Fujifilm Finepix Real 3D W3 - ISO settings and image noise

You probably know that in AUTO mode, the film sensitivity setting is by default AUTO (800), which means that the Fuji W3 can bump the film sensitivity to a value as high as 800 ISO if it thinks it's a good idea. Although these potentially high sensitivity settings can be the difference between being able to take the picture or not under low light conditions, they may produce an undesirable effect called mottling (also known as grain or noise). The "lowest" (max) ISO setting in AUTO shooting mode is AUTO (400), which means the ISO setting the Fuji W3 will use for any given picture can be anywhere between 100 and 400 ISO. Unfortunately, you don't have control over it although it is likely higher ISO settings will be used under low light and lower ISO settings will be used under natural (bright enough) light.

If you don't mind a little bit of tinkering to get sharper pictures under low light condition, it makes sense to me to choose a low ISO setting, say 100, and then play with the shutter speed and/or aperture to make the shot possible (or just see if the shot is possible with whatever the W3 chooses for speed and aperture in P shooting mode), upping the ISO setting to the next one in case the shot is not possible.

The Fuji W3 has 3 "manual" modes to choose from: A (you manually select the aperture), M (you manually select the aperture and shutter speed), and P (automatic aperture and speed selection). In each of those "not so automatic" modes, you can select an ISO setting as low as 100 ISO. If you don't want to do too much tinkering with shutter speeds and/or apertures, the P (PROGRAM) mode is the one to use in order to have direct access to lower ISO settings for the film sensitivity.

To see the effects of the ISO settings in low light condition, I took two 3D pictures (with flash and on a tripod) of my reel-to-reel tape deck from about 2 meters away using the P shooting mode of the Fuji W3, first with the ISO sensitivity setting at "AUTO (800)", second with the ISO sensitivity setting at "100". The MPOs were then loaded into SPM and I went 100% magnification of what my focus was on: the VU-meters. Again, the only difference between the 2 shots is the ISO setting, nothing else.


AUTO (800) ISO mode


100 ISO mode

Night and day! For indoor shots, you clearly want to stay away from the default AUTO ISO modes (unless you don't mind mottling), especially knowing that you can get away with lowering the ISO setting all the way down to 100 without much of a problem, well, assuming you have enough light and a tripod at your disposal. Note that even at 100 ISO, the sharpness is really not that great for a 10 mega pixel camera but it is quite acceptable for a relatively cheap 3D camera (my little Pentax Optio S7 takes much sharper pictures but, to be fair, it doesn't house 2 CCDs).

When there's a lot of light available (outdoor shots), using the AUTO (800) ISO mode is not an issue since the Fuji W3 will probably select a setting of 100 ISO to take the picture (you can see the ISO setting a picture was taken in the lower left of the LCD screen along with other vital statistics like shutter speed and aperture).

OpenCV 2.3.1 and Microsoft Visual C++ 2010 Express

OpenCV is an open source Computer Vision software package. It includes some stereo calibration and matching functionality, and is therefore of great interest to anyone doing 3D.

I am gonna go through all the steps needed to install the 32 bit debug version of OpenCV 2.3.1 using the pre-compiled libraries and have it running on a Windows 7 64 bit machine with Microsoft Visual C++ 2010 Express under the hood.

Get the OpenCV 2.3.1 Windows Superpack and install it. I am gonna assume it's installed in C:\Program Files\opencv.

You're also gonna need the so-called TBB debug libraries because they are not included in OpenCV 2.3.1 (if you have already installed and setup OpenCV 2.3.1 and you're getting an error message about tbb_debug.dll missing when you start debugging your code, that's your problem). Get the windows TBB binaries and place the tbb30_20110427oss directory under C:\Program Files\opencv\build\common (for example). At some point in the future, OpenCV may come with the debug TBB and all the TBB stuff added here may not be needed anymore.

After having created a new Visual C++ 2010 project (make sure it's a Win32 Console Application), right-click on the name of the project and go into Properties.

Click on Configuration Properties->VC++ Directories, and edit the Include Directories. Add:
C:\Program Files\opencv\build\common\tbb30_20110427oss\include
C:\Program Files\opencv\build\include\opencv2
C:\Program Files\opencv\build\include\opencv
C:\Program Files\opencv\build\include

Click on Configuration Properties->VC++ Directories, and edit the Library Directories. Add:
C:\Program Files\opencv\build\common\tbb30_20110427oss\lib\ia32\vc10
C:\Program Files\opencv\build\x86\vc10\lib

Click on Configuration Properties->Linker->Input, and edit the Additional Dependencies. Add:
opencv_calib3d231d.lib
opencv_contrib231d.lib
opencv_core231d.lib
opencv_features2d231d.lib
opencv_flann231d.lib
opencv_gpu231d.lib
opencv_haartraining_engined.lib
opencv_highgui231d.lib
opencv_imgproc231d.lib
opencv_legacy231d.lib
opencv_ml231d.lib
opencv_objdetect231d.lib
opencv_ts231d.lib
opencv_video231d.lib
tbb_debug.lib
Here, I have put all the OpenCV libraries but you certainly don't need to do that, as you can just add the ones you need for your application.

At this point, you should be able to link your code. The only remaining issue is to let the computer know where the dll libraries are by adding the proper paths to the PATH environment variable.

Click on Configuration Properties->Debugging, and edit the Environment.
Type PATH=C:\Program Files\opencv\build\x86\vc10\bin;C:\Program Files\opencv\build\common\tbb30_20110427oss\bin\ia32\vc10 in the window.

You may also change the PATH outside of Visual Studio by going into the Control Panel->System and Security->System->Advanced System Settings->Environment Variables, and changing the PATH there.

Note about Visual C++ 2008 users: vc10 refers to Visual C++ 2010, vc9 refers to Visual C++ 2008. OpenCV 2.3.1 provides both builds.

Note about building a 64 bit project: Microsoft Visual C++ 2010 Express doesn't come with 64 bit support out of the box. You need to install Windows SDK (Software Development Kit) to have 64 bit support. After you have made Windows SDK the default in Properties->General->Platform Toolset, you should be able to go into the Configuration Manager and switch the project over to 64 bit. I haven't tried but there's no reason you shouldn't be able to link against the 64 bit OpenCV 2.3.1 libraries if you want to.

Fujifilm Finepix Real 3D W3 - Macro Stereo

If you look closely at the manual for the Fuji 3D W3 camera, it will tell you that that the focusing range is from 60 cm (2.0 ft.) to infinity. It will also tell you that the recommended range for taking 3D pictures is actually from 1.3 m (4.3 ft.) to infinity.

The Fuji 3D W3 has a 3D Macro mode that enables you to take pictures from 38 cm to 70 cm (1.2 ft. to 2.3 ft.) [wide angle] and from 1.1 m to 2.3 m (3.6 ft. to 7.5 ft.) [telephoto]. Doing macro stereo photography with a camera having a 75 mm lens separation (the Fuji 3D W3) is not a good idea. Why? It's because, unlike our eyes, the lenses do not converge toward the object to be photographed, and you end up with a left and right images that can not be fused by our eyes. But then, how do you do macro stereo (hypo stereo) with the Fuji 3D W3 camera? Well, you don't ... unless you have an optical device to reduce the lens separation (see Cyclopital3D) or a slide bar (more on this later) and a tripod.

You probably know that your Fuji 3D W3 camera has an Advanced 3D mode with one of the two options (Individual Shutter 3D) enabling you to take (with the left lens) the left shot and right shot (or the other way around) in succession, the stereo pair being saved as a MPO file. Now, if you also select Macro mode, you are gonna be able to take pictures of objects from 8 cm to 80 cm (0.3 ft. to 2.6 ft.) [wide angle] and from 60 cm to 3 m (2.0 ft. to 9.8 ft.) [telephoto]. What I am getting at is that if you use a slide bar to take the left and right photo in Advanced Individual Shutter 3D mode, you're gonna be able to do proper macro stereo photography.

