HDR Food Photography Tutorial— Mar 13, 2011
In my last post I gave a few tips for creating HDR food photography. However, in this one I am going to go step-by-step through the process of creating this HDR image of some chocolate chip cookies.
Disclaimer #1: for good HDR food photography, you’re probably going to need some good software. For the purposes of this tutorial, I am using Photomatix Pro, Topaz DeNoiser and Photoshop CS5. (All of these are available as trial versions, so you can try it out before buying.)
Disclaimer #2: I am still quite the amateur at this. I’ve probably only done 5 or 6 HDR food shots, with varying degrees of success. However, I haven’t been able to find any information on the internets in regards to how one would go about creating HDR images of food, mostly because it seems that there really aren’t a whole lot of people doing HDR food photography, at least that I have been able to find. Granted, a stunning HDR shot of the Grand Canyon as the the sun is peeking out from behind some storm clouds, its life-giving rays piercing the suffocating darkness and flooding the world with light and color is probably a far more compelling subject, as opposed to a stack of cookies. (Hmm… I might have to go and take that shot!)
However, food is a compelling subject, and since it’s the little nuances of texture and the subtle interplay of light and shadow and color that really bring food to life visually, the HDR process can really capture the full range of imagery that food can give to the sight. After all, presentation is a huge part of the culinary experience, and so the way that your food presents itself to the viewer makes a huge impact on its eventual taste. HDR has the advantage of being able to capture a larger range of tones, and as such can really make your food imagery appealing. This tutorial will cover mostly the post-processing of the HDR image, but there are a few preliminary things you can do to make the process easier and turn out with better results.
1. Think through how to compose your scene. It’s not just about the food itself, but also about the space the food is in. Everything from the plate it’s on (if it’s even on a plate) to the colors that surround it will dramtically affect the final image. For this chocolate chip cookies image, my wife decided to stack the cookies and tie them with a pink ribbon, and then offset that with a light green tablecloth. These colors were bright and cheerful, and invoke the feeling of delight that you get when you smell the cookies baking in the oven, that rush of aroma and steam when you pull them out to cool, and that singularly sublime instant when the still warm cookie- nearly falling apart- melts in a subtle and silent symphony of chocolate and sugar, gooey and crumbly on the front of your tongue, while the slight bitterness evens it out for a finish that makes you both savor that moment and want to rush headlong into the next bite. You want your scene to evoke these memories and sensations, so think through it carefully.
2. Look for shadows and a contrast between dark and light. HDR functions best when there is a fairly large range of light and dark, light and shadow. That way, the under-exposed shot can catch the tones that would get blown out in a properly exposed shot, and the over-exposed shot can really pull out the colors and highlights. For this shot, I intentionally kept the background darker and had the light source (the sun through the window) coming at it from the side. This way, it accentuated the light and shadows as much as possible.
3. Keep the ISO as low as possible. HDR adds noise, and with food photography this can be a deal-breaker. Not necessarily because of the noise, but because removing noise leads to an inevitable loss of detail, no matter how awesome your noise removal software is. With landscapes and architecture you can get away with a little loss of detail, but with food it can really sap the life out of your image. If you can keep the ISO under 400, that will probably work out best. I personally like a lot of depth of field, so this usually doesn’t present itself as a problem for me, since I can keep the aperture at something like 2.0 on my 50mm 1.4. (Caution- having the aperture at its maximum will also lead to some blurriness, so try to find something below the maximum that you can shoot with.)
Enough with the setup. There are plenty of regular HDR tutorials that can walk you through how to take great HDR shots, so I won’t belabor this any more.
Ok… you have your exposures. Time to process the heck out of them! For this image, I used 3 exposures: Normal at 0, Underexposed at -2 and Overexposed at +2.
1. The first step is to open Photomatix Pro (or whatever other HDR software you have) and load the bracketed exposures. It’s easy enough- just select the three files and drag them into the window. (You can use either RAW or JPEG- I haven’t noticed any real difference between either.) Or you can Browse manually. Hit OK. After you push OK, Photomatix will give you a couple settings you can have it run. If you didn’t use a tripod (I didn’t for this image) it’s a good idea to have it automatically align the images. If you had a pretty steady hand, it usually does a pretty good job. If it doesn’t work out so well, you can always manually align them in Photoshop, resave and then reload them into Photomatix. The rest of the settings you can probably leave off- my experience is that if you have to use them you are better off fixing the images in Photoshop first, since that’s what it’s for, after all!
2. The first thing you will see is a really ugly looking image with all three exposures blended together. Don’t panic- this is not your final image! That will eventually close and you will see your image and a bunch of settings. This is the tonemapping part. When you first open an image in Photomatix, it will automatically use the default settings. However, if you have processed another image previously without having closed Photomatix, it will apply the last used settings. This image is of the default settings. Actually, it doesn’t look too bad, but we’re going to do a bit of tweaking.
3. The default is pretty good, but I want some more contrast to bring out the texture of the cookies. I boosted the Strength to 92, desaturated it just a little bit, (to 46, 50 being ’0′ essentially) raised the Luminosity to 6.7, and raised the Micro-contrast to 100. The final thing I did was to put the Gamma to .97 to brighten it a little bit. Initially I had even more contrast and bumped up the saturation, but decided to dial it back a little.
This is probably the most important thing about HDR- it is really easy to over-do it. With landscapes and architecture you can achieve some neat effects, but with food you have to have some restraint, or it’s going to start looking radioactive. It is best to think of Photomatix as an intermediate step in the process, rather than as the end-result. Photoshop is good for saturation and levels and exposure, so let it handle those things and let something like Photomatix handle the tonemapping. The best route is to find a look you like an then dial back the settings a little bit. Often it will look really cool in Photomatix, but once you get out of it you are left with this really funky image that all of a sudden doesn’t look so cool. Best to be a little conservative with the tonemapping and let Photoshop make your image shine.
4. Now that you’ve got the settings the way you want them, simply click Process and let it create your HDR image.
5. Next, save the image. It will prompt you to open it in Photoshop. (hint, hint. Even Photomaix wants you to do more with the image…)
6. Now that we’re in Photoshop, the first thing we’re going to do is take care of the noise. Since the ISO was so low, most of the image is pretty good. In fact, the biggest culprit is the black background; it has the white noisiness interspersed. Well, since it’s mostly that area, we’re only going to apply the DeNoiser to that area for starters. Take your Lasso tool and draw around that area. It doesn’t have to be perfect- in fact, it’s better if you give it quite a bit of breathing room…we’ll get to that in a bit.
7. Next, right click (or ctrl-click) and select Layer via Copy. This will non-destructively copy that selection and place it on top of the original image.
8. With that copied layer selected, choose Filters>Topaz Labs>DeNoiser.
9. Once you get into DeNoiser, you will notice that there is by default no noise reduction applied, and you can see how noisy it really is. Since this area doesn’t have much detail, go ahead and crank the noise reduction until it disappears. I actually usually leave just a tiny bit, so don’t go overboard.
10. Let’s go ahead and reduce the noise on the rest of the image. In Photoshop select the original image and send it into DeNoiser. (Filters>Topaz Labs>DeNoiser) Since this is the layer with all the detail, you want to be very judicious with noise reduction. It’s not terribly noisy as it is, so we’re only going to set the noise reduction to .08. There is a also an option for Detail Recovery; for this image it’s probably not necessary.
11. Remember how we overshot the boundaries of the copied layer? If you zoom in, you will see that there is a very nasty looking line of demarcation going on between the layers. Let’s blend these together better. First, create a layer mask on the image. Next, choose the Eraser tool, and create a very large brush with 0% hardness. Now simply erase along the edges until it feathers into the bottom layer. If you mess up, simply undo or delete the mask, create a new one and start over.
12. Since we’re done with our noise reduction and have our layers ready, let’s Merge these together to make it easier to work with. If you’re paranoid about messing up, just duplicate both layers, put them in a group and disable them for now. If everything falls apart, you can go back to them later. Once you do that, merge those layers! Select both layers, then right click (ctrl-click) and select Merge Layers.
13. For the next step we’re going to sharpen the image just a little. Actually, this step probably should have been performed before the noise reduction, but I tried it both ways with nearly identical results. Choose Filter>Sharpen>UnSharp Mask.
14. Like any of the other tools, it’s easy to go overboard with this one. The trick is to keep the radius and tolerance very low, and then raise the amount. I tend to use these settings on almost any image I have.
Tip: Digital cameras actually add a little bit of softness to every image they capture, (or so I am told…) so it’s always a good idea to do at least a small amount of sharpening.
15. Now that we’ve got the tedious stuff out of the way, it’s finally time to make this image shine! First, duplicate the image. (Cmd-J) Next, choose Filter>Blur>Guassian Blur.
16. For this step, we’re going to blur this layer. You can use these settings, or you can play around with your own. A different amount of blur will give you different results.
Tip: If you convert the duplicated layer to a Smart Object first, the Guassian Blur will be editable, and you can more easily manipulate it, whereas if it is a raster image you have to undo and then go through the steps again. Once you are happy with the blur, set the Blending Mode to Overlay, and drop the Opacity to 45%.
Doing this blur/blending will do a couple things: First, it will smooth things out a little bit. Secondly, it will help the colors to really begin to come out. Thirdly, it retains the detail while also giving an oh-so-subtle highlight to the brighter areas, which will help pull your subject out a little bit. (If you want, you can even get your image to look a little bit like a painting.)
17. You may notice that the image is kind of dark, so we’re going to use an Exposure layer adjustment to lighten things up.
Note: be sure to use a Layer Adjustment rather than an Image adjustment. Choose Layer>New Adjustment Layer>Exposure.
18. I am going to sound like a broken record, but be careful with these settings. You can very easily blow out the highlights, wash the whole image out, or any number of other disastrous things. Be very judicious in its use.
At this point, it’s really best to let the Exposure settings and the Overlay opacity of the blurred layer determine each other’s settings. If the image is too dark and you can’t compensate with exposure and gamma without blowing it out or washing it out, lower the opacity on the blurred layer. If it’s too bright, raise the opacity of the blurred layer.
Tip: If you really need to get serious with raising the overal brightness, duplicate that blurred layer again and set the blending mode to Screen. That should brighten things up. Adjust the rest of the settings on the other layers as needed.
19. Almost done! Now we’re going to adjust the colors a little bit. I’ve found that not many people seem to use Selective Color, but I find it extremely useful for small and incremental color adjustments. Choose Layer>New Adjustment Layer>Selective Color. This is all a matter of personal aesthetics, and I won’t bore you with all the settings I used. For me, there were two main adjustments I wanted to make. First, I want the greens to be a little more vibrant and light, so I adjusted the Greens. Secondly, I wanted the pink ribbon to come down in vibrance quite a bit, so I adjusted the Magentas. I believe I also pulled a little bit of the Yellows out, mostly because that’s how I roll. (That’s right.)
20. A few final touches. The blacks in the background were bugging me a little, so I took a very soft, large dark grey brush and painted over the area. I then set the Blending Mode to Multiply and dropped the opacity a little bit, so that it didn’t become a distraction.
21. The last thing is a very subtle vignette. First, either create a pure white rectangle shape that fills the whole canvas, or paint pure white onto a new layer. I personally like to create a white shape that is 105% the size of the canvas, give or take. Then choose Filter>Lens Correction. (This is how you do it in CS5. In previous versions I believe it’s under Distortion)
22. I usually put a pretty heavy Vignette on in this step. We’re going to be lowering the opacity quite a bit, so feel free to over-do it this one time!
23. The final step is to lower the opacity to 22% or so and set the blending mode to Divide. This will cause the darkness of the vignette to turn bright, and will brighten the edges, which will add some light to the whole scene. If you want to go dark, simply change the blending mode to Multiply. Again, be judicious with your vignettes!
That’s it! You now have an HDR image of some very tasty looking cookies!