Outdoor Photography and Videography

Tobi Wulff Photography

Home / Tags / howto / Essays

Documentary Shooting: Interviews

We are finally at the stage where we conduct and film the interviews for our TWALK documentary. After a few days of acquiring and testing gear and then test-driving the setup, we now have something that looks great (not perfect but better than many other documentaries I've seen) and is very mobile and easy to set up for a 1-2 person crew.

But first, here is the latest trailer for our project with a bunch of little interviews (sound bites) we took during the event:

Gear

Here is a quick rundown of the gear that I think is required to get a well-lit interview with solid, good sound:

  • 1-2 cameras that can be matched in terms of exposure and color profiles (I struggle with matching them perfectly since one is a Blackmagic and the other one a GH4 but a color chart like the X-Rite ColorChecker Passport (de) helps tremendously to fixup color space and color temperature issues with one click in Davinci Rsolve)
  • A sound recorder because most camera pre-amps are really bad and don't offer enough inputs or control such as rotary dials and limiters - I use the excellent entry-level Zoom H5 Four-Track Portable Recorder (de)
  • Good microphones: it took me a while to accept that it is important to spend a good chunk of your budget on sound equipment - maybe not as much as on cameras and lenses but still enough to get at least one middle to top of the line microphone. I like to always have a backup option in case one microphone experiences cloth rustling, room echo, or simply stops working, so I usually mix a lavalier/"clip-on" mic like the Rode LAVALIER Condenser Microphone and a shotgun mic: the Sennheiser ME 64 Cardioid Condenser Capsule Head is great for indoor interviews, for outdoors think about the more directional ME66.
  • 2-3 LED lights with big enough batteries to work for at least an hour: depending on your exact kit and the conditions of the space you are conducting the interview in, you might actually need more lighting devices (see the section on the detailed setup below). All lights' color temperatures should match as closely as possible, unless you go for an extra warm fill light, or an extra cold hair light, etc. Mine are all daylight balanced (around 5500K) and have a good CRI value (95+). The entry-level smaller Apature Amaran LED lights work a treat indoors if you darken the room anyway, for more power I would look at the Aputure HR672W or Apature's more professional lineup.
  • A reflector and/or diffusion panel (or softbox): your light, in particular your key light, will always look better the softer it is, unless you are deliberately going for a harsher look. To achieve this you have to send it through a (the bigger the better!) Softbox (de) or white semi-translucent panel, or bounce it off a 5-in-1 Collapsible Multi-Disc Light Reflector (de) .
  • Enough Light Stands (de) for all those lights. Also a Microphone Boom Stand (de) for the shotgun mic.

I'm sure this sounds like a lot of gear for just one interview, and if it is not possible to buy or borrow all those things, an interview in natural light might be a better option (I'd still recommend having at least one reflector). However, for the style I was going for I needed a dark space and all those lights to make it work.

Setup

When you do the setup, make sure it's ready when the interviewee arrives. You don't want to bore them with 20 minutes of fiddling with lights and cameras. This means you need a stand-in, e.g. your interviewer, producer, grip, etc.

The first thing to do is find a good spot for the interviewee in regards to space (all the equipment you'll have to set up around them) and background. The background should be not too boring (avoid a plain wall but patterns are good) but also not distracting (no harsh light, no movement). This obviously depends on the subject of the interview and how you want to associate the interviewee with the environment. Typical examples are athletes and their trophies, scientists and their awards and books, and so on. In the end, it all comes down to personal preference and the style you are going for.

Next I will set up my key light. This is the main light that illuminates the subject. The typical rule is to have to about 45 degrees from the camera on the side of the camera the interviewer is sitting. You need a good light stand because the key light also has to be above the subjects head (again, about 25-40 degrees). Eye level can work as well but avoid a light source below eye level because it will make them like starring in a horror or zombie movie. Personally, I actually like my key light to be 90 degrees to the camera line and at about eye level. This gives the face a very strong divided look between the lit and unlit half right down the nose. I use a big diffusion panel to soften the light as much as possible. We will take care of the unlit half of the face in the next paragraph.

At the moment, the shadow line is too strong. Even with some natural light, this is what often makes cheaper productions fall short because the subject will have fairly strong shadows on their face. Our next job is to get rid of them. This is achieved with a fill light which can either simply be a reflector that bounces natural light or the key light, or an extra, smaller LED light than shines some light at eye level at the subject's face. Even at the lowest setting, it does immediately get rid of any hard shadows. The fill light should be closer to the camera to make sure what the camera sees is illuminated enough to give a pleasing and soft look. There are examples of professional interviews well done without any (noticeable) fill for effect, for example my personal favorite In the Shadow of the Moon (just google images for it plus "interview").

We've dealt with the absolutely necessary lights, now for the bonus: add a hair light behind the subject, slightly to the side of the fill light and just out of sight from the cameras. This will add a light onto one side of the hair, head, and shoulder of the subject and make them stand out more against the background. The background at this point should be as dark as possible as to be not too distracting. If it is receiving too much light spill, try to flag the lights with so-called barn doors/french flags or some cardboard. Another option is to move the subject further from the background and use the fall-off of the light's intensity to make it appear darker.

Now it is time for the subject (or "talent", i.e. interviewee) to take a seat. Set up your shotgun mic so that it points down at their chest, and either hide the lavalier mic under their clothing or position it low enough (but not below their sternum) so that it is out of the shot. I use a very tight picture that usually only shows their head and maybe the shoulders so getting the microphones out of the shots isn't very hard. Now do a sound check, get the cameras rolling, use a clapper board or do something similar (clap your hands or snip your fingers in front of the cameras) and you are ready to go for the interview. It can also be a good idea to use a color chart at this point to get a good reference with the final lighting setup.

