Single Image HDR: RAW to TIFF to HDR Efex Pro 2

Why do we do bracket shots and HDR, the normal answer is the dynamic range of the scene is too great for the camera sensor.  Thus, a series of exposures that covered the whole dynamic rage are shot then merge into one single picture and after tone mapping process, the new HDR image covers the complete dynamic range of the scene.  The image from the tone mapping process tends to be very flat, various HDR softwares then has build-in different tools to boost up the image.  Sometimes when it is done too much, it resulted in the very unnatural HDR-look.

Using the bracket shots has several advantages: 1. to avoid the clipping of over-exposed high-light; 2. to avoid the excess noise of the under-exposed shadow.  However, in some occasion, with the modern digital camera, the sensor has already very wide dynamic range that can capture the whole dynamic range of the scene.  How to get the most out from the this dynamic range needs some manipulations.

Direct import one single RAW file into HDR Efex Pro 2 in order to use the tools available to boost up the image has several disadvantages: 1. we lost the Auto Distortion and Auto Lateral Color Aberration;  2. we lost the Exposure Compensation.  These functions are normally built in some of the RAW conversion program, Capture NX2, ACR, etc.  Not all the third party software that has all the distortion profile for all the Nikon camera and lens combination.  The same is for the color fringe (Lateral Color Aberration).  The newer Nikon camera and the Capture NX2 software can do it easily and perfectly, no one knows better than Nikon of their own cameras and lenses.  The functions only work when the RAW file is converted into JPEG or TIFF.  Of cause, if you are not a Nikon user, you will have to relay on Photoshop, etc to do it for you.

As for the exposure compensation, the RAW file has way more information in it than the simple histogram can tell.  When the RAW file of the above image was directly import into HDR Efex Pro 2, There is no detail in the brighter area of the sky.  It was almost impossible to bring it out using global image adjustment.  I know the information is in there, since if I move the exposure compensation to -2 in the Capture NX2, I got every detail out from the sky.  Here is my workflow for this image:

1. Open the RAW file in Capture NX2, change Picture Control to Neutral.  In the Advance tap, move all the slider to lowest: Sharpening 0, Contrast -3, Brightness -3, Saturation -3, Hue -3; 2.  Leave Auto Distortion and Auto Lateral Color Aberration on.

2. Change Exposure Compensation to -2, -1, 0, +1, +2, at each setting save a file in TIFF.

3. Open all 5 TIFF files in HDR Efex Pro 2, merge them into HDR then do the adjustments.

4. Smart Sharpen the HDR image in Photoshop.

5. Adjust the orientation and prospective of the image in PTLens.

6.  Final crop and adjustment in Capture NX2.

Although the ideal condition would be to have bracket shots for the original scene, but the moving canoe made it not practical for merging the shots together without "ghost effect".

RAW vs. TIFF in HDR

Should we load RAW files directly into HDR software: Photomatix Pro, HDR Expose 2 or HDR Efex Pro 2 for creating HDR image or should we convert the RAW files into TIFF files before input into the software.  The argument for RAW is that the software gets the most original complete information to start with thus should give us a HDR image that contains the most information.  However, that may be too much of a good thing.

For us who use Capture NX2 to open our RAW file, we do not even see how the real original RAW file looks like.  Nikon has various default setting in the program and what we have set up (Picture Control, D-Lighting, Auto Distortion, Auto Lateral Color Aberration, etc.) in the camera is also directly brought into the program.  The RAW file open in Adobe's ACR then gives more of the realistic look of the RAW file.

The following HDR images were created from five shots of five different exposures: +2, +1, 0, -1, -2 taken with my D700 and 28-300mm f/3.5-5.6 ED VR lens at Mosaic Canyon of Death Valley.

 This image was created from RAW files using HDR Efex Pro 2 as plug-in of Photoshop with the "Dark" default setting.  It looks pretty good until I look closely.

 I can see the lateral color fringes in the cloud but most noticeably between the rock and the sky.  I do not see them normally when I open the RAW file in Capture NX2, because they are corrected by the default setting.

I then converted the RAW files into TIFF using Capture NX2.  The white balance was as shoot (automatic), Picture Control was as shoot (Neutral), Auto Distortion on and Auto Lateral Color Aberration on.  They were imported into HDR Efex Pro 2 and the HDR image was created using the "Dark" default setting.  I can see this image is a bit more contrasty and also warmer than the previously one from RAW files, most likely due to the Picture Control setting.

The most notable change is the red and blue lateral color fringes between the rock and the sky now disappeared, although I can still see the left over shadow.  Those in the cloud still can be seen that is not too much of a problem.  HDR Efex Pro 2 does have a Chromatic Aberration correction function at the HDR merging step, but the automatic built-in function in Capture NX2 certainly works easier.

May be I should use a better lens next time such as 24-70mm f/2.8 that may give me less chromatic aberration.