The million dollar question is then: By how much should you slide the camera to take the right picture (after having taken the left one)? Well, all I can say with certainty is that Cyclopital3D uses a 25 mm lens separation in their macro close-up attachment. There's probably a magic formula that can give you the correct stereo base for any distance but I certainly don't have it.

Note that no matter what Fuji says, you should not take a picture of something that's too close. 10 to 12 inches is probably the closest you want to go, since it corresponds to the distance of distinct vision for most people.

Manfrotto makes some nice slide bars (used to be sold under the Bogen brand) but you could certainly make your own slide bar if you are of the DIY kind.

Of course, this slide bar business kinda defeats the purpose of having a stereo camera since all you need is a "mono" camera. On the plus side, having a slide bar also enables you to take intermediate frames between the left and right shots for proper lenticular imaging.

Depth Limit of 3D Scenes

Back in the days when the stereo Realist format ruled the 3D world (it was not that long ago), slide mounting for stereo projection was considered the norm. Even if you never intended to show your slides on a big silver screen, you kinda took the pictures (with your clunky old stereo camera) and mounted the slides with the stereo projection principles in mind.

When mounting stereo slides for stereo projection, you had the choice between three types of mask. The windows/apertures in each type of mask have the same center but the size of the windows are different (thus changing the position of the stereo window).

Here are the three types of mask available:
- DISTANT MOUNTS: Cover subject matters from 7 feet to infinity.
- MEDIUM MOUNTS: From 4 to 20 feet.
- CLOSE-UP MOUNTS: From 2.5 to 6 feet.

The main reason to have those masks is that there is no need for projector adjustment when projecting a series of slides because the centers never change (to be honest, I don't fully understand this but it's not really the point here). The other reason (and that's the one we're interested in) is that, when projected, the distances between corresponding points (on the left and right images) are always from 0 to 2.5 inches apart, which guarantees comfortable viewing for the audience. Those masks basically limit the depth, the distance between the foreground and the background in your pictures. In general, when a stereo pair is projected onto a screen, you don't want to have points separated by more than 2.5 inches (the distance between our eyes) because, otherwise, it forces our eyes to diverge (this is referred to as divergence and you really don't want it). If you consider the distant mounts, the points at 7 feet are positioned at the stereo window (the silver screen) and the points at infinity are 2.5 inches apart, which is perfect for comfortable viewing. Same idea for the medium and close-up mounts. Remember that convergence is usually never a problem for our eyes but divergence always is.

These principles are in my opinion still valid today with 3D digital cameras like the Fuji Finepix 3D W3 camera. For example, I don't think it's a very good idea to take a picture with very near objects and distant ones as the convergence range is likely to be too large.

Of course, rules are always meant to be broken.

Stereo Matching - Depth Maps from Stereo Pairs

3D lenticulars made from just the left and right images suffer from a very narrow viewing angle, which makes getting the 3D effect difficult. Intermediate frames makes the viewing a much more pleasant experience. They also make the GIF animations (wiggles) more fluid and therefore more effective. Understandably, there's a lot of interest in generating depth maps from stereo pairs in the 3D community.

If you have a look at the academic papers that deal with depth maps or stereo matching (just using google), you will soon be overwhelmed by the amount of stuff that's being written over the years. The question is: how do you know which method works best?

There's a paper published in 2002 titled "A Taxonomy and Evaluation of Dense Two-Frame Stereo Correspondence Algorithms" by Daniel Scharstein and Richard Szeliski that does a thorough review of stereo matching methods. "Taxonomy" means that they have identified "building blocks" that are common in most algorithms. Doing that always makes things easier since it shows that a lot of algorithms are actually quite similar in the way they are designed. "Dense" means that the matching is done at the pixel level. "Dense Two-Frame Stereo Correspondence" is basically another way of saying "depth map". The authors are kind enough to provide the source code that was used to do the testing (which you can use on your own stereo pairs). If you are interested in trying this out yourself, go to vision.middlebury.com, download the source code, the scripts and the images. You need to compile the code to be able to use the program, called StereoMatcher, which runs on a script that you can customize. On a windows pc, you're probably gonna need Microsoft Visual Studio (the Express version is free). On a linux box, it's a bit easier since the gnu C++ compiler is already there as part of the install. Once you have created the executable and read the scripts that are used to run the provided test cases, it's not that hard to figure out how you can run your own stuff. The only thing is that you're gonna have to save or convert your stereo images in the ppm format (see gimp or photoshop for that). You also have to be well aware that the stereo pairs you can feed to those stereo matching algorithms need to be rectified, that is, for any point in the scene, the two projected points (on the left and right images) must be on the same line, or scan line (see Stereo Matching Rectified Geometry).

I took the liberty to run StereoMatch on the (overused) test image called "Tsukuba":


Left image.


Right image.


True disparities.

Here are the depth maps for each available algorithm:


SSD (Sum of Squared Differences).


SO (Scanline Optimization).


SAD (Sum of Absolute Differences).


SA (Simulated Annealing).


GC (Graph Cut).


DP (Dynamic Programming).

GC (Graph Cut) appears to be the best performing algorithm for this particular benchmark stereo pair which, to be honest, has about nothing to do with the "real world" (piecewise planar objects perpendicular to the optical axes are not too commonly found in nature). If this subject fascinates you, check Stereo Matching - Local Methods, Stereo Matching - Global Methods, and Stereo Matching - Variational Methods for more info on various stereo matching methodologies.

I, Ugo Capeto, have implemented some software to generate depth maps from stereo pairs. It's available at the 3D Software page right here on this blog and it's free to download.

Fujifilm Finepix Real 3D W3 - Focus

There seem to be some issues with focusing on the Fujifilm Finepix Real 3D W3 digital camera. It would seem logical to me that both lenses focus on the same thing but I don't think they do or, more exactly, it looks like they don't. It seems to me the right lens is correctly focused while the left lens is (usually) not. Have no idea whether it's mechanical or software related and whether it is supposed to be that way (under certain conditions maybe). It can clearly be seen at normal (100%) size when adjusting the stereo view in StereoPhoto Maker. Not saying that it's a big deal when you end up printing 6x4 prints but it's definitely annoying to see something out of focus when it's pretty clear that it shouldn't be. Below is a good example of what I am talking about:



This is the left image of the pair. I remember vividly focusing on the large stone in the center of the foreground (confirmed when looking at the parallax in the foreground picture below). Not the greatest focusing point but the camera didn't complain when taking the picture so I can only assume it was good enough.



This is a close-up of the foreground in StereoPhoto Maker (I am using the stereo adjustment view at normal size). You can clearly see that the right picture is focused (expected, because that's where I actually focused the shot) while the left picture is blurry (not good).



This is a close-up of the background. Now, the right picture is blurry (expected) while the left picture is more in focus.

I've chosen this picture because there's quite a bit of distance between the foreground and background and it really shows where the left lens is focusing (it's focusing on something much farther than the right lens). With other pictures, when there's not much distance in the shot between foreground and background, it kinda seems that the left lens is blurry all over (probably because it's focused on a point who knows where).

This seems kinda strange that the two lenses are not focusing on the same point but, for the sake of discussion, let's assume that the two lenses do indeed focus independently. Then, this whole thing could very well be an auto-focus problem when the object that's focused on is not super well defined against the background, especially in low light condition just like in the picture above (I've seen the same thing happen in bright light though). I have no problem with the auto-focus not being able to lock but then, why does the camera say it can focus (focus light is green and camera beeps) and, more importantly, why is the difference in focus distances between the two lenses so large? And this leads to another question: If the two lenses focus on possibly two different things, how can the camera's software figure out the parallax in auto mode? The problem is that you have no idea you're out of focus until you go into StereoPhoto Maker and check the parallax at normal size. If the camera can't focus, no problem, just let me know about it and I can take the appropriate action (hopefully).