This post has got long enough so I will talk about any workflow, interview tips and how to edit an interview in Davinci Resolve in another one. Just remember to do all your backups first thing after the interview is over!

Please use the comment section below or head over to Google+ or Twitter @tobiaswulff to discuss this article or any of my photography and videography work. My Flickr, 500px and Vimeo pages also provide some space to leave comments and keep up to date with my portfolio. Lastly, if you want to get updates on future blog posts, please subscribe to my RSS feed. I plan to publish a new article every Wednesday.

Challenges Adjusting Time in JPGs and RAWs

I recently found myself in a situation where I had to adjust the date and time on all my photos from overseas, JPEGs and RAWs. Lesson learned: it is much easier to remember (if you do) to change the setting on the camera when you are switching time zones.

JPEGs

It is fairly easy to change EXIF and IPTC metadata in JPEGs because pretty much all the tools support it. Apart from just writing data directly, most of the tools (and luckily there are many options) I looked at also allow for automatic and intelligent data/time adjustments, so you only have to specify the offset in minutes or hours or whatever unit you require, and it will set the date and time accordingly. In the end this means the choice of the specific program comes down to personal preference. In Linux, there are several options, both for the CLI and as a GUI.

In digikam, the time adjustment can be found in the batch processing editor. To get there, select the photos you want to adjust, then hit B. You can select the individual destinations for the adjust times and I usually go with all the EXIF tags and the digikam timestamp (IPTC wasn't set when the files came out of camera). After the adjustments have been made to the files, it is important to re-read the photos back into digikam. To do this, select the photos again, then go to the menu Item and click "Reread Metadata".

On the CLI, the job is much easier in my opinion (as is often the case). To get a console for the album you want to edit, right-click on the album in the album view (sorry, this technique can't work when you want to edit photos based on tags or other filters; in that case you have to use the GUI method described above) and select "Open in Terminal". Now we can use (if installed) several programs to fix the date/time:

- exiftool: Does not have a date/time adjust option so for JPEGs I would not use it
- exiv2: Can read and write all the tags in JPEGs (and other formats, but not all RAW formats, see below) and has a handy date/time adjust function: "exiv2 ad -a -10 *.JPG" will subtract 10 hours from the EXIF timestamps. It can also be used to rename the files according to the timestamp ("exiv2 mv") but I like to use digikam for that (it can make filenames unique automatically if necessary).
- jhead: Functionality around timestamps and renaming is similar to exiv2 so it comes down to personal taste and specific use cases: "jhead -ta-10:00 *.JPG" will subtract 10 hours.

RAWs

This is were things get a bit trickier and depending on your camera's RAW format some of the programs will not work, e.g. exiv2 supports ORF but not RW2, and the GUI alternatives (digikam or UFRaw) didn't contain any options to write arbitrary metadata. exiv2 can work on some formats as described above (which is nice because it is the shortest and simplest command) but failed to write RW2 (Panasonic). What did work was exiftool. One slight quirk is that while exiftool displays pretty field names when you print all the metadata within a file (no arugments, just "exiftool file.RW2"), it requires the arguments for time adjustment to be the technical, compressed names of all the individual fields that you want to write, so: exiftool -"ModifyDate"-=10 -"DateTimeOriginal"-=10 -"CreateDate"-=10 *.RW2

I hope someone else who is on the search for the right tool to adjust their photos' metadata will find this information useful. I'll keep it as a reference for the future because I'm sure I will forget to set my camera to the right time zone again.

Please use the comment section below or head over to Google+ or Twitter @tobiaswulff to discuss this article or any of my photography and videography work. My Flickr, 500px and Vimeo pages also provide some space to leave comments and keep up to date with my portfolio. Lastly, if you want to get updates on future blog posts, please subscribe to my RSS feed. I plan to publish a new article every Wednesday.

Home-built LED Panel for Video

Over the last weekend plus a few days I put together an inexpensive DIY LED panel that is so bright it will blind your interview subjects but doesn't break the bank - at all.

I won't explain all the steps and parts for this build because someone else has already done a much better job than me: DIY Perks on Youtube (you'll want to watch most of his builds once you discover the channel!). Watch the video from start to finish, then check your electronics and tools box for the few parts that are needed. I had to get a roll of 300 LED lights, a voltage regulator (or PWM controller if you're not worried about flickering lights but for video work a voltage dimmer is recommended), some MDF and some woodworking and power tools I didn't have (but they will serve me for many years to come). The rest is really simple and if you decide to skip painting it can probably be all done in one afternoon.

The colours here look a bit funny because I had the whitebalance on my camera set to my working light which is a tungsten light bulb whereas the LEDs are clearly much cooler. However, even though it was a cheap roll of LEDs (about $15), it doesn't have that nasty green shift that cheap LEDs are kind of known for. Its colour probably isn't perfect but I'm pretty happy with it. DIY Perks linked to some equally cheap LEDs that he measured at something like 95 CRI (which is really good) but I had already ordered some from another seller on Aliexpress.

The brightness is amazing and while I haven't tested its limits in daylight yet it's definitely too bright to point at your talent (say at an interview) at full power. Note that you need some fairly thick wires to hook everything up until it gets fanned out to the individual LED strips - I chose AWG16 but 18 is probably alright as well as long as you don't power more than 300-400 LEDs. With bigger panels your wire thickness requirements (and the ones for your power source) will obviously go up. I power the panel with 18V from a universal laptop power supply (I would prefer around 14V but the power supply chooses 18V with the 2.1 mm DC connector at the end) but a good LiIon or LiPo battery does work as well.

