Photoshop Product Manager, Bryan O’Neil Hughes, gives you a glimpse of new selection technology for Adobe Photoshop CS5 that offers better edge detection and masking results in less time—even with the trickiest images, like hair. What tools do you currently use to make challenging selections?
See the original video here.
To make great selections of people, I generally trace the person using the pen tool, then use the path as a pixel mask. Then I use the original picture as a reference and repaint the strands of hair into the mask. Then paint the proper colour back into the hair (and discard any trace of the old background).
This method yeilds near-perfect results, but is extremely time consuming. Usually takes me about an hour if I do a really good job.
Nice video. But next time, break a $5 and give us STEREO SOUND!
I can’t wait for Photoshop CS5.
Enormous improvement in extracting image from the background …now we don’t have to use any third party software’s for the same…
The extract tool works just fine for me, i can get near perfect results with it every time. Just have to know how to really use it.
This looks pretty sweet though.
Pen and mask, baby… Pen and mask…
And oh yeah this would make life so much easier.
Before seeing this I could have waited on CS5 – until CS6 perhaps [sic] – but now? Maybe not so much…
Well depends on what image I work, but I use most of the time calculation with add & subtract because you have some additional control of setting [offset and scale value] and it works great with hair detail when for example you have homogeneous background. Other tool that I use is Color range it is extremely powerful, when I want to subtract an subject from a very colorful background, other way is to work directly with channels and choose from R, G, B the channel with most contrast and adjust it using levels and brush in overlay mode or for more accurate paint dodge and burn.
I’m also using Quick selection tool and Magic wand, not so often but you can subtract good selections if you know how to use it properly and of course with help of Refine edge which is very useful in CS4
About the new one, it is interesting, it will save you time, but for me looks more like an automation, and you are not so involved in process of masking and who what to make a mask from a low quality images that ware shoot with inappropriate background?
It will be a great tool for fast approach and for those who will want to create masks easy without much thinking.
Since ver. 6 I’ve been using the old Channel Mask method for doing most tricky selections, but really, that’s just splitting hairs as there are lots of ways of doing it. Looks like this will be a turbocharged version of refine edge – neat idea. If it works as well as your demo with the difficult example image, it should make working with studio shots a piece of cake. Onward and Upward… seems like every new advance opens up a whole new direction of creative possibilities.
Hi, yes it depends by image i’m working on. Usually my favorite tools for masking are PenTool and Layer Mask. Most extreme selections are well done with Chanels, Calculations and a Wacom Tablet that gives me very very good control over the image i’m working on.
Currently I mostly use:
* Pen tool if the edge is not too complex (possible rasterised and selectively refined and blurred to match the photo).
* Using Color Range or a channel as a base, and then touching it up with dodge/burn.
Usually it is a combination of several methods…
The new improvements to Refine Edge looks great, and should save a lot of time on many masking jobs.
Hm, i dont understand the hype around it. With the color pick tool of CS3 i can get the same result.
“I remember there’s a whisker here” Isn’t it better that you paint the selection in one view and see it update in an “output” view (2up-layout) or something? That way it’s a lot more intuitive!