“That was easier than I thought it was going to be.”
I actually said that after completing my brother’s birthday video. I meant it, but it does gloss over some pretty major things. This post is mostly for me to remember some specific details, but also for those who may want to know how much (or how little) effort was required in remaking Enjoy the Silence, one of Depeche Mode’s most iconic videos.
Leaping Into the Void
Mary (mostly) convinced me to do this project while we were chatting before bedtime on December 15. As usual, once an idea grabs hold of my brain, my brain won’t shut up about it until it’s embraced, forgotten, or abandoned. In this case, I had embraced the concept, but there were so many things to do before determining if the project was even feasible:
- Will the music be as exact as I can make it or am I doing a more interpretive cover version?
- Originally, the “band” was going to be Bryan and me, with cameos from kids and spouses. What now?
- Can I find an appropriate costume?
- Is an iPhone going to be good enough to shoot this?
- Is it even remotely possible to put all of this together by New Year’s day?
I knew that I wanted the music to be exact, but that meant finding an instrumental version, naked samples, a multitrack, or some combination of those things. Since my brain wouldn’t shut up about all of this, I got out of bed and did some late-night searching. Amazingly enough, after a few false starts on sites that were clearly scams, I found one set of stems (“STEreo Masters”) for the song. These weren’t real stems or the original, but did have the original vocals isolated and had recreated or sampled the three iconic sounds I needed: the bass line, the synths vocal stabs and sustains, and the guitar riff.
This was really the primary hurdle for me. Knowing I had a passable version of the audio made everything else seem achievable, and I was able to go to sleep. I had three days to get things rolling before accompanying Mary on a work trip to Spokane. Her plan was that we’d use the drive out to scout locations along the Columbia River (or in rural Washington), and then film on the way back. This meant that three things had to be in place before we left:
- The lyrics (at least the chorus) had to be finalized because I’d be singing on camera.
- Filming equipment had to be chosen (or acquired).
- All the pieces of my costume had to be in the car with us.
It was a separate idea, using extended family to be in the band segments, that led to the lyrical solution. While talking to Eric’s wife, she mentioned that he was struggling a little bit with the big 5-0 and didn’t really want to talk about his birthday. That gave me the idea that this should be an non-birthday birthday song. Also, as I read over the original lyrics, the depth of emotion really came through. Not only would it be a non-birthday song, but would highlight something many of us learn too late: achievements are great, things are great, but it’s the people who surround us that are the most meaningful things in life. “I” to “you” refocused the song, and it only took a couple of minor tweaks in other places to make it all fit together.
To identify the specific shots I was trying to recreate, as well as get a feel for how complex the overall project would be, I loaded the original video into Final Cut Pro (FCP) and burned a timecode window on it. I then took that version and put it in Digital Performer (DP), to map out the tempo and cuts. I noted both the SMPTE time and Measure / Bar / Tick points for each cut in the video. I created a spreadsheet that had a screenshot in addition to all of this information, and carried a printout with me for the rest of the project. I should have referred to it even more often than I did, but it was a huge help for both planning and reminding me that things weren’t as complicated as I kept making them out to be in my head.
Living On Video
I had three options on hand for filming: a Sony HD camcorder from 2006 (it uses DV tapes!), a GoPro 3 with dodgy batteries, or one of our phones. The camcorder and phones were the best bet, but both suffered the same problem: decent lenses and zoom. Mary and I had talked about getting a “nice” camera for years, and it didn’t take much discussion before we agreed that Santa would be getting us a nice Canon EOS 90D for Christmas. This wasn’t just a boondoggle – I have video plans for the coming year that really did require a better / more modern camera – but it was a nice excuse to finally make the acquisition. I also ordered a backdrop and lighting kit, along with a ring light. They wouldn’t show up until the 24th, but that was plenty of time.
As for the costume, Amazon to the rescue. Sort of. I found a cheap crown that looked about right, but couldn’t find a full-length red cape with ermine. Apparently, there is only one supplier for the entire United States, and nobody would have stock until after Christmas. I took a chance on a purple one from Amazon, but it was too small and see-through – definitely not the heavy cloak I was wanting. After a bit of panic in the wee hours of Saturday morning (the day before we left for Spokane), I started looking for local costume shops, and found one just a few miles away that had what I needed for rent. Since I knew a snow scene was in the cards, I decided to use my black snow pants, and found a high-neck shirt at Target. A pair of $3 leather work gloves from Home Depot completed the ensemble.
Majesty, Fences, and Weather
The drive along the Columbia River is gorgeous. As we went along, I was looking at the scenery and seeing plenty of places that would work, especially if we got off the highway and on the backroads. I noted these places and then used Google Maps / Street-view images to virtually scout locations while Mary was working. It was all looking good, but then a huge winter storm blew across Oregon and Washington. As it turned out, everything was covered in fresh snow until we got almost all the way to the Portland area. Not an ideal start, since only the final mountain scene needed snow. It was looking like the entire video might be snowbound.
I identified three places for the drive back that would be good choices. The first was gated and fenced, and the snow would have made the gravel track impassable even if we could get in. The second location, Blalock Canyon was actually a target of opportunity. We were in a large area of farmland, freshly covered in snow, on our way to a different area, and decided to go ahead and get some footage. Mary dropped me off at the top of the hill, then drove back down the road about a mile. This is the first of the King shots in the video, and that area ended up being a good one.
The intended second location, Philippi Canyon, ended up being inaccessible. On Google Maps, it looks like a beautiful scenic hiking area, but it’s actually a privately owned, working ranch. To say that they didn’t want people walking around is an understatement. You don’t notice how much of the world is fenced until you’re looking for a place you can walk. I wonder how many miles of fencing there are in the United States alone, much less the entire world.
We didn’t have enough light left to film at the third site, Rowena Point, but I knew that this was probably going to be one of my locations and we earmarked it for later. When we got back, I was in a semi-panic because the forecast for the region was awful. Rain at best, snow at worst. I actually thought about flying someplace like New Mexico or Colorado, but the entire western US had pretty much the same forecast. To make things more pressured, a huge snow system was forecast for the 25th or 26th, even in the lowlands. If I didn’t get my footage before Christmas, I probably wasn’t going to get it.
The next day, we drove to nearby Ridgefield, WA, and completely lucked out with a break in the rain, just in time to get some beautiful sunset skies and river shots. While on site, it seemed like I’d have far more than I needed, but that would prove to be a mistaken assumption when I got to editing. On the way home, though, I was starting to relax a bit – I knew Rowena would be a good place, and the beach was already set. I still intended to head up to one of the lodges on Mt. Hood to get my mountain shots, so Ridgefield would be a safety site.
Initially, I was planning on the northern part of Cannon Beach, along with the iconic Haystack Rock featured in The Goonies, but was worried about two things: first, it’s a popular place and getting a shot without any people in it might be a challenge and, second, a dead blue whale had washed up on the beach the night before. I definitely didn’t want that in the video. I also didn’t want to smell it. Mary suggested a few places to the south, and we found a great location. The trail down to the beach made for some great footage as well. This was a good thing, since we were there at high tide and couldn’t get to some of the caves / passthroughs I had originally planned on using.
It became apparent that Mt. Hood wasn’t going to be possible – even if we could find a location without people, it was unlikely we’d have clear skies, and the whole point of going was for that magnificent “top of the world” shot in the Alps at the end of the original video. At this point, I started compromising a bit on my perfectionist vision, and decided that the view of the Columbia River Gorge from Rowena Point would have to be my closing shot.
The final day of location shooting was on Christmas Eve – we returned to Rowena with friends, Brian and Mary, had a nice hike, a reasonably fun time (when I wasn’t stressing about getting shots during breaks in the weather), and finished up with one of the best Chinese meals I’ve had in a long time.
Some technical details about the filming. On the first day in Blalock Canyon, I had correctly set my frame rate to 23.976 (for that film look), but mistakenly set the camera to record 4k footage. This ended up being a blessing, and I shot all of my other footage the same way. The main reason it was beneficial is that it gave me the ability to crop, zoom, and rotate footage without any quality loss for my intended end-resolution of 1080p. This saved a couple of shots, especially in Ridgefield, where I didn’t have the tripod level.
I was worried that my ten-year-old Mac would have trouble with the high resolution, but it never broke a sweat. Takeaway: if you can shoot in a higher resolution than what you need, do it. I didn’t know enough about the camera or videography to use anything other than full-automatic mode. This was probably the right choice, though there are a couple of times where automatic focus or exposure changes posed a problem in editing. I did use my old Sony Camcorder’s tripod, and that was definitely the right choice.
Green or Black
The backdrop and lighting kit came with three cloths: black, white, and green. My initial plan was to use a green screen for my “studio king” and band shots, but didn’t have the right lights / skill to adequately light it and get a good key. When I loaded the test footage into Final Cut Pro, it was pretty obvious that it was going to be far more work than just shooting in front of a black backdrop and adjusting the exposure level down.
While I’m really glad I have the light kit for future projects, I ended up using two Hue lightbulbs in gooseneck lamps to get the stark side-lighting for these shots. My one questionable purchase was the ring light, not because it’s not great to have it, but because it was both huge (it’s an 18″ ring, which is a lot bigger than it sounds), and was too bright for the look I was going for. It overpowered the side lights. It would have been more useful on some of the exterior shoots to light my face and costume against bright skies.
One of my favorite shots in the entire video, presenting the birthday cake, was a shower thought, but I knew it was right the second it entered my brain. We got the cake and candles at Safeway and, yes, there are 50 lit candles on that cake. In retrospect, it probably didn’t matter because you can’t see anything other than FIRE, but I really liked the effect.
A Rose By Any Other Color
This was a “seems simple enough” side-project: recreate all of the color combinations of roses on boards from the original. I thought about just grabbing frames from the original video, but they were highly compressed SD and would have looked out of place. So I bought a dozen roses from the grocery store, a couple of boards and spray paint from Home Depot. I didn’t have optimal weather for spray painting roses, but it all worked out well enough. If you ever want to do something similar, multiple thin coats is the key.
I was really pleased with how these turned out, and even put the camera into manual mode and experimented with various exposure levels to get the saturation and highlights correct. This was definitely one of those “Ray loses his battle with perfectionism” moments, but I think using the frames from the original video would have bugged me forever. I’m glad I went to the extra effort and expense in this case.
I won’t go into too much detail on this, other than to say that my “cool, I’ll just use stems and drop my voice in” plan imploded almost immediately. First, the files I found weren’t exactly right, and followed the arrangement for the radio single. The video had a different arrangement and mix, notably the introduction. Additionally, things like the snare/clap and guitar weren’t original recordings, but short samples, often flooded in effects to disguise the inconsistencies. In the context of the mix, they were pretty close, but not close enough for my ears. I ended up recreating most tracks from scratch. In the case of the guitar, I used EQ to isolate mostly the high frequencies that give the original its character, but then used a different guitar instrument at a lower volume underneath to smooth it all out.
The vocals went pretty quickly. I decided to have Dave Gahan singing softly in my ear while I recorded so I’d get a little closer to his original timing and inflection. Initially I had way too many effects on my voice. When I got to mix down, I took them all off, did a little EQ work, shortened the reverb, and called it good.
The Family Sends Their Regards
Video from the various families started trickling in. At first I was frustrated that not everyone was able to shoot in front of a dark wall, but that quickly gave way to being happy that I had lots of options. This might be rationalization, but having every band shot with a black background might have made people less identifiable and the overall effect more monotonous. It made sense when there was only one band in the original, but this version would have several.
I did what I could to repair video, including removing windows, electrical plugs, and light switches. Most of these fixes were far easier than I expected. For the video shot against a black backdrop, I’d do a quick shape mask around the people, then lower the exposure until the backdrop disappeared.
It was fun seeing the raw footage, and I decided to make a “Behind the Scenes / Blooper” video, which is included at the end of this post.