The box that houses the regulator as well as a switch and a potentiometer is actually a clear Raspberry Pi case that I had lying around. Due to the thick wire it was a bit of a squeeze and a solder joint popped off at one point due to the strain on a cable so better give yourself plenty of room for the electronics. This will also help with dissipating heat but with only 300 LEDs the heat sinks on my regulator don't get too hot.

One section where I deviated from the instructions in the video above is the dimmer circuit. While DIY Perks uses a multi-K Ohm resistor and a 22K pot, I used a 10K pot and no fixed resistor. I get the perfect voltage range out of it (I think my trim pot on the regulator is actually turned fully clockwise), meaning 7V (when all the LEDs just turn on) to about 11V which is as high as I want to go without compromising the lifetime of the LEDs.

To mount it on tripods etc I sandwiched a coldshoe with 1/4" thread between two pieces of MDF glued together, than screwed (and glued) that onto the back of the panel. Because the panel isn't very heavy, this is an extremely sturdy connection. Just watch the video to see exactly how it's done. I've since applied the same technique to make some adjustable mounting "bricks" for my motorised timelapse slider.

Please use the comment section below or head over to Google+ or Twitter @tobiaswulff to discuss this article or any of my photography and videography work. My Flickr, 500px and Vimeo pages also provide some space to leave comments and keep up to date with my portfolio. Lastly, if you want to get updates on future blog posts, please subscribe to my RSS feed. I plan to publish a new article every Wednesday.

Fast Editing and Color Grading with a Gaming Mouse

There are many control surfaces out there to help with editing and colour grading but they are mostly geared towards professionals and are very expensive. Examples are surfaces from Tangent and Blackmagic.

North RouteburnSunset in the North Routeburn, Aspiring National Park, NZ. Cropping, exposure and contrast adjustments, monochrome processing in Darktable.

For amateurs and enthusiasts there are multiple cheaper options. I've written about my DIY controller for Davinci before but there are also consumer devices that can greatly speed up editing and grading. Because I have to use a mouse anyway (for lack of a complete, fully-features control surface), the Logitech G700s gamging mouse(de) is one of my favourite tools. Its main feature are the four thumb buttons on the left side that can be assigned arbitrary actions or shortcuts through the Logitech software. The configuration is stored in the mouse itself so once set up the Logitech software is no longer needed. This means the settings will work the same on any computer or any operating system. The mouse wheel is kind of special on this mouse because it can be pushed left or right (very useful for scrolling on a timeline), and it can be put in a free spinning mode with the button next to it. This is useful for browsing websites or scrolling through long documents such as the Davinci Manual (PDF).

Edit

For editing I use three of the buttons to switch between Davinci's three edit modes (pointer [shortcut A], trim [T], razor blade [B]). The fourth button is used to toggle snapping on or off since I constantly find myself switching between shifting clips around (in this case I want them to stick to the next clip so that there is no gap) and making fine adjustments to the length of clips or exact cuts. Another good option for those four buttons are the clip modes Insert, Overwrite, Replace, and Place on Top.

The top buttons aren't used as heavily because they are a bit awkward to reach while holding the mouse. At the moment I have the three buttons on the top left set up to set in and out points and to toggle video/audio linking for the selected clip. I almost find the I and O keys on the keyboard easier to reach but I also miss them sometimes when I don't look down. The two buttons in the center of the mouse switch through the profiles and turn the free-spinning mouse wheel on or off.

Colour

The most common action when colour correcting and grading, and the central piece of Davinci's colour page, are the nodes. So I set up the four buttons on the side of the mouse to quickly add serial, parallel, layer nodes, and also to add a serial node before the currently selected one. I haven't found a specific use for the three top buttons yet as there are so many possible shortcuts but none of them are used as often as handling nodes. Maybe I'll go for undo/redo, or for handling clip versions or the gallery.

Summary

Using a gaming mouse together with a few keyboard shortcuts or a simple control surface (I'm looking forward to see what the Tangent Ripple can do) can greatly speed up your editing and colour grading work. If you've only been using a normal three button mouse so far I highly recommend giving a gaming mouse with 7-10 additional buttons a go.

Please use the comment section below or head over to Google+ or Twitter @tobiaswulff to discuss this article or any of my photography and videography work. My Flickr, 500px and Vimeo pages also provide some space to leave comments and keep up to date with my portfolio. Lastly, if you want to get updates on future blog posts, please subscribe to my RSS feed. I plan to publish a new article every Wednesday.

RAW Timelapse Workflow with Darktable and Davinci Resolve

I shot a new timelapse in the mountains, this time exclusively recording all the frames in full resolution and RAW (unlike my previous outdoor timelapse). It was recorded with the Olympus OM-D E-M1 (de) and Olympus 12-40mm F2.8 PRO (de) lens.

Out in the field

Basic outdoor timelapse 101: all manual settings, that is ISO, white balance, aperture, shutter speed. White balance was obviously daylight and I kept the ISO at its minimum (200). Shutter speed should be set to something "video-like". Video and film cameras usually use something called a 180 degree shutter which essentially means that the shutter speed is 1/(2 x frame rate). So for a 24 fps video that means 1/48 or (because photo cameras usually don't offer this setting) 1/50. Anything faster than that runs the risk of making the timelapse feel jittery and too sharp. For fast movements, like people or clouds, I like to go even slower and aim for something like 1/20 - 1/40. This gives the video a more dreamy and pleasing look.

