru
- Messages
- 3,595
- Likes
- 3,322
Not too sure what you're asking there but the third image is being compressed, 'quality' = 10.
If you don't compress it, ie save it with 'quality' = 12, is there still an issue?
The only reason I can see for the difference is the artefacts created by compressing the file with 'quality' = 10.
Regards.
MrToM.
If you don't compress it, ie save it with 'quality' = 12, is there still an issue?
The only reason I can see for the difference is the artefacts created by compressing the file with 'quality' = 10.
Regards.
MrToM.
"Original PC png" is a template. The main problem was why this image titled "first Mac jpg -> png", doesn't look like the template?
Following your way of thinking MrToM, that it has artifacts and it was a .jpg file, not .png file initially, I saved two more files on Mac to check if you right - "second Mac jpg (10=quality)" and "second Mac png".
It seems that, apart from the artifacts, there's no big difference between these two last files and the template, unlike the "first Mac jpg..." and that's my question now...
Why does the first image from the Mac look different than the others? Is it about the quality of the .jpg?
Following your way of thinking MrToM, that it has artifacts and it was a .jpg file, not .png file initially, I saved two more files on Mac to check if you right - "second Mac jpg (10=quality)" and "second Mac png".
It seems that, apart from the artifacts, there's no big difference between these two last files and the template, unlike the "first Mac jpg..." and that's my question now...
Why does the first image from the Mac look different than the others? Is it about the quality of the .jpg?
Last edited:
thebestcpu
Guru
- Messages
- 3,117
- Likes
- 2,873
Hi chraqek
It would help to what you are trying to achieve with your project or is this just an academic exercise.
If at any point the image is turned into a JPEG, there is quality lost than cannot be recovered even if it is converted back to a JPEG.
So bottom line, once turned into a JPEG, it will never be as clean as the original uncompressed PNG template.
JPEG compression focuses its compression on color first rather than luminosity so artifacting and changes in color can occur.
Also, I took you image through a program JPEGsnoop and you are also using a JPEG compression with the progressive scan option set to 3. This is worse quality than having the compression set to baseline standard.
If you want an exact template, do not use JPEG compression anywhere in the work flow. If you want to minimize the differences, the use quality level 12 and baseline standard option. It will still have artifacts and some color shifts though.
To know exactly how any one image is different and why, one wold need the "exact" steps that were taken in workflow with all option settings. Yet I think that would just be an academic exercise and not sure that helps with the end result you desire.
I hope this helps.
It would help to what you are trying to achieve with your project or is this just an academic exercise.
If at any point the image is turned into a JPEG, there is quality lost than cannot be recovered even if it is converted back to a JPEG.
So bottom line, once turned into a JPEG, it will never be as clean as the original uncompressed PNG template.
JPEG compression focuses its compression on color first rather than luminosity so artifacting and changes in color can occur.
Also, I took you image through a program JPEGsnoop and you are also using a JPEG compression with the progressive scan option set to 3. This is worse quality than having the compression set to baseline standard.
If you want an exact template, do not use JPEG compression anywhere in the work flow. If you want to minimize the differences, the use quality level 12 and baseline standard option. It will still have artifacts and some color shifts though.
To know exactly how any one image is different and why, one wold need the "exact" steps that were taken in workflow with all option settings. Yet I think that would just be an academic exercise and not sure that helps with the end result you desire.
I hope this helps.
Is it ok now...?
"Original PC png" is a template. The main problem was why this image titled "first Mac jpg -> png", doesn't look like the template?
Following your way of thinking MrToM, that it has artifacts and it was a .jpg file, not .png file initially, I saved two more files on Mac to check if you right - "second Mac jpg (10=quality)" and "second Mac png".
It seems that, apart from the artifacts, there's no big difference between these two last files and the template, unlike the "first Mac jpg..." and that's my question now...
Why does the first image from the Mac look different than the others? Is it about the quality of the .jpg?
MrToM
Guru
- Messages
- 3,595
- Likes
- 3,322
Tom Mann
Guru
- Messages
- 7,223
- Likes
- 4,343
Guys, a caution: before you suggest that there are significant differences between uncompressed files (or compressed by a lossless algorithm) compared to files saved as JPGs at Q=12, then please read through this post and the surrounding thread:
Save jpeg at quality 10 vs 12...
In it, I describe the results of a simple experiment that shows that when saved as a JPG at a quality setting of 12, any differences between it and the original are less than 1 quantization level at 8 bpc. Also, take a look the demo of the incredible sensitivity of this subtract-then-amplify method by comparing two versions of the same image, the only difference being that the second version was shifted by 1 pixel.
Tom M
Save jpeg at quality 10 vs 12...
In it, I describe the results of a simple experiment that shows that when saved as a JPG at a quality setting of 12, any differences between it and the original are less than 1 quantization level at 8 bpc. Also, take a look the demo of the incredible sensitivity of this subtract-then-amplify method by comparing two versions of the same image, the only difference being that the second version was shifted by 1 pixel.
Tom M