lechat

Is that the saddle i sold you?

f4rrest

I've got an extra Nifty Fifty I don't need (and a few other lenses). Got any interesting bike stuff to trade?

simonaway427

My son decided to one-up the OP:





No photoshop

andrewluke

For me it is. Photoshop is another tool for photographers but too many think it is now the standard that fixes all. Those of us that still attempt to do this in the camera have to deal with people assuming we suck at photography and are just great in PS and I'm tired of it. It's a tool, not the answer.

Sorry... little hyperbole. Don't take it so damn seriously. Everything besides the drastic creative stuff can be done old-school. Color filters, distortions, and everything have been done for decades.

dewaday

I think you're hanging on to outdated notions of image making. If I can create a more emotionally evocative image with absolute control over my vision, inside the box, why care how it was made? Clients don't, viewers don't, I don't, as long as it satisfies the intent. Sure lots of people use PS as a crutch, but they're still crummy image makers without vision who occasionally get lucky, but long term it's still the same old boring derivative crap.

I suppose there's a place for the purist approach, but it's pretty tough to deny the expanded visual opportunities offered by digital. Everything's a tool.

waterrockets 

Sorry for the confusion, the Nifty Fifty is part of the list of things I already bought that cause me to wait for some more budget. Now, if you want to trade a 580EXii for cheap...

waterrockets 

Here's the ELA for the original shot. The bright areas are those that were modified. Note that the background grass and tree blur is not modified.



Here's a photoshopped motion blur (note the streaks in the ELA):
https://www.errorlevelanalysis.com/permalink/afc2b92/

dewaday

What if he wasn't working in jpeg format when the manipulations were made?

waterrockets 

He was working in jpg format when he sharpened things, and that's one of the more basic functions you'd want to do when working in RAW...

I just don't think the motion blur was shopped into this image after the original exposure.

SlimAgainSoon

Nice shots. Especially the pan ... well done.

The second one, the static, looks little flash hot to me.

But I'm not pro ... I just like to look at the pictures!

gregf83 

How do you know that? I have no idea whether the motion blur was added but that ELA method of detecting changes doesn't seem like it would be useful for someone shooting in RAW. Why would someone fool around with editing JPGs when they have a higher quality version available.

dewaday

Unless he saved out a jpeg image for general distribution, then resized and re-saved a smaller web version and re-sharpened for the smaller size. Pretty common workflow that's a well used action on my box. All kind of academic, since the image is what matters and the image works nicely regardless of how he got there.
Doesn't this pointless man swinging belong in foo or P&R?

gregf83 

I thought the pictures were fine. I don't care if they were manipulated or not as the original purpose of the photos was to experiment with lighting and they looked good from that perspective.

I was just curious about this ELA technique which was something I had never heard of before. I did a little reading and it doesn't seem like a definitive technique for determining if a file has been manipulated.

waterrockets 

Well, it would appear that there were portions sharpened in .jpg, because they are brighter. It could be that those edges were just sharp enough to go white with ELA too, but the helmet doesn't seem like it would pop like that on its own w/out modification. The blurred areas all look to be about the same color though.

You're new to the internet? This is in every thread on every forum everywhere. Nobody has said anything like: "You know who else liked fake blur, right?" and then posted a picture of Hitler. That would be a P&R conversation. Foo would be a bunch of flirty tease girls rating the subject's hotness, so this doesn't belong there either.

True on both counts. I don't care if it's manipulated or not, but I just haven't seen any evidence that was. The blur looks natural to me.

ID64

Yep After proper fitting session is the most comfortable seat ever. Thanks!

Nerull

I'm not quite sure you're understanding how ELA works. I'm not sure the guy behind that website understands it either - on different pages he explains bright edges as both a normal byproduct of jpeg compression that doesn't mean it was edited and later explains at as signs it was edited.

I put in several photos that I know were not edited (because I took them) and it showed many edge highlights like those in the ELA you posted. That seems to be completely normal with edges in JPEG images. Red seems to cause the brightest edges in my images - and what do you know, the bright areas in the images posted are red.

Then I put in a few images that I know had a photoshopped motion blur, and it didn't seem to detect any of them.

From what I've seen, ELA is more or less useless at detecting anything except which parts of the image are red.

thegunner

tsk tsk tsk WR... iso bumping will only take you so far on the t2i. if you had a 5dii or a d3s you could essentially play with iso as you please, but there's a marked dropoff in image quality at 3200 on the 550d. you're better off risking a shot at f/4.0 with a narrower DoF than trying to use iso to account for shutter speed.

waterrockets 

Looks pretty obvious with my sample that I just ran on a photo of my Fuji:

https://errorlevelanalysis.com/permalink/7578f99/

The t2i can perform pretty well at 3200, but that's about the end of it. The marked dropoff is 6400. The original image was taken at ISO 100, so I've six stops to play with.

thegunner

we must have differing opinions on what's usable then the pic you posted was just fine, i was talking in general.

rollin

You're new to the internet right? if a mod rambles off topic then it's ok.



I think Hitler did like fake blur and the OP is kinda dreamy.

waterrockets 

I dunno, now I'm seeing the blue sky is a really different color in the original ELA (similar to my faked blur example).

waterrockets 

Well, I'm no expert, but 3200 seems acceptable to me (not ideal, but certainly good enough): https://www.smugmug.com/gallery/11350...97928646_CcRXz

Nerull

So why did you have to photoshop the logos on your Fuji?

Sure, a completely out of place blur is going to show up. Now try one that's well done.

Here is a completely untouched image I took earlier this year. Now, pretend you had no idea about the authenticity, what would you say, based on just the ELA?

https://errorlevelanalysis.com/permalink/10c1f80/

It has the features of the original image being discussed - edge highlights, dark sky, red standing out like a sore thumb - things you've cited as evidence parts of the image were retouched.

Nerull

On a whim I looked up some more info about the guy behind ELA. It seems even he can't keep his story straight about which images are edited and which are not. It is far, far from foolproof and if the guy who invented it has trouble interpreting the output, trying to use it as hard evidence of anything is foolish.

In fact, you might wonder if Dr. Neal Krawetz ever gave a presentation on ELA in the first place...
https://errorlevelanalysis.com/permalink/84c6402/

waterrockets 

Yeah, it's looking like ela is better for real alterations like adding lipstick or changing the shape. Especially pasting from other images.

I still think the original on page one is legit motion blur.