This is the wiggle (Piku-Piku) version from Start 3D (kinda like what StereoTracer from triaxes.com does). Obviously, if you only intend to show wiggles from your Fuji W3 mpo files, then being slightly out of focus on one lens may be the least of your worries.

Well, since then, I have taken a few more pictures in better light conditions and using better focus points, and it seems that the auto-focus behaves better. The two images are usually equally blurry under 100% magnification. Conclusion: the lenses on the Fuji w3 are not very good, to put it nicely.

Lenticulars - Interlacing with SuperFlip

SuperFlip is a free lenticular software available from vuethru. If you can't read the help file on your computer, I've taken the liberty to convert the help file into a pdf file that you can download here: SuperFlip manual.

We are gonna go through all the steps to create and print a 3D interlaced image from a left and right image (taken, for example, by the Fujifilm Finepix Real 3D W3 digital camera).

Start SuperFlip and add your left and right images (left first, right second). They should be at the correct aspect ratio.

Click on "Process" to get into the "Process Setup Dialog" box.

"Line Screen (LPI)": enter the result of your pitch test for your printer, paper, lenticular lens combination.

"Width" and "Height": enter the width or height you want (the other dimension will be calculated automatically if "Preserve Aspect Ratio" is checked) or you can click on "Calculate Output Size from Input Size" if the images are already at the correct size. If you plan on putting alignment marks, make sure you account for that because they are added when you actually print (see below).

Uncheck "Resample Output with banding eliminator".

Check "Alignment Marks". Click on "Alignment Mark Options" to check or modify the width of the alignment bands. You must take those into account when you set the width/height of the interlaced image. It's easier to mount the interlaced image to the lenticular lens when you have alignment marks but it's not necessary.

Uncheck "Color Separations".

"Number Images": 2.

Check "Preserve Aspect Ratio".

Uncheck "Use Separators".

"Interlace Direction": check "Vertical".

Click on "Process". You should be taken to the "Resolution Enhancement" window. Choose the highest starred enhancement level and click "OK Enhance!".

Save the tif interlaced image and if it looks ok, click "Print" in SuperFlip (select some kind of photo paper and the best print quality). If you are printing on 6x4 photo paper, you probably will have to select "6x4 borderless" and then, the printer may automatically select "landscape" depending in which direction the 6x4 photo paper is fed to the printer (that's the case for my HP Officejet printer). I don't particularly like the idea of printing interlaced pictures in landscape mode so I don't use 6x4 photo paper but it may work for you. In all cases, make sure you let the paper dry before doing any kind of manipulation.



This is what the interlaced image should look like with the alignment marks all around. The stereo window has been set at about the level of the 1st "$5 Each" sign (you can easily tell because it's not fuzzy like the rest of the picture). This is the Brattle Book Shop in Boston, one of America's oldest and largest used book shops.
All that's left now is to mount the printed interlaced image onto the lenticular lens.

Lenticulars - Pitch Test with SuperFlip

Lenticulars are in my opinion the best way to view 3D pictures (taken for example with a Fujifilm Finepix Real 3D W3 camera). You don't need that much to make your own lenticulars, mainly an inkjet printer and some lenticular lenses (with adhesive attached).

As you probably know, lenticular lenses have a certain number of lenticules per inch (LPI). For 3D comfortable home viewing, it is common to use 60 LPI. You get the 3D effect by printing an interlaced version of the left and right images and placing the lenticular lens on top. The left eye will only see the left image and the right eye will only see the right image thanks to the interlacing and the geometry of the lenticules. In the real world, lenticulars have intermediate images, not just the left and right images but that has no effect on the pitch test, of course.

Before you can even start to think about making 3D lenticulars using an inkjet printer, you need to calibrate the printer, paper and lenticular lens you are planning to actually use. This calibration will determine the proper LPI (lenticules or lines per inch) for your setup.

