I might be superfluous in my method here, but it works well for me 🙂
I had this idea for a couple of days and finally decided to do it yesterday. This was the process (very very roughly)
Step 1: Getting the initial photos
This is the embarassing part. You’ll look silly almost no matter what – trying to get a self-portrait in a suitable position that you envision. A camera on tripod, remote control and bounce-flash later though (thanks, I like my bedsheets too):
Got some stock-photography from Pierre on DeviantArt
Step 2: Putting the 2 together
I had to give the photo a bit more room at the top, so I photoshopped some more wall there. With composites, sometimes I’ll mask out the individual parts as precisely as I can initially and save the respective masks in my channels. In this case, I put them together first and then masked it all together. I think that’s a mistake and I should have masked the subject first (me) then masked the television. Could have saved myself quite a bit of time afterwards.
Step 3: Magic
OK, so I skipped a lot of steps – my bad – I didn’t take any screenshots. I was too busy wondering “how the hell am I going to get this to blend”. I guess there’s always a point in these highly-processed photos where I say “I should stop right now, this is just not working”, but I keep on going because it’s taken so much effort already. Eventually, there’s a point where I say “oh, I might have something here”. I suppose that makes it all the more rewarding in the end.
I use a lot of different filters and textures and try to burn-and-dodge a lot.
Step 4: Colouring and painting light
I hate masking and I love “painting-light” and manipulating it. I think that if I actually loved masking, I could have some really amazing work.
At this point, I also did a bit of liquifying and stretched/distorted parts of the photo.
Almost 3 hours later, this was the final shot: