If I want to clone-stamp over a noisy area and make it less obvious (not sure if it’d fool the software), I’ll make duplicate layers and clone each one separately, using different sample points, then overlay them using soft-light/hard-light/overlay or whichever suits for the image. Then, when flattened, no parts are mathematically identical.
I used to have a jpeg compression program years back in Geocities and Altavista days (pretty sure it was coffeecup software), that allowed you to ‘paint’ different regions of the same image with different compression levels. This allowed you to massively compress the sky for instance, whilst only lightly compress your focal point.
I cannot for the life of me find it though.
3 Likes
That program sounds awesome. I’d spend the effort to compress quite a lot of photos that way, simply to reduce storage costs. Because my time at work is worth nothing to the business, so I can do what I want.
I’ve even had a look around on an old hard drive, to no avail. Even my Google-Fu is failing me today. I’ve got a couple of other hard drives I can check later on.
If I come across it I’ll let you know. It was a godsend in the days of dial-up.
2 Likes
Here you go; Image Optimizer, by Xat.com
The technique is called ‘Weighted optimization’ or ‘regional compression’.
And here’s a Photoshop method.
ETA
In video terms they appear to refer to it as Region-of-interest (ROI) coding
3 Likes
4 Likes
I remember the selective compression you could do with Image Optimizer, similar to how you paint a masking layer in photoshop in mask mode. That was the same era when ProJpeg (very similar) and ImageVice both from BoxTop software were the top of the game.
These day’s tools like JPEGmini and ImageOptim that rule as best in class image optimizers.
JPEGmini: PAID
low visual loss jpg compression.
jpegmini for jpg files isn’t free but their compression results in big filesize reductions that have very low visual loss.
tip: If using photoshop I recommend save for web setting high quality for clear low compressions jpeg that you then feed to JPEGmini for best over compression to quality level. Don’t pre-compress jpg in photoshop any lower then high quality or photoshop will reduce the visual quality too much.
ImageOptim: FREE
lossless png optimization.
imageoptim is for png files, and it utilizes all the cutting edge lossless png optimization and compression tools and some custom code of their own, no other png compression tool does as good of lossless compression.
tip: turn up the number of passes up to max in the settings which makes the optimization take longer but you get smaller files…it really isn’t that long on most modern computers.
tip2: you can hit run a second time and sometimes get a little additional compression.
1 Like
Optimization’s not something I use much these days though, Storage is cheap.
Yeah, I don’t optimize my personal photos.
I use it mainly for images for websites, mobile sites, or mobile apps where deployment size and speed loading are benefited by these techniques and that these techniques do not compromise visual quality to achieve compression.
@awjt for example, your banana image optimized through adobe very high save for web and imageoptim takes it from 40,977 bytes to 29,578 bytes with near lossless quality, you could likely get even higher compression if you started increasing the loss slightly where it isn’t perceptible in certain areas.
1 Like
That is absolutely bananas. You truly are the master. 
2 Likes
That banana’s becoming perceptible everywhere.
have you difference mapped the two images? the difference map is solid black.
(edit 1: wait, weird…that file is much larger then the one I upload and compressed differently…does the bbs have image compression? darn it is messing things up!)
(edit 2: reuploaded and it is showing the right image again…for now.)
1 Like
Meh. Clearly fake. Moran.
5 Likes
You should lower the banana liqueur intake 
1 Like
Wait a minute–I thought the “Falcor” one was just some in-thread funning by @funruly (post #13, above)…methinks someone didn’t see the joke? (Or am I missing something altogether and seeing jokes where none are intended??)
Interesting discussion either way, though!
5 Likes
I keep hearing this but when their output is actually recorded and transcribed they usually seem unable to put together a coherent sentence. Many of the Republican candidates seem well versed in the art of non-sequiturs.
In the UK, class is conveyed by accent and vocabulary. In the US, where PR has been more exploited, it seems to be more the choice of words. You detect someone’s class by the way they refer to people from other social groups, and the words they use around employment, taxes and so on. In effect, it is not education (which teaches you to think) but training, which teaches you to perform.
6 Likes
I first read that as someone claiming to be a descendant of the Founding Fathers, or perhaps just a white supremacist. Then someone pointed out that the signwriter meant “dissent”.
A lot of the people who write these signs read little and exist in a largely oral culture, so I find the mis-spellings totally understandable. Having spent a little time in Kentucky, I am well aware that a phonetic transcription of Kentucky English would give some very odd results.
After all, in Chaucer’s time while Latin and French were the “educated” languages, with pretty regular spelling, English was largely oral, and Chaucer or his printer couldn’t even make up their minds how many esses there were in Englishe/Englisshe.
6 Likes
Ohhh, these things drive me nuts in the most hilarious way!
At least the people making the signs don’t write that odd “S” that looks like an “F”…or is the the other way around?
