Archive

Archive for the ‘Tutorial’ Category

BBC Life – Behind the Scenes

November 14, 2010 Leave a comment

The Life series on Discovery Channel (aka BBC Life) offers some of the best time-lapse footage I’ve ever seen.  That’s why I was so excited to find this ten minute video explaining in detail how the most ambitious time-lapse tracking shot was captured.  Anyone who enjoys time-lapse and wants to learn to do it better should see how a team of experts can take two years to capture one shot.

Windows Live Movie Maker – Time Lapse Tutorial

April 12, 2010 3 comments


Last year, Microsoft came out with an updated version of its free movie making software now called Windows Live Movie Maker.  With the introduction of Windows 7, all new PCs should include Windows Live Movie Maker.  If you’re using an older version of Windows, the new Movie Maker should still work (download link).

The good news for time lapse enthusiasts is that this new version simplifies the process and no longer requires the workaround that I described in a previous post.

Getting Started

Begin by importing all of the photos in your time lapse series into Movie Maker.  I prefer to number them sequentially, but I believe Movie Maker is less picky than VirtualDub in this regard.

Change the Duration

Select all of the photos you’ve just imported and click on the “Edit” tab at the top of the screen.  The only available selection is “Duration,” which is set to 5.0 seconds by default.  For the video I’ve used in this tutorial, I’ve set the interval to 16 fps or 0.06 seconds when rounded down.  Here are some other times for easy reference:

16 fps = 0.06

20 fps = 0.05

24 fps = 0.04

30 fps = 0.03

Saving the Video

Go back to the “Home” tab and scroll through the “Sharing” window until you find the proper settings.  I photographed this carnival ride in a 16:9 aspect ratio with over 1920×1080 pixels, so I’m able to select the “High-Definition (1080p)” option.

After you’ve saved the video, you can open it in Movie Maker and edit it the same way as any other video.

Here’s the final result of the video I made during this tutorial.  The people in the foreground toward the end are picking up the pieces of someone’s mobile phone after it flew off the ride.

Share

Categories: Software, Tutorial Tags: ,

Star Trails Photoshop Action

March 29, 2010 8 comments

Music: Stars by Rivers

(Attention Photographers:  If you’re interested in “stacking” photos, simply use this action as described and select one of the final photos in the series, such as the one below.  Also note that this was made using a 6 megapixel point-and-shoot camera, so your photos will probably look better.)


Click here for the free “Star Trails” Photoshop action

Can you find the North Star?  I found the Big Dipper and Cassiopea and then just tried to point my lens in the right direction.  I got lucky because I almost edged out my focal point.  I find it amazing how little the North Star moves.

Why I Made this Action

Last week I posted a video of stars crossing the night sky to show the benefits of post production dark frame subtraction.  This week’s video uses similar photos and a simple action I made using Photoshop CS3.  This is my first time posting a Photoshop action, so I’d appreciate any feedback.  The small gaps in the trails come from a short delay between photographs.

I came up with this technique after watching jcmegabyte’s star trails videos on Youtube.  His videos are better than mine but I hope to refine my star trails in the future.  Notice the small patch of purple in the upper-left corner that’s indicative of post production DFS.  Without this post editing, his star trails would pulse instead of being constant.

My Photoshop action didn’t take long to make and is very simple, so you might want to edit it until you’re happy with the results.  It features a slow fade-away because I liked the result better than persistent lines.  You can see both types of trails in jcmegabyte’s videos.

Install the File

To begin, download the file to anywhere on your hard drive.  Run Photoshop and open the actions box.  When you click on the drop-down menu within the actions box, you should see an option to load an action.  Click this and find the star trails action.

Process Your Photos

Copy your stars photographs into a new folder.  This isn’t completely necessary, but I find it helps in case I accidentally process the original photos.  Open the first photo in Photoshop.  Select all, copy, and then close the photo (Ctrl+A,Ctrl+C, Ctrl+W).  With this first file in memory, you can begin batching your photos.

In Photoshop, select “File,” “Automate,” and “Batch…”  Find your stars folder and pick the star trails action.  Click “Ok” and then go make a sandwich.  Because this is a simple action, it shouldn’t take too long to finish.  You can open your output folder while you wait if you want to make sure the action is working.  Your first photo should look the same but the subsequent ones should feature growing light trails.  Now turn the photos into a video as you would normally.

I’m hoping to use this same action to create other kinds of trail videos.  I’ll post the results if anything turns out.  Good luck!

Update:  Erik Røstad made a much better video than mine using this script.  Good job!   http://www.flickr.com/photos/12328282@N04/4869251405/

I’ve also used this script for my video of planes landing at SFO.

Share

Categories: My videos, Software, Tutorial Tags: , ,

Dark Frame Subtraction

When I began taking long-exposure photos of the night sky last year, I noticed that any shot over three seconds required an equal amount of time before the camera was able to take another shot.  This was a mild annoyance at three seconds, but it became a real problem as my exposure times approached one minute.  At first, I thought the camera was running some complex compression algorithm, but a little research on the internet told me that it was due to Dark Frame Subtraction (or DFS).

What Is DFS?

My technical knowledge of camera sensors is limited, but here’s the basis of my understanding:  When a camera takes a long exposure, the sensor begins to generate enough heat to affect the image.  This creates a purple haze and random hot pixels, as in the photo below.

To compensate for the unwanted noise, the camera takes a second photo of equal length without actually exposing the sensor.  The image produced contains only the unwanted noise and is called a “dark frame.”  Here’s what a 50 second dark frame looks like from my camera.  Click on the image to see all of the hot pixels.

After both photos are taken, the camera automatically subtracts the noise in the dark frame from the original image.  The resulting image looks like this:

How Does DFS Affect Time-Lapse?

If you snap a few pictures a night, in-camera DFS might benefit you.  If you want to make time-lapse videos, DFS is prohibitively long.  Star photography with a point-and-shoot might require one minute exposures, so a five second video could take over three hours to photograph.  Without DFS, that same video would take one and a half hours.  Even if your battery lasts for hours of constant use, the sky doesn’t stay completely dark all night.

The Solution

The good news about DFS is the same dark frame can be used on all of your photos as long as the exposure time stays the same and the camera doesn’t change temperature too much.

Begin by acclimating your camera to the outside temperature.  Then, turn off DFS.  You can check your manual to see if this is possible, or you can set the “Dark Frame Subtraction” in CHDK to “Off” if you have it installed on your camera.  You can find this setting from the main menu under “RAW parameters,” but you don’t have to shoot in RAW.

Next, choose how long you want your exposure to last.  Once this is determined, take a dark photo by leaving your lens cap on or by otherwise completely blocking light from reaching the sensor.  Now that you have your dark frame, take your photos as normal.

Manually Subtracting the Dark Frame

I went through a bunch of websites and trial software before I found this website with a simple Photoshop action that does the trick.  The Photoshop action is about halfway down the page.  The website also has a good overview of DFS that’s probably worth reading if you have any further questions.

Finally, batch process all of your photos using the DFS action and you should have a pretty good series of photos that took only half as long to photograph.

Webcam Capture – CamStudio

March 10, 2010 2 comments

I recently posted about making a webcam time lapse from a photo feed found on the internet.  The method I used in that post involved using a piece of free software that was designed for the task.  Today’s post is about an alternative process for when the first method doesn’t work.

First, take a quick look at the video above (webcam found here) and you’ll see the major advantage and disadvantage of this new method.  The time frame for the video is about fifteen minutes in real time.  Using the previous method, this would mean the entire time lapse video would be just under one second because the minimum capture interval is one minute.

The downside is that there is noticeable freezing in the video.  Even the best camera streams will occasionally lag due to heavy traffic.  Because this new method turns the webcam photos directly into video, it’s difficult to edit the repeating frames without video editing software.

The Software

This new method uses a free screen capture software that you might use for making a video tutorial of something on your computer.  It’s designed to capture real-time video but can be adjusted for our purposes.  The software is called CamStudio (download link).

