John's Photo tips and tricks

Making better selections

2011-10-05T08:50:00.000-07:00

Today’s blog is a tip or actually a technique that will appear to be so obvious, but honeslty I have never seen it documented.

One of the hardest thing to do in compositing, that is puting images from various locations together in a seamless montage that looks real is the ability to make a great selection.

Looking at Photoshop, notice that a LOT of the tools are there to help you make good selections, and in CS5 we have been giving the “refine edge” tools set that makes life a lot easier, but still not perfect.

Well I had a job recently that make me rethink the entire way I make selections, and in the process have come up with a rather interesting way of doing selections that give incredible results.

First the problems with make a selection:

getting all the area you want selected to begin with
getting areas that are “straight lined” smooth and not rough
getting hair or tree branches to show the background through
makes lines not so harsh that it looks cut out

Okay I know there are more issues, but these in my mind are the biggies.

Now lets look at a common workflow and see where the problems come begin

First we make a gross selection, today that is with the quick selection tool, which does a great job, about 75% of the time, get you in the ball park for sure
Then we go in and refine the selection, using either other selection tools or quick mask to make sure that all areas are covered.
Then we go to refine edge to fix the problem areas, just as hair, glass, trees along with feather, smoothing, and if necessary shifting the line.
After all of this we create our cut out, and I generally create a layer with a mask so that I can fine tune after the fact.
The next step further refinement is based on what the cut out is being used for, if it looks okay then we end, if not we modify the layer mask to make it look right. And here is the problem, quite often you end up with a less then desireable edge, and you have to spend a long time refining, and often messing up your layer mask to make it look right.

Now I want to suggest a slightly modified workflow that I have had great luck with.

First we make a gross selection, today that is with the quick selection tool, which does a great job, about 75% of the time, get you in the ball park for sure. Nothing new here, except you don’t have to be fanatical.
This is the change comes, right click on the selection with selection tool selected, and select from the pop menu, “Make a working path”. This takes your selection and turns into a path. When you get the tolerance dialogue use the default. Now for those of you that have not used the path tools at first they are intimidating. This blog no on how to use the tools, you will have to go someplace else for that.
the first thing you do is go the paths pallet and drag the work path layer to the new path icon, and create a path, and name it.
Then with the direct path tool, edit the path so that it matches exactly the areas you want. You will see immediately how inaccurate most selection tools are. Using the path tool, you will be able to fix smooth curves and make them smooth, and get detail that is virtually impossible with normal selection tools.
When you come to problem areas, like hair get as close to the bulk of the area you want so that in refine edge you only need to make small refinements. This is tricky at first and might take a couple of times to make a good both for these areas, but once you make a couple of these you will bet the idea.
Always error on the inside of the area to be selected. What I mean here is that if you are doing a straight line or smooth curve, move the path a couple of pixels inside the line, this will avoid the “edge” often seen in selections.
once the path is completed and all areas are the way you want them, with a path tool selected make sure the path is not selected, and right click on the image and the submenu will come up that says “make selection”, choose that, and give a feather of only 1. Your path is not a selection. Be sure that the path is NOT selected, that is all the anchor points showing or your selection will not end up correct.
No select any selection tool, and select “refine edge” either from the tool bar or the sub menu.
use the refine edge pallet to correct the problems of hair and trees, and study the cut out carefully. I like to use a feather of 0.8 and a smooth of 1, but that does vary. If the cut out looks good, create a new layer with mask (this is my default and highly recommend it). If the mask does not look good cancel, go back to the selection and create and edit the path. loop this step until you have a perfect mask.

That is all there is to it, if you spend a little bit of time doing this you will be amazed at how good your cut outs will be. And even better is since it is all non destructive you can go back and fix any problems.

A better way to do HDR

2011-09-27T07:50:00.001-07:00

A better way to do HDR

There is not doubt that HDR (Hight dynamic range ) photography has taken the photo world by storm. Not since 4x5 color transparency have we been able to achieve beautiful results in photography. And look at those "HDR" photos that have the surrealistic effect, those are tremendous example of HDR.

Well not really, what is so interesting is that a technology like HDR which allows photographers to achieve some incredible results, has taken a life of its own. What I am referring to is the genre of photography often called HDR Effect, which is really a special effect using on an image using the HDR techniques and tools, but really is not HDR. Now before many of you photographers get out the "tar and feathers" let me explain.