Nikon V1,10mm 2.8, HDR Efex Pro 2: 緣麵屋enishi

This is the only Taiwan outlet of the famous Japanese Noodle shop, 緣麵屋 (Enishi).  We went once a month during the past several months.  The usual wait is one hour.  The people on the left were having the noodle in front of the kitchen, the people on the right were waiting patiently.

Here is the kitchen, we can see how the noodles are made.

It is not me who was eating the noodle, but I got the seat next to him after one hour wait.

Not only the noodle is cooked right in front of us, but each piece of meat is fired before it is served.

This is what we came for 沾麵.  The soup is hot but the noodle was rinsed in cold water to keep it "Al Dente" or in Taiwanese "Q".  We dip the noodle in the soup before we eat it, very nice indeed.

HDR Expose 2 vs. HDR Efex Pro 2

 I have detailed before in one of my blog (http://graceandbradley.blogspot.tw/2012/07/her-for-interior.html), how I generated the above image using HDR Expose 2 from five images: -2, -1, 0, 1, 2, and white balance the resulting HDR image in Photoshop using one more image lighted by a flash.  HDR Expose 2 has a different way of generating HDR image, which is quite good and interesting.  Although I have further adjusted the image in Capture NX2, it seems the resulting image still lack a little "punch" that we normally see in HDR image.  In a way, it is very good and looks like a normal photograph!

I did the whole thing again now using HDR Efex Pro in Photoshop CS6.  I also did the white balance for the interior as described before using an image lighted by flash.  This time I have been able to generate an image with more "Punch", more detail contrast, better reflection from the floor, etc.

Photomatix Pro vs. HDR Efex Pro 2: Zabriskie Point, Death Valley

 I have often troubled by, although the histogram show that I have all the information in the RAW file, how can I actually get the image I want without extensive adjustment using Photoshop or Capture NX2, which sometimes is beyond my ability.  The above image was taken at Zabriskie Point, Death Valley in the early morning last January.  The image was under-exposed one stop and the histogram looked perfect.

This is a HDR image generated by Photomatix Pro from three images: -3, -2, -1 stop exposure.  Now, it is quite nice.  Since "-1" image already contains all the information, why do I have to use HDR to get the image as I want.

This image is generated using one single "-1" RAW file by HDR Efex Pro2, used as plug-in in Photoshop CS6.  Almost as good as the HDR image from Photomatix Pro.

HDR Efex Pro2 allows me to play with all the HDR adjustments (Tone Mapping, etc.) for just one RAW file.  The result is quite good, as long as the original RAW file already contains all the information, no clipped high-lights or shadows.

Nikon V1 with FT1 Adapter and 80-400mm, 105mm Micro Lens

Nikon V1 with a 400mm lens makes a combination equivalent of 1080mm.  The image size is about 10.1 MP.  At 300 dpi, it can make a print of 13x18.6 inches.  In Taiwan, the largest size print that I can get from commercial Fuji Frontier printer is only 8x12 inches.  Thus, the image size is more than enough if I do not do any cropping.  The caveat is Nikon 80-400mm VR lens is not a G lens thus can only be used as manual focus lens on V1.  The ISO was set at ASA100 in order to get best image quality.  The combination of camera and lens is quite light weight, but since I need to manual focus, I set them up on my Gitzo GT2541 tripod with G2272M fluid head.  Most shots were done at 400mm (1080mm) thus the VR function of the lens must be shut off.  The focusing aid in the eye piece was quite useful.  When using FT1 on V1, only the center focusing point is working.  That was not too much of a problem.

After toyed around the 80-400mm for a while, I changed to 105mm Micro lens.  This combination is equivalent to 283mm.  The 105mm Micro lens is a G lens, thus auto focusing was working on V1.  Compare to 80-400mm, the 105mm V1 set up is so light weight and easy to use, I shot quite a few images off-hand.  The following images were all shot with 105mm lens except the last one.

When I was ready to pack the camera, I saw what most photographers would hope to see in the nature, something to add interest to the image.  In the distance, a tiny frog was sitting by a water lilly.  Thus I took out and set up the whole thing with the 80-400mm lens again and got the following image.  It is worth the trouble.

The image quality was incredible for these pictures, as good as if not better than my D700.  When I took out my older images that were taken with D700 at the same location.  The size of the images actually was much smaller than the ones from V1, because I had to do a lot of cropping.  Using the 80-400mm and 105mm on V1 given me such a long reach, I hardly need to do any cropping of the pictures.  All of them are around 8x12 inches at 300 dpi.

One famous photographer once said, to make an appealing photographs, we have to think like a bug, getting close and into the scenery.  The long reach of the V1 with 80-400mm and 105mm gave me just that.  Although I was on the solid ground by the pond, looking through the eyepiece, it gave me the feeling that I was a little bug flying in between the flowers,  Although, I would not get too close to the frog!

The take home massage from this exercise is that V1 is capable of delivering very high quality image.  I have all the V1 lenses, they can give images as good as smart phone, which is quite useful.  However, in order to really realize V1's potential, we have to use the lowest ISO possible and FT1 adapter with the much better FX or DX format lenses.