The Webcam Feed

If you’re using this method, it’s probably because the action is too fast for the first method or because you’re having difficulty capturing the image with the first method.  If the problem is the latter, you need to determine whether you’ll need additional software.  If the feed refreshes at a regular interval without problem, you can skip this step.  If the feed is static or freezes, you will need to find software to reload the image.

I use the Chrome browser extension, ChromeReload (download link), for my webcam feeds.  It allows me to refresh a specific tab at whatever interval I choose.  There are similar extensions available for Firefox, IE, and probably other browsers.  Search for “reload” or “refresh” and you should find them.  Once installed, pick an appropriate interval so that the image can refresh as often as needed but not so often that the image capture software picks up unnecessary reloads.

CamStudio Settings

By default, CamStudio is set to record real-time streaming video.  To use it for time lapse, we’ll have to change the settings.  Select “Options” and then “Video Options.”  Pick a compression filter at the top.  For this video, I’ve used the default codec.  Next, uncheck “Auto Adjust” so that you can change the settings under “Framerates.”

Now comes the tricky part.  Find how often the webcam refreshes.  Some webcams refresh every second, others can take minutes.  Even if the webcam refreshes frequently, watch to see how often the image fails to refresh.  If the picture changes every few seconds but isn’t reliable, choose a longer interval.  For this video, I’ve selected five seconds, or 5000 milliseconds.  Next, choose a playback rate.  Remember that a rate of at least 16fps helps smooth out the action.

Capture the Feed

Once you’ve accepted the settings, click on the red record button in CamStudio and drag your cursor over the area you want to record.  The webcam should now be framed by green corners that flash when an image is captured.

Troubleshooting

There are a number of problems with this method that make the earlier method preferable.  The most obvious is that you can’t use your computer while the software is recording.  You’ll also need to disable your screensaver and any windows that might pop up and ruin your video (IM, Windows updates, etc.).  If you want, you can disable recording the cursor under “Options” to make things a bit easier. It’s also a good idea to turn off your monitor if you plan to record for long periods without a screen saver.

Start small before trying to record a video.  My first few videos were ruined by unexpected problems or by my own actions.  Any great webcam finds or videos?  Post links in the comments.

Share

Categories: Software, Tutorial Tags:

Make a Webcam Time Lapse

February 8, 2010 3 comments

Since I began making time lapse videos, I’ve occasionally come across a high quality webcam feed on the internet and thought that it would make a great time lapse video. It’s basically the same setup that I use; a camera set to record some interesting and dynamic subject at timed intervals. Sometimes the webcam even has its own time lapse video, but usually it doesn’t. Over the last month, I’ve been experimenting with different webcam capture software that’s available on the internet. Today’s post is a quick tutorial that uses only free software.

Download webcam capture software

I’ve been using a program appropriately called Web Cam Time Lapse (download link). Web Cam Time Lapse (WCTL) is a no-frills way to save an image from a URL address. Even though the user interface is simple and there aren’t many options, I’ve found the program to be fairly stable even when running multiple captures for days at a time.

Find a good webcam

There are two types of webcams–live streaming video of low quality and jpegs or gifs refreshed occasionally.  For high quality time lapse videos, we’ll want to ignore the streaming video and locate large photos.  Most still images are refreshed between five seconds and fifteen minutes.  There are plenty of websites devoted to webcams, but here are three of my favorites:

Webcam Cruise – a thorough collection broken down by country and subject.

123Cam – cameras with thumbnail previews by geographical region and subject.

Webcam Galore – thumbnail previews by region and theme.

Begin recording

Once you’ve found a good webcam, you’ll have to find the image URL.  This is usually different from the page URL and can often be found by right-clicking the image and selecting “Copy image URL,” or something similar.  You might have to dig a little deeper and view the source code to find the source of the embedded webcam feed.  Search the source code for “jpg” or “webcam.”  These two methods usually work but not always.