High Dynamic range photography refers to a process that allows a series of picture to be blended together to increase the "dynamic range" of a scene, that is to capture areas of high and low light that would ordinarily be lost.

Now what does that mean, well in reality it is pretty simple. Take a picture and look at it. is there any areas that have the highlights (light area) washed out? How about the dark areas, or shadow areas? If the answer to either is yes, then your scene has a low dynamic range and could possible be improved with the aid of HDR.

What HDR is supposed to do, is increase the detail in the highlight and shadow area that would ordinarily be lost. You can tell when this is necessary by doing these easy to follow steps.

Take a picture that is best possible image, get the exposure as close as you can. This is very important. The baseline image must be the best it can be, otherwise you will get unpredictable results. This is also where a lot of haloing comes from.
Look at the highlight area, and if you have the blinkies turned on you will see a flashing, black and white area. Or an all white area if you don't have blinkies on your camera.
These are your highlights that will not have any detail.
Now look at the dark areas, and see if there is any detail in those areas, those are your shadow area.

Now that you have done this you can determine if you need to increase the dynamic range of your picture. The general rule of thumb is that the brighter the day the greater the chance you will need to apply HDR. On a cloudy day the difference between the bright and shadow area is slim, thus increasing the dynamic range of the photo is different then on a bright shiny day, but for right now let assume we have a bright shiny day and our picture reveals that we have lots of white and black areas on the picture. Now the HDR process can help us solve this dilemma.

The HDR process involves sandwiching more then one digital image together to increase the range of detail in the shadows and more importantly the highlight areas. Generally speaking photographers are told to take three or more images, and to bracket the images 1 to 3 f/stops and then process the image.

While this will work, it might not always provide you with the best images. And this technique often yields strange side effects such as haloing on edges. but not always, sometimes the images come out looking incredible and other times well with some strange edgy effect, that has become know as the HDR effect.

What is actually happening is that this "HDR effect" is actually a special effect generated by the HDR process, from insufficient mid-tones in the images.

Also the most often question I get on HDR is how many images, and what exposures bracket should I use. Most of the time people are told to just try it, and I have hears all kinds of suggestions, 3,5,7,9 images, and anywhere from 1 to 3 f/stops intervals.

In reality there is a very easy way to determine how many images and what bracket you need to extend the dynamic range of any picture, wether it be a high contrast scene or a low contrast scene.

From now on rather then using f/stops for the interval I will use EV or exposure value. This is found on most cameras, and for canon owners this is called EC, or exposure compensation.

Before we start there are some perquisites you need to know about.

You must use a tripod, this is essential for if you get the picture out of registration no alignment will help
while not necessary I strongly recommended two extra pieces of hardware Hoodman's loupe so you can actually see the images during the live view process and during the evaluation. Also I prefer to set my own custom white balance using the expo disc the results are noticeable.
You have to use aperture priority or manual, do not use any automatic mode nor shutter priority, you have to keep the f/stop the same. Now with that being said, you can get some very interesting effects, by varying the f/stop but that comes after you know what to do.
Use the lowest available native ISO on your camera. With HDR any form of noise is deadly especially in the the images that have extreme highlights
Your camera must have a live view mode. If your camera does not then don't bother going beyond this point for this technique will not work.

Now lets get started.

Take your first picture. Look at it. Note the shutter speed.

Now using the live view mode, adjust the camera EV (or if using manual mode the camera shutter speed) so that the highlight areas of your image are perfectly exposed, dis regard all other areas of the picture, just the highlights. When you are satisfied with the image, take another picture. I have to emphasize here that don't just get rid of the blinkies, make sure that the highlight area is perfectly exposed. this is where a lot of mistakes with HDR occur, people end up with the highlight area just down from the over exposed area, in stead of completely exposed, and this one of the causes of the haloing.

Now this is where this technique differs from most. Normally you would just set a bracket exposure of +/- 1 or 2 EV of the image and be done. There is a danger in this that valuable mid tone information will be lost, so using the EV compensation again, lighten the image by one full EV. If you are back to the original exposure then stop, if not take another picture, and repeat this step until you have an image that is very close to the base line you took. Do this even if you are within 1/3 of an EV of the baseline image.

Now that we have done the highlights we need to do the shadows.

You should be at your original exposure setting. Now using the EV lighten the image until the shadow area is perfectly exposed, and yes the highlight areas will be way over exposed. but remember we took care of the highlight already.