The first step is to print a line screen test using SuperFlip from the fine folks at vuethru:
- click on Utilities->Print Line Screen Test...
- input a proper "starting lines per inch"
- select "just doublet"
- print
In the printer "Properties...", select something that looks like photo paper, set the print quality to best and use portrait orientation. Of course, you have to print on good photo paper! You are also supposed to use the same paper for the pitch test and the actual interlacing of your image sequences because the pitch test is affected not only by the printer itself but also the paper.

What's a proper starting lines per inch (LPI)? Well, personally, I choose the LPI of my lenticular lens minus a few cents. For example, if my lens is 60 LPI, I am gonna start at 59.96 and go from there depending on the results given by that first print.

It's a bit hard to explain how to actually do the pitch test once the line screen test has been printed but the following video explains the process quite well (It's actually quite simple.):


The key is to put down the lenticular lens at an angle and then rotate it slowly until you see the solid bands. In my case (60 lpi lenticular and hp officejet pro 8000 printer), the lines 60.02 to 60.06 produce the best flips so 60.04 is probably the correct lpi to use in the interlacing process. The whole process is a bit harder when you have lenticular lenses on the smaller side, like a 4x6 lens for example. Bigger lenses, like a 8x10, produce more accurate results.

Check the Creating and Using a Pitch Test by microlens for another interpretation of the pitch test.

Fujifilm Finepix Real 3D W3 - Stereo Window and Parallax

When you view a 3D picture in the autostereoscopic LCD screen of the Fujifilm 3D W3 digital camera, the screen itself is the stereo window. The term "stereo window" is used extensively in 3D stereo photography and it's a good idea to know a little bit about it. It's a real big deal when mounting Realist format (or other format) slides, especially for stereo projection.

When you take a picture and "auto parallax control" is on (it's the default), as you press the shutter button half-way and focus on whatever object is in the middle of the frame, the parallax adjustment is made (automatically) such that this object is positioned right at the stereo window. Keep in mind that you can always focus (and set the parallax) and then re-frame while the shutter button is half-way pressed (it's called "focus lock").

When you adjust the parallax while viewing your 3D picture on the LCD screen, depending in which direction you press the button, things seem to be moving either towards you (possibly past the stereo window) or away from you. The question is: What's the proper parallax for my 3D pictures?

Usually, and as a starting point, it's a good idea to have the nearest object right at the stereo window. In stereo slide mounting, it's referred to as "near point" mounting. But what makes 3D stereo photography so spectacular is to have objects "pop" in front of the stereo window. Keep in mind though that you absolutely cannot have something in front of the stereo window that's clipped by the frame or screen. It's a cardinal rule of stereo photography that should never be violated as our poor brain cannot process too well the idea of an object being clipped by a window that appears to be actually behind that object.

When you play with the parallax button, you are simply moving the two "frames" closer or further apart. In stereo slide mounting, it's like moving the two film chips horizontally in the mount. You can see the left and right images slide (either in or out) as you keep the parallax button pushed. Note that when the two images coincide at a given object/point, that object/point should appear to be in the plane of the LCD screen, aka the stereo window, when the image is rendered in 3D. As the manual says, it's never a good idea to have too much parallax, so don't slide those images too much apart!

If you change the parallax of a 3D picture you are currently viewing and want to "save" the parallax settings, you need to do the following:



So, technically, it's not a "save" but rather a "copy", which is a good thing since you will always have the original parallax settings for that 3D picture (until you actually delete it).

Note that the parallax can also be adjusted off camera, in StereoPhoto Maker, for example.

Reel 3-D Catalog 1998

This is the 1998 catalog from "Reel 3-D Enterprises", once the premier supplier of everything stereo (3-D). These days, "Reel 3-D" only sells to dealers and is not online anymore as a business. The catalog has lots of basic info about stereo photography and it shows all the goodies that were available to amateur stereo photographers back in 1998 (when digital stereo photography was still just a dream). Not really sure where you can get what's shown in the catalog these days but a pretty safe bet would be ebay (and the dealers that "Reel 3-D" is talking about).