I record every frame in RAW. I like to store JPEGs as well so I can generate a quick timelapse when I get home without having to go through the RAW workflow (described below) first.

To do the actual timelapse recording, there are several option depending on your circumstances and your equipment:

  • Using the camera's in-built timelapse function: most compact solution and works well on the E-M1 except when you want faster than 1 second intervals;
  • using a remote shutter release or remote timer: works great but you have to dial in the intervals using the anti-shock functionality and it's an extra cable flapping in the wind;
  • a slider or panning head triggering the camera: whenever the E-M1 sits on the panning head (see next section), it will receive it's shutter releases from the Genie. The result: accurate intervals perfectly timed with the stops between motions of the moving parts of the timelapse setup.

For filters I often use a graduated ND filter to make the bright sky and the darker ground a bit more even. This is particularly important at sunrise and sunset because the ground will be really dark. I also have a circular polarizer that lives on my lens 95% of the time: vegetation looks more lush, colours more vibrant, and annoying reflections of leaves or glaring surfaces disappear. It can also cut through a lot of haze and mist on a more cloudy day. Time in Pixels just released an excellent article about filters for video with many visual examples.

Getting moving

I've written about my DIY slider before and it is actually undergoing some major upgrades right now to make it more usable and flexible. However, I don't usually take it very far because it is heavy and big. In order to have something that always fits in even the smallest bag, is compact and rugged (not weather-proof, though) and "just works", I got myself a Genie Mini which is actually being developed here in NZ. It's controlled from a smartphone via Bluetooth so setting it up takes a few minutes since my phone is usually off when I'm in the outdoors (no reception anyway) but it's very intuitive and flexible (watch the videos on their website). All the shots in the video at the top of the page that have some side-to-side movement are done with the Genie Mini.

RAW workflow

The out of camera JPEGs are alright but (especially for landscapes) don't look nearly as good as they could when I develop my own final images from RAW: better colours, more dynamic range, more wiggle room in the highlights (and some in the shadows). This is particularly important when photographing sunsets, sunrises, or rapidly changing lighting conditions because the exposure can be adjusted so much in post. I load all my RAWs from one scene into Darktable, then do all my adjustments on one of them (shadows, highlights, general exposure, Velvia/saturation filter, contrast, noise reduction, but no cropping - I can always do that later when editing the video). Then, I copy the settings to all the other RAWs and export everything to bitmap files with a high bit rate, such as 16-bit PNG or TIFF. In theory, one could also make fine adjustments to individual frames at this stage.

Editing

The last step is to edit the photos into a timelapse video and maybe add some music and sound effects. I mainly use Davinci Resolve for editing because it also has colour grading built in but the colours should already be fairly correct and good looking from the last step. Davinci can directly import image sequences (i.e. individual files) and display them as video clips.

Please use the comment section below or head over to Google+ or Twitter @tobiaswulff to discuss this article or any of my photography and videography work. My Flickr, 500px and Vimeo pages also provide some space to leave comments and keep up to date with my portfolio. Lastly, if you want to get updates on future blog posts, please subscribe to my RSS feed. I plan to publish a new article every Wednesday.

Weather-proofing the GH4 with Speedbooster

When I upgraded from an Olympus OM-D E-M1 (de) to a higher-quality kit, it was pretty clear that the weather-sealing of the camera body, lens and accessories won't be as good - it doesn't get much better than the Olympus OM-D E-M1 (de) plus Olympus 12-40mm F2.8 PRO lens (de) (rain, dust and freeze proof) so it could only decrease. However, watching most professional documentary and expedition shoots, apart from when using GoPros, video cameras are simply not made for rough conditions. Usually you work around the weather-related issues, often with a crew, using umbrellas and camera bags/sleeves.

The Panasonic LUMIX GH4 (de) is a great video camera and it is weather-sealed. For a full weather-proof package you would have to get the Panasonic lens which isn't really the best option for video shooters for multiple reasons:

  • MFT lenses deliver amazing images and they certainly make sense in a lot of situations (my landscape photography camera uses one and I wouldn't dream of putting Canon L glass on it when going hiking) but they are not a good long-term investment for film;
  • going wide with a GH4 in 4K mode (2.3x crop) or with a Blackmagic Pocket Cinema Camera (2.88x crop) is quite difficult so a focal reducer is a must for wider than normal angles;
  • smaller sensors such as MFT and Super16 still have quite a disadvantage over full-frame or cinema cameras such as the C100/C300 when it comes to low light so a speedbooster can greatly help to get those f1.0 or better apertures.

All of this combined means that vintage or modern Canon EF or Nikon mount lenses, both APS-C and full-frame, are the best choice for film. Good glass greatly improves the image and will work on many other, higher end cameras as well, for instance Blackmagic Ursa Mini or RED. Canon L lenses are great because not only do they deliver the goods optically, they are also weather-sealed, one example being the very versatile Canon EF 24-105mm f/4 L IS (de) .

This, of course, leaves only the Metabones Canon EF to BMPCC Speed Booster (de) in between the camera and the lens vulnerable to the elements. It's a shame that Metabones hasn't implement better sealing because the front and back end of the speedbooster is a pretty standard EF and MFT mount, respectively, and it is made of solid metal. I think the lens side of the speedbooster might actually be slightly sealed because of the rubber gasket from the Canon L lens, however, the camera side of things is just metal on metal and therefore not sealed at all.