As before come down 1 full EV and see it the picture matches the baseline. If it does then you are done, if not repeat this until it does.

Now at this point your dropping to baseline, might be less then an f/stop and that is okay, but if more then one full EV you need those mid tones.

Before we go further there is an interesting phenomena that I have noticed. When taking the first highlight pictures, you will notice that a lot of the time, your shadow pictures will be the same number of EV but in opposite direction. This is why the auto bracket method works a lot, but not always.

Also notice that the midtown pictures are not needed about the time amount which is why the three picture method works a lot of the time, but not aways.

Now that you have all of your images you need to assemble them. I know I am asked which method do I use, and honestly it is hard to say. I know that I like the control of HDR Effects pro, but their alignment and ghosting leaves a lot to be desired. So if there is an issue with that, I use the HDR in Photoshop, and found it works really well.

So assemble you images in whatever program you like you should notice that the images looks a lot like a really nice 4x5 transparency, with brilliant colors and little any any halo effect.

But before we finish we have to discuss two more problems that is what do you do with a picture at base line that is just fine, how do you get the brilliant HDR effect with that, and what do you do with an image that is flat, can HDR solve that problem.

Now that we have extended the range of a high contrast scene what can we do with a low contrast scene.

Well at first you might think that HDR can't help this situation, and honestly in most situation you are right, but I learned that in some situations HDR can help.

Shoot your baseline photo, and if the scene is flat that is there is not real contrast to the scene try this.

Adjust your EV up until you get the blinking highlights, this will give you a shot with your shadows with no detail, sort of the reverse of what we were doing. Generally this will be less then 1 EV.

Now do the same in the opposite direction until you get a harsh dark shadow, this will give you a highlight area. And this will generally be less then 1 EV.

How this trick will not always work but worth a try on a cloudy day that you want ot give some ompf to .

But what about the perfectly exposed day, should you use HDR, and the answer is no, you can achieve that same look but adding a bit of contrast, and some vibrance.

And remember, these are not absolutes, but starting points.

The problem with photographing blondes

2011-08-27T09:20:00.000-07:00

I know it has been a while since I posted but I will make up for that, I have a whole lot of new material to share with you, starting with today

Okay, I know blondes have been a problem all through history. And yes, I can hear it now. But I do have good news, at least in photography blondes provide very few problems, at least in front of the camera. What happens before or after the shoot, is well a different matter While it might appear that blondes skin tone will cause metering problems, in reality my experience has been that it is less then 1/3 of an ev off, and in most cases right on, seems that like brunettes there skin tone almost matches 18% gray.

But there are two things you need to watch for with blondes, there hair color, and freckles. Blondes are very proud of their hair color so be sure that you watch the exposure and don’t blow out the detail, nor make it darker then it actually is. I have found if you use dodge and burning techniques to add extra highlights this works very well. But this only works on blondes, if you try and add highlights to red heads, brunettes, the result is less then desirable.

If you error error on the side of keeping the hair lighter then darker.

Now for the problem of freckles. While most red heads love there freckles, blondes hate them with a passion. Now this is not a 100% guarantee, but I would say 90% that you need to assume that they wan the pesky brown dots to disappear in the final photo. I think the reason for this is that the blondes freckles are not as distinct as on a red heads, and to the blond appears to look like they need more make up or ?

If this is the case, and you subject is screaming at you to make sure that her freckles don’t show, never fear, on blondes getting rid of freckles is quite easy, and can be accomplished with a simple gaussian blur, as discussed in the chapter on retouching.

Photographing Hair

2010-11-15T17:07:00.000-08:00

Recently I have seen a lot of models I know showing me pictures of hair. Well that is because they are hair models, and they are supposed to be showing off the creations of local hair dressers.

Also I have seen some photographers I know doing the same thing.

The problem is that in the attempt to do "part" photography they forgot the cardinal rules of photography, "Know what you are shooting" or "Know your subject".

Here is the problem. The photos were generally good, some great... of the Model. But they were supposed to be of the hair, not the model. In this case the model is not the subject, rather the hair is the subject.

That was the first problem.

The next problem was that a majority of the photos shot face front, head and shoulders. While this is great for models, and portraits, this is not that great for hair.

Here are a few things to consider when doing hair photography

1. The hair is the subject, not the person that is attached to the hair. You are showing off a hair dresser's creation, not the model.

