LiquidAcid |
Aug 2, 2007 06:09 PM |
Quote:
Originally Posted by Overkill
(Post 483319)
But why keep it in JPEG if it's going to adversely impact the graphical data?
|
It's not impacting the graphical data. It's only impacting the manipulation of the graphical data. And this problem seems to be solved because such transformation tool exists.
Quote:
Originally Posted by Overkill
(Post 483319)
Did he have a choice initially to not use JPEG at all on the data?
|
Because it's a bad idea to start with a source image that is in JPEG format, decompress it to a raw uncompressed bitmap, do the transformation and then compress it a second time, either with a lossless codec or a lossy one. Both methods are not optimal because
a) your destination image is generally larger than the source image (I doubt that any lossless image codec can use the fact that some image data was previously DCT-transformed and quantized)
b) image quality is degraded when re-compressing with the same or some other codec
Quote:
Originally Posted by Overkill
(Post 483319)
If it isn't greatly affected, who cares.
|
Maybe he cares....
Quote:
Originally Posted by Overkill
(Post 483319)
I can tell a JPEG when I see one, and for the most part I don't cringe, unless it's on something like pixel art.
|
That's your opinion. There are a lot of people which need this lossless transformations, otherwise such software/code wouldn't exist.
It seems a bit funny to me that on the one hand people are crazy when it comes to this new HD technologies with higher and higher resolution and on the other hand they recompress JPEGs, edit MP3s by decoding them first, use the onboard integrated soundchip , etc. (long list here)
Maybe the meaning of quality has gone absent...
|