Taping over the speedbooster with Gaffers tape works pretty well since it is easy to work with, and it comes off without leaving and residue behind (unlike Duct tape and other, stronger tapes). Unfortunately, it is kind of hard to get the camera to speedbooster interface taped off perfectly because of the awkward protrusion that is the GH4 EVF/flash housing. I added a layer of cling film as a first layer of protection, then taped everything in place. So even if the sides of the tape let some water in (which they shouldn't), hopefully the plastic will still keep the openings in the speedbooster dry.

My test consisted of a roughly 40 minute walk in moderate to heavy rain, keeping the camera level or pointing down (mostly to keep rain off the front of the lens) and always horizontal. I also decided to throw my SmallHD 501 On-Camera Monitor into the mix which also meant exposing the HDMI interface of the GH4. A plastic bag and more tape covered it nicely and I could still operate the buttons and joystick on the monitor. Swapping batteries on the monitor also still worked but it'd probably try to cover it more completely in the future and just cut a little opening for the HDMI cable (it was easier if it attached to the bottom of the monitor instead of the back at a 90 degrees angle).

Overall, it worked beautifully and I feel like I could venture out into a few hours of rain if I had to shoot something without the help of an assistant holding an umbrella. Swapping batteries or lenses or fiddling with any of the accessories is almost impossible without additional cover, though. Also, using a microphone would require a broader rain cover as well.

Please use the comment section below or head over to Google+ or Twitter @tobiaswulff to discuss this article or any of my photography and videography work. My Flickr, 500px and Vimeo pages also provide some space to leave comments and keep up to date with my portfolio. Lastly, if you want to get updates on future blog posts, please subscribe to my RSS feed. I plan to publish a new article every Wednesday.

Use a MIDI Controller as a Video Editing and Color Grading Surface

Control surfaces can greatly speed up editing and color grading work and also avoid issues like CTS because your hand can move around more freely instead of clutching a mouse all day. However, like most things in the video/film world, they can be very expensive. At the top end there are full suites like Blackmagic's Davinci (multiple $10,000s), the Avid Artist line (many $1,000s) and smaller devices like the Shuttle Pro V.2 (de) ($129). While they are often very well made and can be worth it if your profession is to produce multimedia every day as efficiently as possible, the actual cost of the functional parts is actually much lower. So I decided to build something similar to the Shuttle Pro but with a few key differences:
  • Sends MIDI messages instead of registering as a keyboard
  • Fewer buttons but a shift function that doubles the number of functions, including the jog/shuttle wheels
  • Different placement so that the hand can rest on the left and easily access buttons on the top and the right

MIDI

The MIDI protocol has been around for decades and is primarily used in the audio and lighting world for input devices and synthesizers. As opposed to using an input device that emulates a keyboard, MIDI has got the advantage that the incoming messages can be easily translated to keyboard shortcuts, whereas an Arduino emulating a keyboard will always send the same shortcuts. This gives much more flexibility, e.g. when changing programs or modes within a program (think media, edit, and colour pages in Davinci Resolve).

Within Resolve, shortcuts can be configured (or are configured by default) for pretty much all functionality apart from curves and colour/lieft/gamma/gain wheels - without a dedicated control surface one still has to use the mouse to modify these parameters. In order to process incoming MIDI messages on Linux I use mididings. I actually gave a talk at KiwiPyCon 2014 about using mididings to control photography (or any) software on Linux. On Windows I wrote my processing code in C++ and used the rtmidi library which is easy to compile (I use MinGW gcc) and comes with many excellent examples.

Assembly

The physical parts come down to a few buttons (cents to a few $), LED(s) (cents) and the jog/shuttle made by ALPS ($15-20). Figuring out the pins on the jog/shuttle was pretty straight-forward but this article goes through the process in more detail and might be useful to anyone trying to get a similar part working. I already had the plastic case and an Arduino to power the project lying around. In order to turn the Arduino into a MIDI device you'll have to replace the firmware on the ATmega used to communicate with the computer via USB (which is different from the main ATmega on an Arduino Uno!). The firmware and detailed instructions can be found on the HIDUINO Github page.

The shuttle controls forward and backward play at different speeds (J, K and L in Resolve) and the jog dial advances or rewinds the playhead one frame at a time (left and right arrow keys).

I still need to figure out a way to make an outer wheel for the shuttle and a knob or inner wheel for the jog rotary encoder. 3D printing might be the best way but first I'll have to learn how to create the virtual parts for it. At least the shuttle and jog wheels have sturdy grooves that should make it fairly easy to attach knobs or wheels to it.

At the moment, the Arduino will be connected to each button, LED and ALPS shuttle/jog through cables that go into the female headers on the little green board. Eventually, the Arduino will have to move into the enclosure and more sturdy, soldered connections between its pins and the components will be made.

Please use the comment section below or head over to Google+ or Twitter @tobiaswulff to discuss this article or any of my photography and videography work. My Flickr, 500px and Vimeo pages also provide some space to leave comments and keep up to date with my portfolio. Lastly, if you want to get updates on future blog posts, please subscribe to my RSS feed. I plan to publish a new article every Wednesday.

My (rather late) Guide to some Olympus OM-D E-M1 Features and Quirks

During my first big outdoor trip with the Olympus OM-D E-M1 (de) I found a few new ("hidden") features and quirks that I want to share.

HDR modes

In the dense Fiordland forests I experimented with the E-M1's HDR modes and found that, unfortunately, the camera itself doesn't really produce any useful, good looking HDR images (JPEGs). However, whether HDR1/2 modes or one of the multi-frame +/- EV modes from the HDR menu are selected, the camera will also automatically engage the high-framerate interval shooting mode so that the shutter only has to be pressed down once to capture all the exposures at a (theoretical) speed of 10 fps. If you want to make use of this feature you shouldn't set the "H" framerate to anything lower than the maximum. Note that in the modes HDR1 and HDR2 you won't get the individual exposures as RAWs - you only end up with one HDR image produced by the camera.