Reel 3-D catalog cover


Dear 3-D Friends


Do you have a Question about 3-D Photography?


Easy to Use 3-D Mounting Supplies


Easy to Use Stereo (3-D) Slide Mounts


Stereo (3-D) Viewers


Stereo Slide Storage Pages and Stereo (3-D) Viewers


Books about Stereo (3-D) Photography


Books about Stereo (3-D) Photography


Books about Stereo (3-D) Photography


Books about Stereo (3-D) Photography


Books about Stereo (3-D) Photography and View-Master Reels


Books about Stereo (3-D) Photography


Books about Stereo (3-D) Photography


Argus 3-D Camera and Print Viewer Outfit


Books about Stereo (3-D) Photography


Anaglyph 3-D Glasses


Projector Lamps and Polarizing Filters for 3-D Projection


Polarized 3-D Glasses


Halogen Bulb for Battery Stereo Viewers


Silver Screens for 3-D and Gepe Slide Mounts for 2x2 3-D Pairs


Franka 3-D slide Adapters, Twin 35mm Viewer, and 35mm Slide Frames



Twin 35mm Slide Viewers


Nissin Electronic Flash for Stereo Cameras


Stereo Instruction Manuals


Table of Contents for Reel 3-D Catalog

Fujifilm Finepix Real 3D W3 - Viewing/Sharing Movies

Recording movies is quite easy with the Fujifilm Finepix Real 3D W3, it's the viewing and sharing that's a bit difficult when you don't have a 3D TV.

MyFinePix Studio, the software that came with the camera (which you can get on the Fujifilm website as well), is supposed to allow you to view and edit your movies but, to be honest, I could not get it to work on my HD movies (1280x720 16:9).

No worries because there is StereoMovie Maker that's available for free. After having installed StereoMovie Maker, you will probably be told that it cannot open any of your movies. It's because you don't have the motion jpeg (mjpg) codec installed on your computer. Google "motion jpeg codec" and you should be able to grab one, hopefully for free. Personally, I am using "motion jpeg codec 3.2.4" from free-codecs.com but I have the bad feeling it's just a demo. There's also a mjpg codec from morgan-multimedia.com but that one is definitely a demo.

Update on mjpeg codecs: forget about all those demo mjpeg codecs! All you need is ffdshow. A big thanks to David Kesner of the fuji3d yahoo group and dddphotography. Install the ffdshow codec, open the "VFW configuration" screen, select the "Decoder" tab, select "Codecs" in the list on the left, search for the "MJPEG" format in the list on the right, and change its "Decoder" to "libavcodec". The ffdshow codec has been around for a while, is free, and is highly recommended for all kinds of video formats, not just mjpeg.

StereoMovie Maker is real easy to use and there's a pretty good html manual that comes with it so there's no need here to delve into what each button does.

StereoMovie Maker has very basic editing capabilities as it appears that it can trim a movie (actually, it doesn't cut anything, it just moves the start and/or end markers, meaning that it only trims when you save) but I don't think it can merge 2 movies together.

It's cool and all to view/edit your movie with StereoMovie Maker but it would be even nicer to put those movies on the web somewhere. That's where youtube comes in with its 3D capabilities (it's pretty recent). So, let's look at how to save movies for youtube and the whole uploading and setting up on youtube.

Before you save your movie, make sure to swap the frames so that the Right frame appears on the left and vice versa. It's because the default in youtube for side-by-side viewing is Right-Left (cross-eyed). Technically, you don't really have to switch frames but you have to remember to change the "stereoscopic video options" when editing your video, once it's uploaded (I think). Now, you're ready to save the movie you've just edited: use the "side-by-side" option, pick a video compressor (divx or xvid, for example), configure the compressor so that you have a nice balance between image quality and file size, and press ok.