Next, open WCTL and click “Add,” at the bottom of the screen.  Pick a name so you’ll remember what you’re recording, paste the image URL, select a save folder, and choose an interval.  For some reason, WCTL doesn’t allow intervals of less than a minute, so you’ll have to pick a subject that can be recorded over hours or days.  Many webcams say how frequently they’re updated, otherwise, spend a few minutes watching the screen to find out how often the image is updated.  Click “Test URL,” to make sure the image loads properly before saving.  Here’s what my settings for Mt. Fuji looked like:

Cleaning up the images

Webcams take a lot of bandwidth, so even the larger images have small file sizes.  Because of this, you can record thousands of images without significantly filling your hard drive.  The 304 images of Mt. Fuji for this video only amounted to 58.2 MB.  The bad news is that bandwidth limitations mean that the image won’t always refresh on schedule and that some of the images will be partial or in some way unusable.

I begin preparing my images for video by first going through and eliminating any distracting photos.  Then, I run Duplicate Cleaner (download link) to quickly delete duplicate files.  I find this much faster than hunting down multiple duplicates.

Because I like to process my photos into time lapse videos with VirtualDub, I have to make sure that the images are saved as jpegs and are in numerical order.  Most of the time, I don’t have to format my webcam images because they’re usually in the jpeg format, but occasionally I’ll find that they’re gif images or that VirtualDub has some problem identifying the format.  When this happens, I quickly batch convert the files into jpegs using Format Factory (download link).  Finally, I select all photos in the set, right-click and select, “rename.”  Now the photos are ready to be processed.

For step-by-step processing directions, please read my post on using VirtualDub for time lapse.

Share

Make Video Wallpaper for Your iPhone

January 16, 2010 Leave a comment

If you have a jailbroken iPhone or iPod Touch, you might have seen an app for running video wallpaper.  Or maybe you’re running VWallpaper already and want to know a free way to make your own video wallpaper.  This post is about turning photos into a time lapse wallpaper, but most of the steps are the same for other video.

Before I begin, I should point out that moving these videos to your iPod takes a bit of skill and patience.  I’m only going to explain how to make the videos in the proper format because this isn’t an iPhone hacks blog.  However, if you’re interested in jailbreaking your iPhone or iPod Touch and installing these videos, you’ll have to take the following steps:

1 – Jailbreak your iPhone

2 – Install Cydia or Rock

3 – Install Winterboard

4 – Install VWallpaper app

5 – Install an SSH program (and please change your iPhone’s root password!)

6 – Transfer videos to the proper folders

There are plenty of tutorials online to show you how to do these steps, but I couldn’t find one authoritative website and so I’m not providing links here.  I used bits and pieces from over a dozen sites when I first did this, so feel free to provide links to your own tutorial or a favorite how-to website in the comments.

Take the pictures

Like all time lapse videos, you’ll have to begin with a series of photographs.  The only differences are that the camera should be tilted 45 degrees to a portrait orientation and the resolution can be much less than normal.  All current iPods play video in 480 x 320 pixels.

Your video will be playing in the background every time you start your iPhone, so you want to make something that isn’t too distracting.  Take your photos at a smaller interval to create slower motion and use a higher framerate to avoid choppiness.  Try to make a video that’s at least thirty seconds so it will be less distracting when it loops.  The good news is that you can fit many more photos on your memory card because of the small resolution.

Edit

Remove the occasional bird or plane using a photo editing program such as Photoshop or Gimp.

Make the video

I use VirtualDub for most of my time lapse videos.  In this case, it works especially well because it allows me to crop my video to the desired dimensions.  Select “Video” from the top menu and then choose “Filters…”  Choose the resize filter and adjust the aspect ratio to 320 x 480 by selecting “Crop to aspect ratio” from near the bottom and changing the absolute pixels.  Remember that the numbers are reversed from the normal aspect ratio because the photos were taken in a portrait orientation.

Click “OK,” select your framerate, and save the file as an AVI.  Don’t worry about compressing the file or changing the file type, we’ll change that in the next step.

Convert

This final step requires another piece of software.  For video file conversion, I prefer the free Format Factory because it’s easy to use.  Download it here.

Drag your AVI file into Format Factory.  Choose “All to MP4” and click “OK.”  Don’t use the default iPhone settings or your video will be converted to a horizontal format and won’t look right on your iPhone.

Now use the SSH program to upload the video to your iPhone and the file name should appear in the VWallpaper app.  Good luck!

Share

%d bloggers like this: