10 lessons I learned, observations I made from making of "Home"
My last project "Home" was fun to work on and it was a great learning experience. I learned a ton during making of this little film. Small projects like this, are invaluable for beginning filmmakers just like myself. Best way to learn anything is to "get the hands dirty". I completed the project but I realise it is not perfect and there is a ton of things I wish I did differently. Let's take look at the project and break it down, talk about equipment, techniques used, mistakes made and considerations for the future projects.
I shot most of the footage on Fujifilm X-T2 with a variety of lenses such as mainly XF 14mm f/2.8, XF 50-140mm f/2.8 and XF 35mm f/2. Rest of the footage came from GoPro Hero 5 and DJI Mavic Pro. For controlling exposure I used Tiffen variable ND filter on Fujinon lenses and NiSi ND filters on the Mavic. Except few locked off tripod shots, I captured the rest of the footage handheld. I mounted the GoPro on Joby Gorillapod and used it instead of a GoPro stick.
Image profile setting
I set my Fujifilm image profile to Pro Neg Hi picture profile with sharpness, contrast and saturation at their minimum values. I saved these settings as a custom profile so it was easier to recall the settings whenever I needed them. There are a handful of shots in the edit I mistakenly shot with Astia profile.
DJI Mavic footage was shot in 4K using D-Log picture profile with the sharpness set to +1, contrast and saturation to -2. There are a couple of shots in the edit where I had to use the cache file from my phone (which is only 720p in resolution) since my SD card got corrupted in mid-flight.
ProTune at 2.7K is my preferred picture profile and resolution for GoPro. It is also the highest resolution possible with image stabilisation engaged. I used a linear field of view to avoid the GoPro fisheye effect.
For editing, grading and sound design I used Davinci Resolve 14 Lite. It is a no-brainer. I cannot believe the application of this calibre is available for free. Obviously it has some minor limitations compared to paid version but the differences are minimal. My final output was in 1080p anyway (output resolution limit is one of those limitations).
10 points I wanted to talk about. Challenges, lessons, observations.
As I mentioned before, these small projects are great for learning about the filmmaking process. I learned a ton myself and I will use this newly acquired knowledge in my next project. Now that I discussed the cameras, settings and resolution, let's talk about "stuff".
1. Fujifilm X-T2 as video camera
The quality of the footage coming out of this camera is excellent in my opinion. The APS-C sensor size is very close to the Super35 size and the aesthetics of it is very pleasing to my eye. I thought the lack of log profile will hold me back but actually, I didn't miss it as much. My preferred picture profile (Standard Neg Hi) gave me enough dynamic range and it didn't need to be pushed too hard in grade. it just looked good out of the box. A little bit of contrast and saturation adjustment got me quickly in the ballpark. Even few clips I shot accidentally with Astia picture profile looked great, although bit contrasty for an additional grade.
We all know that Fujifilm X-T2 is an excellent camera for photography and its usability is a feature that many photographers praise. Including myself. The video is a little bit of a different story. As discussed, the video quality is an excellent however camera's usability for video production left a lot to be desired for. From a hybrid shooter point of view, switching from shooting stills to shooting video is a bit clunky. The camera does not remember the video/stills specific settings. Every time I changed the mode from shooting stills to shoot video I also had to change a bunch of other settings. When shooting run-and-gun style (which is how I mostly shoot), I sometimes forgot. Hence few clips I shot accidentally with Astia profile rather than Standard Neg Hi.
Lack of IBIS really hurts (rumours say Fujifilm is working on it). No video aiding tools make the shooting experience also less pleasant and somewhat difficult to make educated decisions on exposure and focus. As of now, the only exposure aiding tool is histogram which is barely sufficient for video. Another area I want to touch on is focusing. It is really hard to use the back screen for focusing. Even with focus aids such as focus peaking. All these issues can be solved by adding an external monitor such as SmallHD Focus. That already makes the setup less nimble though. But it is a solution.
Fujifilm shooters came to love the Fujinon glass. Excellent build, optical quality, reasonable pricing are things we all appreciate. For video work, they are less ideal. All Fujinon glass uses focus by wire. Pulling a manual focus is very hard and it is not repeatable. It is the nature of focus by wire. So if possible, it is better to use proper manual focus lens. At this stage, it has to be adapted lens since Fujifilm doesn't have any lens like that in the lineup. This is not always possible, for example when on road and I have to pack light, I just skip on those lenses and work with what I have.
2. Using auto-focus
It is pretty obvious that when doing serious video work, I should use manual focus. I used auto-focus on X-T2 number of times and as expected, it cannot be really trusted for video work. For a professional shoot, it is a big no-no. Although it is among the better autofocusing cameras in the mirrorless world, it is still not reliable for a serious production. The performance will vary on the frame, subject, it's size in the frame or it's movement. Generally, for my next project, I will use manual focus as much as possible even with the limitations of focusing by wire. Perhaps I sacrifice shallower depth of field just to be sure the subject is in focus.