You should be ready to upload to youtube now. It's just like any other video in terms of uploading but, in order for youtube to know it's actually a 3D video, you need to add yt3d:enable=true and yt3d:aspect=16:9 (if 16:9) to your tags in the video description. That's pretty much all there is to it. When you view the video you have just uploaded, there should be a 3D icon with a drop-down menu where you can choose how you want to view the video (anaglyph, parallel, cross-eyed, and more). There should also be a settings button right next to the 3D button in order to select the definition (remember that the Fujifilm Finepix Real 3D W3 digital camera can shoot movies in 16:9 high definition).


Video I took of backyard birds with my Fujifilm Real 3D W3 camera (featuring a nice stereo window violation). You probably have noticed that there's no 3D button in the embed player. That's because you need to press play in order for the 3D button to show up.

Fujifilm Finepix Real 3D W3 - Viewing/Sharing Pictures

So, you have been taking pictures with your Fuji Finepix Real 3D W3 camera but you have this lingering question on your mind: how do I view them off camera and share them with the rest of the world? You're probably aware of the software that should have come with the camera called MyFinePix Studio. You can load the MPO pictures you have taken and view them as either anaglyphs or LineByLine (interlaced by row) but there's apparently no way to save the pictures (besides hitting PrtSc and pasting the screen dump into Gimp or Photoshop) and upload them on the web.


Anaglyph obtained from MPO 3D file using MyFinePix Studio via PrtSc and Gimp.

Fortunately, there is stereoscopy.com downloads where you can find all kinds of useful 3D programs and utilities. Let's look into StereoPhoto Maker and see if it can do what we really want: view and save as anaglyph, side-by-side, and possibly wiggle animated gif. Well, it sure can do all that and much more so it's highly recommended in lieu of MyFinePix Studio (sorry Fuji).


Anaglyph obtained from MPO 3D file using StereoPhoto Maker. Same result but much easier than MyFinePix Studio.

I personally don't like side-by-side stereo whether it's parallel or cross-eyed on a computer monitor. Can't say I am a big fan of the wiggle animation gif either because, in a lot of cases, it just does not work. My favorite method for displaying and sharing 3D images on a computer monitor remains the good old red-cyan anaglyph but it assumes the person viewing the anaglyph has a pair of glasses (they're easy enough to get though). Yeah, I wish there was a better way to view 3D on the usual computer screen but, unless autostereoscopic LCD displays become the norm, it's gonna be tough for quite some time to share 3D on the internet.

It's nice to display on a computer but sometimes you want something more tangible, something you can hold in your hands. Well, you can print an anaglyph but, in my opinion, it's better to print side-by-side stereo views which can then be viewed with a stereo slide card viewer (you know, the kind of stuff that was so popular in Victorian times). But, in my opinion, the best way to have "hardcopies" of your 3D pictures is with 3D lenticular imaging, mainly because you don't need any glasses or viewer to appreciate the 3D stereo effect, and that's a big plus.

Yeah, I don't have the FujiFilm FinePix REAL 3D V1 viewer (costs a real pretty penny) and I don't have a 3D TV either (might get one at some point with a Blu-ray 3D player), so those viewing options are not discussed here.

Fujifilm Finepix Real 3D W3 Digital Camera

The Fujifilm Finepix Real 3D W3 digital camera is a digital camera that can take 2D and 3D still pictures as well as record 2D and 3D movies. Just like the Nintendo 3DS, you don't need glasses to preview your 3D pictures or movies because the LCD screen is autostereoscopic.





Below are snapshots of the Fuji Real 3D W3 manual pages which deal with formats for both still picture and movie recording.



3D pictures can be saved in MPO (multi-picture) format or in "MPO+JPEG" format (a jpeg of the left picture is saved in addition to the MPO 3D picture).



Size and aspect ratio options for taking pictures. Default is L (large) 16x9.



You have the option to save pictures either in fine (low compression) or normal (high compression) quality. Default is N (normal).



You can record 3D movies in 1280x720 (HD quality), 640x480, or 320x240 formats.