Realistic Photomanipulation -What areas of photoshop to focus on?

[therue]

Hi guys,

I'm really interested in learning realistic photomanipulations of characters (for example, taking a character from an image and compositing it seamlessly into another), but i've just started learning photoshop at the moment, so i was wondering if anyone could point me towards the right directions perhaps?

There's so many areas to learn when it comes to Photoshop, and due to the wide variety of art professional who use this software, surely some techniques or areas of photoshop are more relevant to certain type of artists- so when it comes to realistic photomanipulation, what areas of photoshop should i learn or focus on? and does anyone know of any good tutorials/guides/web sties/ or any learning resources they could recommend?

thanks in avance.

 

[Igrecman]

There are two tips I can give .
Perhaps the most important one, Work on the highest possible image resolution you have. It is very important for masking and cutting precisely the subject to paste on the second picture.
One mistake many people make, is not paying attention to the direction the light is coming from in both pictures. Lighting should be similar in both images.
You have to try and learn from your mistakes. It can take some time to get good at it. It took me a few years to be more than average at that.

This is an example of mine.
The two dogs were pasted in of course. The resolution was not actually good enough for this but it worked fine anyhow. As you can see, the brown dog lit by the sun and from the front works, but the same dog in shadows wouldn't work at all

 

[SCTRWD]

First things first: try to get rid of the mental block that you "compose one image over another"...

You compose TWO images together at the same time. You are free to change both of them.

Then come things like size, light, shadows, focus, tone, noise, etc. compatibility

One important thing is to decide what you really want: you want two images "look" seamless, or you want to make a viewer "believe" they are seamless and what you can do to make it happen. Sometimes that calls for different techniques and approaches.

The name of the game here is try to make two images interact in some way: cast shadows on each other, obscure each other, reflection, color cast... things like that...

If there are no natural openings for it in your images consider adding them... In movie biz they call them "props": some little thingies whose sole role is to turn a viewer into believer That works, too...


And, of course, there is a range of trade cheats. Like toning the whole image to some color(they say it's for drama look but we know better It's just a quick way to shortcut the tedious work of tonal addjustments).

Or introducing some attention distractors. Believe it or not, back in the old days there was a technique of making hopelessly blurred images look sharper by simply adding noise

That's just from the top of the head. Maybe I will come up with something new later Or the fellow guys here will make some additions...

 

[therue]

what if one image is a lot better quality than the other? in terms of resolution, etc
like if one image is a taken by a professional camera with with more pixels and details, i guess just changing the tone and things in that nature like you said?

 

[SCTRWD]

It's pretty much always the case Clients always give you the sources of different resolutions and sizes

Nothing can be done with that, your final image will have to be the lowest resolution of the two images. Transform will make it all for you, just choose the right Transform intent and deal with the consequences. That's part of the job....