Photo enhancement

This really has nothing to do with New York State railfanning directly, but we have some pretty talented people computer-wise and photographically, so I'll post here in hope of finding one of them.

We've all seen the TV cop shows where they have a blurry image of some car's license plate, and the cops use computer enhancement to clarify the image enough to read the number and catch the crooks. I'm sure there is a degree of artistic liberty there, but to some degree that CAN be done in some cases.

I was looking at a photo of a NYC loco and trying to read the number of the wooden caboose in the background. You can SEE the number, as you enlarge the image you can read the "NYC", but the number gets just a little too fuzzy to make out.

Is there some program or method that somebody other than Dick Tracy can "adjust" a photo to sharpen it up? I'm thinking that since we know what the "NYC" is supposed to look like, if you could make an adjustment to bring THAT into sharp focus, whatever else you have nearby - like the number just below - should also show an increase in clarity.

Any suggestions, or is this a lost cause?

The smart fix and exposure setting in Adobe Photoshop Elements 11 seem to clean up a great deal of "fuzzyness". Brightens/darkens and gives you a beautiful clean image. Depending on the size of the original photo though there may be nothing you can do.

It's really hard to get any kind of sharpness back in a photo. Sharpening a photo works by increasing contrast in the edges -- however, if the photo is fuzzy to begin with and thus lacking in clearly-defined edges, you'll find yourself getting to the point where the text in in image becomes pixelated before becoming readable.

Any chance you could post the image in question to see what we're up against?

While it was my curiosity on the caboose number in this particular case, what prompted me to ask was the fact that old pics often have some detail we wish was plainer - a milepost which would help with a location, a station sign we can't quite make out, or some detail on a locomotive. On the cop shows, there's always some chick with wacky clothes and weird hair who can make three strokes on her keyboard and instantly make the fuzziest photo crystal clear!

I'm a computer illiterate so I'll need to be talked through the photo posting. I don't know where I got it from - it's a B&W publicity photo, staged, of NYC FM switcher 9108 in a yard, the loco is on a yard track while a switchman in the foreground operates a Racor stand and signals another engine which is out of sight behind the photographer. I have it saved as a jpeg on my computer but can't seem to copy it to bring it to this page - right-clicking does not give a "copy" option.

BR&P wrote:While it was my curiosity on the caboose number in this particular case, what prompted me to ask was the fact that old pics often have some detail we wish was plainer...

Any Photoshop tricks really involve making something lighter or darker, or increasing contrast between edges (as explained earlier). Sometimes these are enough to job the shape recognition pathways in our brains to figure out that blurry car number or milepost. As stated before, you really can't "add" data that isn't there, that's where your brain comes in!

-otto-

Email it to me if you can. I've got Photoshop at home and will give it a shot.

Sent.

Of course with Photoshop, it can be any number you want it to be!

So if I have that photo in my Documents, how do I post it on here?

BR&P wrote:Sent.

So if I have that photo in my Documents, how do I post it on here?

Normally, the image has to be on some sort of web server. Many folks use the several photo sites that are available, either railroad related or not.

For one-on-one (such as CPSmith suggests), send a PM via the webboard and exchange emails. Then you can just email it to him as an attachment.

Thanks Tree. I already sent it to CPSmith and another member who contacted me off list. I do have Photobucket so I suppose I could upload it to that and then bring it here. I thought there might be a simpler way but apparently not.

BR&P,

I tried zooming in and cropping and other things.....it became too pixelated to make out the number.

Charlie

Oh well...worth a try. Thanks! I could not quite get it either. Fortunately it's more a curiosity thing, not something I really HAD to determine.

Things like this:

https://www.youtube.com/watch?v=3uoM5kfZIQ0" onclick="window.open(this.href);return false;

are pure fiction..so ridiculous they are laughable..
The eye from a photo like this:



when magnified gives you this:



and there is nothing more anyone can do with it..even the FBI.
you cant bring out detail that doesn't exist in the first place.

For BR&P's caboose photo..the version on-line is only 640 pixels wide..
when you try to zoom in on the caboose number you get this:



No one will ever be able to get any more detail than that, ever, because more detail simply does not exist in the image.
What is really interesting is that everyone *thinks* they can make out N.Y.C.!
It really does looks like N.Y.C. is really there..I see it myself!
but in reality, no one can actually see it at all..
we only *think* we see the letters N.Y.C. because we already know they are there..
If it was a more generic caboose, from an unknown railroad, no one would know what those letters were..

For this particular photo, a more detailed version might exist!
the resolution might be higher on the original negative, enough so that the number can actually be read..
but we would need a higher resolution scan of that original negative to know for sure..

Scot

Good points, Scot. I actually CAN read the "NYC" when it's at a lower size but at some point the distance involved with the number - and thus needing it bigger - crosses the line with clarity as your eye example shows, and further enlargement makes it worse instead of better.

If they had painted the caboose or at least the number more recently, we'd probably be home free!

On the subject of "photo enhancement", here's a tip which might help someone. While digital is the current standard, there are a gazillion slides still out there. Sometimes you want to see some detail on a transparency but projecting it on a screen loses a little clarity. There's a device which can help.

Remember microfiche files? little square pieces of film with detail shrunk down, and you'd use a special reader to see them. I have a reader left over from those days and it is perfect for studying fine detail on a slide. It's maybe the size of a computer monitor (the old ones, not the flat screens) and not very heavy. The one I have is an Eye Com 1000, and I don't see any on flea-bay. They are a dinosaur, but that means if you CAN find one you can probably drive a good bargain.

They also work well on B&W negatives.

BR&P wrote:Good points, Scot. I actually CAN read the "NYC" when it's at a lower size

well no, you cant..not with the photo in this thread anyway..
that was the point I was trying to make
the letters NYC are physically not there in the image..
This is actually what they look like, from the actual photo:



making it smaller does not make it clearer..you cant see letters that are not physically there..
they could be ABC or XYZ or GGD..your brain makes them look like NYC because you happen to
know that is what they actually are..but you dont *really* see NYC..because the letters NYC do not physically exist in the image..
blowing them up does not make them clearer, and making them smaller does not make them clearer..
it is what it is..and what it is is a fuzzy blob with no discernible letters or numbers at all..
very interesting stuff!

Scot