2. Be sure to take 4 shots: front, left side, right side and back. And, remember when I talk about "talking to the model", in this case, you should talk to the hair dresser. Find out what they are doing with the style, what they want shown, and more importantly anything to be avoided.

3. Avoid face front. On your front shot have the model angle her face so you can show off the hair. Remember, the hair is what you are photographing, not the face.

4. Avoid using flash. Flash and elaborate hair dos normally don't go together, the flash produces way to much light and contrast. Make use of natural light, preferably a window light after about 3 PM, and north facing. Hmm, can't do that, then I suggest using studio light, you can find some that are very portable, white balanced, and cool. If you have to use a flash, turn down the power a lot, if the flash is in EV, down about 3 ev to begin with, if power, start with 1/8 power.

5. Use white reflectors to focus light, not colored ones. You don't want to add color to the hair, you want it just the way the hair dresser created it.

6. Use a longer lens. Most of the time I see photographers using a wide angle lens and they are very close to the subject. The problem with a wide angle lens that close is you can get distortion, not a lot but in this case enough to where it could change the creation, or worse make it look funny. I prefer to use a medium telephoto. That way you can get far enough away to see what is going on, but close enough to get good photos of the hair.

7. Last, use a tripod, and watch your focus. The hair must be tack sharp, and that is only achieved with a tripod and manual focus, don't trust auto focus on this.

Well that is it for now, see you all at the workshop this Saturday visit www.jbphotonvideo.com/workshops if you need more info.

Digital workflow for Portraiture

2010-10-28T10:34:00.000-07:00

Yesterday I was helping a friend work on some portraits. During this session I realized that most people are not aware there is a sequence that should be followed when editing digital portraits. And this sequence varies depending on the software you are using.

If you are using a program such as Photoshop or Photo elements or equivalent

Selection

Transformation

Weight

Anomalies

Cropping and straightening

Exposure correction

Color correction

Compositing (if required)

fix eyes

anything else that can't be fixed by other means

Selective corrections

Exposure (dodging/burning)

Blemishes

Skin smoothing

Color correction

In this context selection means selecting the picture or pictures you want to work on, this is the first step, and honestly is probably the hardest. Make sure you spend time here. On a normal portrait shoot count on 1% of the photos you will finish, that is one in 100. Well what that means is that if you get more then that you are doing well, but that is what you count on. I normally run about 5-10%, but remember be critical only work on the best you have, and remember don't delete anything on the camera unless it was a mistake. This past weekend one of the best shots of the workshop was deleted because the person thought the exposure was off.

The compositing step and transformation step are critical. The transformation is where you fix the weight that the digital camera puts on, and it is a lot, but remember do that only to bring the image back to what you saw when you took it. The compositing step is where you fix things like a closed eye, back background, adding jewelry etc. and be sure to do it before you do the selective corrections that way you get the colors to balance correctly.

This order is important and will provide the best overall end product, but understand this sequence changes if you use Lightroom or Aperture. Since this programs are non destructive you need to do as much correction as you can before you have to do destructive editing.

If you use those programs use this sequence

Digital Workflow for Portraiture Lightroom or Aperture

Selection

Cropping and straightening

Exposure correction

Color correction

Transformation (photoshop or equivalent)

Weight

Anomalies

Compositing (if required) (photoshop or equivalent)

fix eyes

backgrounds

anything else that can't be fixed by other means

Selective corrections

Exposure (dodging/burning)

Blemishes

Skin smoothing

Color correction

In future blogs I will discuss more on each one of these areas, but just as a teaser the secret in all of them is be subtle, most portrait corrections should not be noticed by the view, only you should realize that you corrected something, and remember Never but NEVER allow the subject to watch you making corrections. That can be disastrous.

Composition 101 and pet peeve time

2010-10-07T08:28:00.000-07:00

This last weekend I went to a college band reunion, it was great. And as you might suspect I took lots of pictures. So I could be in some of them, not like my college days were i was in almost none, I gave my camera to different people to take pictures of me and my friends from days gone by.

With out exception every picture taken by various people all had the same problem.

They centered the heads in the middle of the picture. That left over half of the picture with open space. I ended up cropping and in some cases a lot, all those pictures. The actual poses were fine, but all were centered.

I am not sure where this comes from but is the single biggest problem I see with beginning photographer and snap shot shooters.

They invariably put the face in the center of the photograph and almost always take a face front shot.

Okay as far as event photography, I do a lot of face front shots, and in event photography where you are trying to capture a moment it is okay. But having so much open space in a picture is not.

