aislinbluemage wrote:No. Just... no.
A lossy compression algorithm can never be better than a lossless compression algorithm. One removes data that the compression algorithm deems we cannot notice due to the rounding effect our minds produce. Our brains "fill in the gaps" but things still look or sound... off.
Lossless compression algorithms, however, remove no data. They generally just change the way the data is stored and remove the "gaps" of empty data.
Less data will always be worse than the complete data.
Didn't look at my link did ya. The guy I referenced shoots JPEG for a living. He knows his stuff.
A JPEG, done right, will look the same as a PNG for photographs (but NOT for text or cartoons or things like that). And by the same, I mean that even a professional will have trouble telling the difference until you zoom in to a pointlessly high zoom. If you really wanted to go crazy, you could go for RAW or shoot TIFF. But a TIFF file is huge and generally is less versitile than a RAW (as is the case for the Nikon NEF, where my D70 produces a 5 MB NEF and ViewNX produces a 15 MB TIFF). But even the TIFF throws data out to produce a larger file.
See this link for a comparison of RAW vs JPEG --
http://www.kenrockwell.com/nikon/d200/q ... ttings.htmA JPEG at 'Fine' level is either lossless or near lossless. A Photoshop created JPEG at quality 12 is lossless. A Photoshop 3-4 is good enough for the web. Modern professional cameras can produce a lossless JPEG. Not that it is needed.
My current camera can take 1.6K photos on an 8 GB card at the highest quality. When I drop down to the lowest quality, I can take 4.8K. If this quality is 'good enough', and I might be in a situation where I need to take 2K+ photos before I can off load them, then this 'good enough' quality becomes the better one. Besides, larger files start to get in the way. If you double the file size, you double the number of CD/DVDs and hard drive backups you need. Also, running out of space on the cards is not a good option. Is it better to take a small quality hit that no one will notice or miss a good shot because you were worried about something that can be fixed with 3 clicks in photoshop later.
An example. I took this photo a few weeks ago. Original was RAW and I made 3 versions of it, a PNG, a JPG at Photoshop 12 quality, and a JPG at Photoshop 5 quality. In that order.

PNG lossless, 301K

JPG lossless, Quality 12, 136K

JPG Quality 5, 31K
I'll give you that the bottom JPEG version does not look as good and artifacts are present. But no one would say that the top JPEG doesn't look good. The airplane is a few shades redder in the PNG than the JPEG, but it is otherwise the same. The actual Photoshop version falls somewhere between the two versions, so I'm saying they both reproduce what I am going for equally. Since there would be color correction made before printing, I believe that both sources would come out the same. The bottom image is fine for web use, and is about 11% of the file size of the PNG. The interlaced PNG version I made is 362K for some odd reason. Paint.NET produces a 500K file doing an open and save-as on the first .PNG. In addition to the larger file size, compressed PNG images require additional CPU cycles to process. But this is not nearly as much of a concern anymore.
One of the first things I learned in photography that you always need to know what you are using and why. There is always a tradeoff.
Another quick example -- 62K file, although some of my highlights are blown.