As described in the next section, the last used HDR mode cannot be toggled on and off with a press of the HDR button on the left-hand side of the camera the way you can toggle the other bracketing modes. Therefore, I had to sacrifice another Fn button and chose the one on the Olympus PRO lens since this is the only lens I regularly use for this kind of photography. My smaller lenses that lack the L button are typically used for portraits, wildlife or timelapses.

Quickly toggle bracketing

As opposed to the HDR functionality (which annoyingly has to be switched on and off with a separate Fn button), bracketing can be toggled with a quick press on the HDR button when the button lever is in position 2 and "lever2+left buttons" (the last item in the gear B menu) is enabled. The really cool thing is that a long press (more than 1 second) of the same button will take you to the bracketing menu where you can select between different modes (ISO bracketing which uses different ISO levels, AE bracketing which is similar but in addition varies the shutter speed as well, ART bracketing among others, and the new focus bracketing). Why can't toggling HDR work exactly the same way, it's basically the same thing??

ISO bracketing

I think this is my new favourite HDR mode for handheld shooting because it takes one exposure but delivers 3 different RAWs at +/- 1 EV (e.g. at ISOs LOW, 200 and 400). For many situations, that is enough coverage unless you're dealing with extremely high-contrast scenes. However, this also means that there is only one exposure and the camera delivers 3 different files with different light amplification so it is not the same as actually producing different exposures with longer or slower shutter speeds. Anything other than the base ISO of the camera (200 on the E-M1) will introduce noise and/or limit colours and dynamic range. Here are some interesting, yet slightly confusing discussions around this topic on Photography StackExchange.

One more note on AE bracketing which is similar to ISO bracketing because it does vary the ISO but also takes multiple true exposures at different shutter speeds: contrary to the E-M1's HDR modes, high-framerate interval shooting does not engage automatically so you'll have to set this separately if you don't want to press the shutter multiple times.

Interval Quirks

This, to me, is clearly a bug (although I can see why Olympus did it) and destroys the expectations of precision one might expect from the built-in interval timers used for timelapses and multiple self-timer exposures. Which features exactly do I mean? First, there is a self-timer mode where the user can enter a custom value (in seconds) for the camera to wait, and how many pictures to take at which intervals (also in seconds). The other feature is the timelapse mode where the user can set the interval between shots. Unfortunately, both modes suffer from one common problem: the interval timer only starts once the image (RAW, JPEG, or both) has been written to the SD card. Not only does this introduce an additional delay, it's also inconsistent as writing to the card does not always take the same amount of time.

I can see why Olympus did this, at least for continuous shooting like it is done in the timelapse mode: if it takes a while to write all data to the card, shooting more photos during that time will result in the internal buffer slowly filling up until the camera won't be able to take any more pictures. However, I didn't actually find this to be a problem, and here is both how I tested it and how I work around it when I want to shoot timelapses with interval times faster than the normally achievable 2-3 seconds.

You'll have to use a remote trigger than plugs into the camera. These can be bought for very little money on sites like Amazon. Then, enable the H framerate anti-shock mode (the one with the rhombus) and set your desired interval time (which can be fractions of a second) as the anti-shock delay time. Unfortunately, you'll have to go into the main menu to do this. An option to access this mode quicker is to save it as a MySet and make it accessible through a Fn button or a position on the mode dial. However, to change the interval time you'll still have to do a lot of navigating in the gear menus. Once the shutter remote is pushed down and locked, your camera will happily shoot timelapses at interval times below 1 second. This has worked fine with a reasonably fast and large SD card like the Transcend 64 GB UHS-3 Flash Memory Card (de) down to 1/2 and 1/4 s for many hundreds of shots.

Quickly access focus peaking settings

While the viewfinder or LCD is displaying focus peaking (either because assist is enabled and you've turned the focus ring, or because you've pushed a Fn button that has been assigned to peaking), a press of the INFO button will pop up a small menu that lets you adjust peaking colour and strength without having to trawl through the main menu. I wouldn't be surprised if this quick-access feature works for other things as well but I haven't noticed any yet.

Please use the comment section below or head over to Google+ or Twitter @tobiaswulff to discuss this article or any of my photography and videography work. My Flickr, 500px and Vimeo pages also provide some space to leave comments and keep up to date with my portfolio. Lastly, if you want to get updates on future blog posts, please subscribe to my RSS feed. I plan to publish a new article every Wednesday.

DIY Motorised Dolly Slider

Motorised sliders (or dollies) bring motion and a nice parallax effect to timelapse shots. Commercial sliders have the advantage of being well built (hopefully roughly proportional the amount of money spent) and easy to set up. On the other hand, costs for rails and card alone can be several hundred dollars and adding motors and a control unit quickly pushes it $1000.

A home-made DIY slider can be made with $150-200 for materials and the Arduino board or other micro-controller, depending on what you already have lying around. I'd like to take mine hiking more often but it wasn't built with minimal weight in mind so this is definitely a point where a good commercial carbon-fiber slider and a cart with less metal could come into play one day.

There are many different designs out there but the biggest distinctions between them are:

  • continuous motion vs stepper motion
  • two rails vs monorail
  • motor mounted on end truss vs motor mounted on the cart

Continuous motion is cheaper because a very simple motor can be used and the expensive timing belt can be replaced with a wire to pull the cart. However, this does not work for longer exposures since the camera has to be absolutely still while the shutter is open so I opted for a stepper motor right from the start. Keeping a cart stable on a monorail requires more engineering than with two rails but it can cut down on weight and makes it easier to mount to a single tripod with a screw hole in the centre of the rail. I didn't know how to build this so I went with two rails and wheels on either side to keep it stable.