When taking pictures even of couples and groups, remember the rule of thirds. Keep the heads in the top third and bodies in the bottom two thirds. That generally gives you your best pictures.

And don’t forget to get in closer. Don’t leave all that open space, make every pixel count. Sure it is okay to leave a little space for cropping to size but no more then 10%. If your camera has a grid turn it on, and use it.

And remember it is okay to turn your camera 90°. If the photograph is vertical then shoot it vertical.

Your good and bad side

2010-09-29T16:27:00.000-07:00

How often have you heard people say, "be sure and sure on my good side", or ask "which is my good side.

I know a lot of photographer have no idea what this means, but there really is a simple way to determine a persons good and bad side. You will be amazed at how much your portrait and general people photography will improve but this simple trick.

Take a good look at your subject, face front.
One side of their face will be distinctly thinner then the other. Very few people have truly symmetrical faces.
The thinner side, and this is regardless of how heavy set the person is, should be the side towards the camera, and the not thinner side, should be away from the camera.
Here is the trick, the heavier set the person, the more you should put the not thinner side in shadow. This will give the illusion of removing the weight that the digital camera put on a person. remember this can be as much as 40 lbs.
Never shoot a person from the not thinner side, it will ad a consider amount of apparent weigh to the person face.
Never shoot a person start on, always have them turn there face, not thin side away from camera.
The more symmetrical the face the flatter the picture can be, that is the difference between the light and dark sides of the face. A person that is totally symmetrical can be shot face front and loot great. the greater the difference, the greater the ratio between light and dark.
The classic difference is 3:1 and while this is idea, it really only works in a few situations. Customize the lighting to the subject.

And that is all there is to it. Try this next time you are taking pictures of people and notice the difference.

Fill flash secret and New blog site for tips and tricks

2010-09-19T05:50:00.000-07:00

I have decided to change the way I blog the tips and tricks blog. Originally I was using iWeb on my mac, which works great, except I have to be on my home mac to do a blog. And I found out very quickly that I needed to be able to update the blog anywhere I happened to have a revelation. :)

Now for today's or week's tip depending on how you look at it.

Fill flash is one of the most frustrating things for a beginning photographer to master. No one ever gives you all the steps necessary to get it right. Well after a lot of trail and error have one for you, a virtually fool proof set of steps that will get you fill flash pictures virtually ever time.

First what is fill flash.

Fill flash is use in out door photography when you have to brighten a face where the contrast ration between the subject and background is so great that you either have the background exposed correctly or the subject, but no both.

What fill flash does not do well, and often you see this, is get rid of shadows on a face. While it will lighten them a lot, but will not get rid of them. So before you use fill flash make sure your lighting on the subjects face is good, the flash will only fill in light.

Here are the steps to virtually perfect

Compose your picture, yes this is the first step make sure your picture is properly composed
Take a picture of just the background, adjust the exposure until it is perfect, just the way you want. Note the f/stop and shutter speed. Be sure the shutter speed is 1/250 or less, for most flashes doe not sync correctly faster then 1/250, some do so you might want to find out what your camera flash does. Now this might get tricky if the background has heavy shadows and highlights, if this is the case exposure in error of the highlights, we can fix shadow a lot easier then blown out highlight.
Manually set the your camera to the exposure for the background.
Turn your flash on, if you have a diffuser use it.

Turn your flash power to either 1/2 or down one ev. Most high end pro cameras have this ability but not all. If your camera can not adjust the flash power you can not use this technique.
And yes leave the flash on the camera it is not necessary to take the flash off camera. but make sure you are within the distance your flash works. In most cases try and be between 6-15 ft from your subject, further away then that your flash will be so weak it won't work, and any closer you stand a good chance of blowing out the high light. If you are within 6 ft, start out at 1/4 power of -2 ev on the flash
Take a picture.
Look at what you just took does the subject look correct, and not like you used a flash. If it is perfect stop,
If not change the flash power up or down until the subject is perfectly exposed and they do not look like they have been shot with a flash. I move my in 1/3 ev increments and generally get it in two or less shots.

At that point start shooting. As long as you don't change lighting conditions you can use these settings.

There you have today's tip, flawless flash photos. Try it and tell me what happens. I used it on a shoot this past week with some great results and when I get them finished. Yes all photos need some editing, but the more you get done in the camera the less time behind the computer.