3. Mixing frame rates
All footage was edited on 1080p, 24fps timeline. However, I shot many different frame rates and resolutions. Footage from Fujifilm X-T2 was mostly 24fps. For slow motion, I used everything from 47.952fps to 59.940fps. All GoPro fotage is captured at 47.952fps and clips from Mavic at 23.976 fps (although it was set to 24fps in the DJI GO4 App). I used few clips at 30 fps where the Mavic didn't record on the SD card. I was forced to use cache files from my phone which came in at 30 fps and 720p resolution.
4. Need more stable footage
Since Fujifilm X-T2 doesn't have an IBIS, it is very challenging to get stable footage handheld. Even with an optically stabilized lens. Any kind of stabilizer (glidecam, 3-axis gimbal etc.) would improve the production value. For the future projects, I will need to employ a whole bunch of tricks to get as stable footage as possible in camera. GoPro footage was shot with internal image stabilization turned on, but it is still far cry from acceptably stable footage. Mavic is obviously the best out of the three in producing stable smooth clips.
5. Davinci Resolve stabilizing function
One way to save a shaky footage is applying some sort of software-based stabilization in post-production. Personally, I have no experience with Warp Stabilizer from Premiere but I can say the Davinci Resolve built-in stabilizer was not good enough to save most of the shots. Applying the stabilisation in Resolve often resulted in really bad looking jello effect and it was not usable. Of course, the software cannot do a magic. It is designed to improve a little shakiness or micro jitter. If the footage is a mess from the start, there is obviously no way to make it usable. Lesson learned. I need to use any kind of stabilizing hack I can possibly come up with for my next project. Time to be creative folks.
6. Sound recording
I admit. I slacked on this very crucial aspect of filmmaking. I had a shotgun microphone with me but I barely used it. Mostly because I didn't have it at hand when I grabbed the camera to capture something. Same goes for the sound recorder. It mostly stayed in my bag for the very same reason. This is obviously a big mistake. My initial idea was only to use a background music for this video. That quickly changed as I wanted to challenge myself with something more complex. I designed most of the sound FX in the movie in the post-production stage from sound clips I collected over the years. I also used sounds from Freesound.org.
7. Disk space
This proved to be a problem. I stored the footage on external USB 3.0 hard drive. My MacBook Pro doesn't have big enough SSD to store the clips while editing. So I was editing from external hard drive. Pain in the butt, but I had no other option. Davinci Resolve 14 got much better at handling H.264 files so I sort of got away with it. I turned on caching and it helped a lot. At some point, I had issues with final renders from Resolve (discussed in the next point) and one of the steps I've taken to mitigate the problem was generating high quality optimized media (ProRes 422HQ). That quickly filled my SSD drive though. I had to offload some other files to various different external drives. I lost a lot of time on this. Lesson learned. Always have enough space, more than you think you need. And USB 3.0 is less than ideal when editing from external hard drive. I knew this but I had to work with what I had at my disposal.
8. Davinci Resolve output issues
I had an interesting issue with Davinci Resolve. It worked flawlessly up until I tried to render my final output. Upon reviewing the file I noticed, that grade dropped on certain frames. I had no idea what caused this issue. It drove me crazy. This appeared again and again and I noticed it mostly happened on GoPro files and occasionally on X=T2 clips. Footage from the drone was unaffected. I ended up generating hi quality optimized media to help alleviate this problem. It worked partially. After many tries, I got a version which was free from flicker but unfortunately, I haven't cracked this mystery. It might have been just my setup or a result of mixing many different frame rates, or even my external hard drive as my working drive.
9. GoPro performance in low light
The clips from GoPro were extremely noisy when the light levels dropped even just a little bit. I set the GoPro to automatic exposure setting which was probably a mistake but anyway, there was just too much noise in the footage. It shows in the final movie. Perhaps I shouldn't use those clips at all. It is worth to note that Davinci Resolve Lite (free version) does not have the built-in noise reduction available.
10. Challenge of editing own footage
I think many can relate to this. I noticed, sometimes I have a such and emmotional connection to the shot that I want to have it in the movie and I don't care if it fit's there or not. That is not a good thing. And as it often happens, once the movie is live, I realize those shots were not necessary. They don't really contribute to the movie in any way or they simply break the flow of the edit. It is a good idea to have another pair of eyes to take a look at the edit and give constructive criticism. I never showed the edit to anyone. This was a small personal project so it was probably ok but I know for a fact that for a paid job I will get a second opinion. If I can afford it I even hire an editor.
One final note
Filmmaking is definitely something I want to keep doing. I am most and foremost a photographer but a video is a different medium to tell the story and it is very intriguing. As it happens, it is in much higher demand than photography. I think I can do both and I am really excited about new projects on the horizon. This medium is there for me to explore, especially the small niche of one-man-band productions and hopefully come up with some nice looking short films. And I also hope if you reached this point of this making of article, you learned an thing or two as well.
Please visit my gear page if you consider buying some of the equipment I used to shoot this video. There are affiliate links to Amazon. It cost you nothing and it can help me. Your support is appreciated.