At first I thought I can keep the cart lighter by mounting the stepper motor to the end of the rails. While this is true, once you add up the weight of the metal cart itself, a camera body and lens, and ball head, a single stepper motor wouldn't make much of a difference any more. Having the motor on the cart has several advantages: everything, from the camera to the motor to the control unit is in one place and you won't need to run cables all over the place; only half the length of timing belt is required since it won't loop around the ends. Here is a link to a good example of a DIY monorail motor-on-cart slider.

    Here is a quick test video I shot using the slider. Unfortunately, I bumped it a bit at towards the end but you get the idea. There will be more exciting timelapses in the future that actually use the motion/parallex effect in a meaningful way.

    Material List

    • Arduino: $5-20 depending on original vs clone and capabilities - it's easier if it fits a shield
    • Motor driver shield for Arduino: $20
    • Battery pack and switching regulator: $10
    • 12V NEMA-size stepper motor and mount $23 - alternatively a smaller stepper motor
    • Timing belt and pulleys $40 - one could probably find much cheaper spare parts somewhere else
    • Aluminium rails $10-20 - I can't remember exactly
    • Steel or aluminium cart and ball bearings - can't remember how much it was, maybe $20; I had the ball bearings already
    • Cheap, small to medium-sized ball head - don't use the really small ones like the Giottos Mini Ball Head if you have anything bigger than a compact point&shoot because it will wobble a lot and adjusting it will be very hard: $20-$30

    In the photo above - once you look past the rat nest of wires - you can see the motor driver shield sitting on the Arduino. All connections come out of the shield (they are fed through from the Arduino), so the rainbow ribbon cable is for the rotary encoder, there are some wires for the LED, ground and 5V, and also the stepper motor itself which is hooked up to the left-hand side 4-pin screw terminal. The two "things" encased in plastic in-line with the wires are a fuse and a switching regulator to bring the input voltage (9-12V) down to 5V for the Arduino. A linear regulator like the one that is on the Arduino would work too but might generate too much heat for a closed up enclosure.

    Photo above: a cheap but decent-sized ball head that unfortunately wobbles a little bit but does the job. Since all ball heads (and all decent tripods) use the same screw sizes, you can mount whatever you want, small and cheap or big and fancy. I like that the ball head has got a two-way water bubble level built-in.

    Programming

    On the Arduino platform I use the Adafruit_Motorshield library. To move the stepper motor the minimal distance and as smoothly as possible I run:

    Adafruit_MotorShield *motor_shield = new Adafruit_MotorShield();
    motor_shield->begin();
    Adafruit_StepperMotor *motor = motor_shield->getStepper(200, 250); // steps and speed
    motor->step(distance, FORWARD, MICROSTEP);

    I also set up an LCD and a rotary button using the SoftwareSerial and ClickEncoder libraries, respectively. Text can be written to the 16x2 LCD directly over the serial line, plus there are some special characters that move the cursor, clear the screen, and so on. The ClickEncoder uses up one of the timers of the Arduino and unfortunately it is the same used by the MotorShield library so I can't use both at the same time. This is ok because I only use the rotary encoder to set up all the timelapse parameters, and once the slider is moving and the camera is taking pictures I don't want to touch it again anyway. It's basically two separate programs: first the menu/settings, then the timelapse.

    Issues

    I found that the 200 steps per revolution that the stepper motor provides aren't quite enough for super smooth and slow motion, so after about 15-20 minutes with one frame every second the cart will already reach the end of the rails. It is possible that I haven't configured the motor correctly yet but it is set to micro-steps in the code and as far as I know this is the smallest possible rotation. I use micro-steps instead of normal steps because the provide smoother movement: a normal step would yank the cart and make the heavy camera wobble too much. It also makes sure that the motor is always enganged in case the rails are on an angle. This way the cart can't slide back down. To solve the problem of step sizes being too big I might incorporate some model kit plastic gears but for now I have avoided it since it isn't that easy to get everything lined up correctly without making the whole construction incredibly flimsy. There are also other stepper motors out there that provide twice or more the amount of steps per full revolution, usually through internal gears (see the alternative smaller motor I've listed above).

    Currently I'm running the whole setup off of 8 AA batteries, specifically Panasonic Eneloop AA Ni-MH Rechargeable Batteries(de). However, since the stepper motor requires 12V and 8 rechargeable AAs only provide a maximum of 9.6V, it does have issues climbing an incline stepper than about 20 degrees. On the flat it works great, though, and it lasts for many hours as well. In the future I might upgrade to a 12V battery or boost it up with a converter, maybe using a LiPo battery for their amazing energy density.

    Future Developments

    I've got a little micro-switch that I want to mount at the end of the rails so that it detects the cart hitting the end. This will eventually stop the timelapse. I also want to refine the menu system and hopefully improve the software side of driving the motors, i.e. better speed and step control.

    In case I've already exhausted all possibilities regarding the motors, I might actually have to add two differently-sized gears to the system to bring the speed down so that I can take hour-long timelapses with many hundreds of frames.

    And finally, panning while moving sideways would make my timelapses look much more impressive so adding a second motor is high up on the agenda. However, I'm not sure yet how to fit it between the camera and the top of the ball head (glue it to the quick-release plate?), or alternatively if it could or should live under the ball head in which case I'm worried about stability. The latter would see the motor mounted under the cart, however, which would be a very clean looking solution. Either mounting point will give very a different result when the rig is on an incline and depending on the subject it can work well or look really out of place.

    The most important improvement, however, to be made is getting it off the ground. At the moment grass or plants can easily get caught in the wheels and there are no points to screw in a tripod quick-release plate (the end trusses hook quite nicely into the top of my Manfrotto BeFree Travel Tripod(de) , though). Often something can look good at eye-level but having to put it all the way down on the ground limits my possibilities so a good 1/4" screw hole at either end would make it immensely more useful and stable in vegetation.

    Please head over to Google+ or Twitter @tobiaswulff (see links on top of the page) to discuss this article or any of my photography and videography work. My Flickr and Vimeo pages also provide some space to leave comments and keep up to date with my portfolio. Lastly, if you want to get updates on future blog posts, please subscribe to my RSS feed. I plan to publish a new article every Wednesday.

    Long-run timelapses across multiple seasons - Part 1

    Some links to products in this blog post are Amazon Affiliate links that earn me a few cents or dollars if a reader buys any product on Amazon through this link. The price of the product does not increase so it is a free way to support this site by using the links provided. The main product link goes to Amazon.com and the "(de)" leads to Amazon.de.

    This is the first part of a longer series of blog posts about timelapses. I have started planning for and taking long-run timelapses that span many weeks and months, and I want to talk about how these ideas and visions can be accomplished in a reasonably efficient workflow by an amateur photographer. I say "reasonably" because processing timelapses from RAW files and working on such long running sequences will always involve a lot of work.

    Who, and why

    Apart from an article on Photo Sentinel, there aren't many interesting articles or howtos available. I highly recommend reading the article if you're interested in timelapses because it showcases different techniques and links to some great videos in each category. However, it belongs to a company that sells specialised long-term timelapse equipment which does not really fit the kind of subjects I'm shooting. On Youtube and Vimeo there are only a couple of videos that portrait certain subjects in nature over the course of many months but there are some amazing and award-winning short videos and films that I will link to further down.

    The most impressive executions of this sort of timelapse - and the aforementioned howto talks about it as well - are several features by the BBC such as The British Year and of course Planet Earth. The team which shot the timelapses for The British Year talk at length about planning, shooting, editing and various tips in a blog post. I highly recommend Chad's blog and all the content on his website as it is a wealth of timelapse stories, workflow tips, and kit reviews.

    How-to

    The obvious but most time extensive way to shoot a timelapse across multiple seasons is to take individual photos of the same subject under similar lighting conditions and from the same spot over a long, long time. Another technique is to take multiple "normal" timelapses, that is sequence of an hour or a few hours, and then blend them together such as in the Youtube video "4 Seasons 1 Tree". Unfortunately, the blending will be very obvious, and there also isn't much movement or change within the individual sequences themselves. On the other hand, there is no flickering due to abrupt changes in lighting or weather. This could be enhanced by doing some masking and selective blending to change some areas of the image before others which can also be seen in the video as the ground changes before the tree does.

    The easiest way to accomplish a long running timelapse is to have a camera that can be left in a fixed spot and orientation. The photo above is actually a blend of two individual frames, one with different lighting and more leaves on the tree. It shows that blending and aligning photos on the computer can produce a very smooth result even if the original photos are totally unaligned and taken in completely different conditions. In amateur nature photography, it usually isn't an option to have an absolutely fixed camera spot because the locations are too exposed to the elements. Even in urban environments you wouldn't leave your camera or tripod anywhere except inside your own house or apartment - and then you wouldn't be able to take it somewhere else.

    Therefore - unfortunately - we have to re-set up the camera and point it at the same spot every single time. This gets very complicated if movement of the camera is involved but even with a static shooting position there will be slight variations due to uneven ground, zoom lens variations (zooms are not "clicked" after all) and inaccuracies when pointing the camera at the subject. A very sturdy tripod is important but because I usually travel on foot or bike and also take my equipment on hikes into the mountains, I couldn't just go for the most sturdy one out there. So my tripod is the Manfrotto BeFree Compact Aluminum Travel Tripod which I love because it fits even into a normal day pack, yet it can extend to eye level and is reasonable rigid. However, pushing down on it will bend the legs in their joints so it is tricky to get it set up 100% exactly the same way every time.

    So there will be variations in tripod position, tripod height, camera attitude and focal length. Luckily, those issues can be resolved almost completely in post-production and I will talk about methods and tools to align photos and blend them in the next blog post in this series. Apart from dedicated software and plugins for Lightroom there are also a bunch or free tools available that do a very good or even perfect job at the expense of a maybe not so polished user interface or some efficiency.

    Ongoing work and ongoing articles

    Something like the video "Fall" from NYC Central Park is probably the closest inspiration to what I am planning to achieve. I didn't know the video when I started my project. There is also a year-long timelapse from the Canadian Rocky Mountains which employs some really nice blending and obviously beautiful outdoor scenery.

    As I shoot individual frames and sequences for my own long-run timelapse video, I will add more parts to this series talking about specific shooting tips and releasing some more snippets of the ongoing work. Towards the end I'm sure it will all become fairly editing and video post-production heavy.

    Please head over to Google+ or Twitter @tobiaswulff (see links on top of the page) to discuss this article or any of my photography and videography work. My Flickr and Vimeo pages also provide some space to leave comments and keep up to date with my portfolio. Lastly, if you want to get updates on future blog posts, please subscribe to my RSS feed. I plan to publish a new article every Wednesday.

    1 of 2
    1. 1 2