Putting It All Together
With all of the footage and the music complete, it was time to assemble the final project. This went far more quickly than I anticipated, and I had the rough video laid out in just a few hours. Initially, I referenced the original video and closely matched cuts and angles down to the exact frame. Because I didn’t have the exact same shots, though, this became a problem and I decided to err on the side of making an enjoyable new video instead of an exact reproduction. It’s still pretty interesting how close things are, though. Here’s a version that has the original overlaid on the new one:
All things considered, this was a pretty easy and simple video to recreate. There isn’t any CGI, there’s no crazy choreography or camera work – it’s almost closer to documentary or corporate videos with simple cuts. That said, there are a couple of shots where there’s more going on than may seem.
In the shot where I’m walking away from the camera in an open field, there were two signs and a light pole. In theory, this should have been easy to remove with simple masks and shifting layers, but FCP isn’t able to do multiple masks in any sane way. I ended up using a demo of the Slice-X plugin from Coremelt, which made quick work of everything. I don’t have any immediate needs for that plugin, so I didn’t buy it, but will definitely do so the next time it goes on sale or is required for a project.
There are a couple of shots where I used color grading to bring out the sky a bit, mostly in the snow scenes. Additionally, there are a couple of shots where I replaced the sky entirely with one from another shot. Some of the Rowena location skies are actually from Ridgefield.
One of the most technically impressive effects was also one of the easiest to do. In the two Rowena shots where I’m setting up the chair and looking at the gorge, there was a lot of traffic on the highways on either side of the river. I used the FCPX Removal plugin from Pixel Film Works for this, and a lot of it was fairly automatic. The plugin analyzed the clips and removed anything that was moving. I then went back and replaced the things I wanted – like me – and then had to do a little mask and color adjustment work. The reason for having to do this was that I filmed with fully automatic exposure settings. Small movements in the foreground could make the camera brighten or dim the image jussssssst a bit, but enough that it was obvious when another part of the image didn’t change. In all, I had to manually change the exposure on about 15 frames – not too bad.
Against the Grain
The final effects I used were to apply some color grading to simulate old film stock (the “Vintage” plugin from Sheffield Softworks) and put some over-the-top film grain on all of the walking footage using the Kingluma Granularity plug. It was a seemingly small thing to add, but was one of those small touches that made a big difference in the overall effect and quality of the remake.
It also made a HUGE difference in the file size of the end product. When digital video is compressed, footage that has lots of areas that don’t change can compress down quite a bit. But “film grain” is really just random noise. Noise that covers the entire screen and changes every single frame. This is a worst-case situation for compression.
But what’s the real-world impact? The full-quality, 1080p direct export master file from FCP is 3.54 GB, and 4.5 minutes long. For comparison, when I rip one of my DVDs (480p), a feature-length film is normally between 1 and 2 GB, and a Blu Ray movie rip (1080p) is normally somewhere between 3 – 8 GB at the quality settings I use.
So using the same settings I use when compressing my Blu Rays, I got the video down to 341 megs. That’s roughly a 10x reduction, which isn’t bad, but it’s still a lot of data to be moving for a 4.5 minute video. For streaming, and on the devices that I assumed most people would be watching (phones or web browsers), a 720p video is perfectly adequate. By reducing the dimensions, I was able to shave off a decent amount, but it’s still 253 megs. And that’s almost entirely due to the fake film grain.
How do I know? Compare that to the Behind the Scenes video below, which uses only the original footage with no added grain. Like the music video, it’s also 720p, but is 30 seconds longer. It compresses down to just 38 megs! The 1080p version is only 60 megs. Was the film grain worth the extra size? It definitely was to me – when I look at the video without the grain, it looks… wrong. Even “fake” in a way, which is hilarious.
Is It Finished?
One of the oddest moments in a project like this is when the end of it sneaks up on me. I don’t mean the deadline, I mean crossing the finish line. I was doing some minor tweaks on the morning of January 1 and realized that I was pretty much finished. I experience this sometimes with Song-A-Day tracks – there’s just nothing else it needs. Stick a fork in it.
Where Do I Put All of This?
I will be writing more on digital storage and archive later this year, but there’s too much to discuss in this already massively overlong post. Suffice to say that this project generated a lot of raw content. The music project ended up being only about 3GB – largish, but I’ve done worse. But the raw video? Over 100GB. My initial impulse was to throw it all on a cheap storage drive, but… why? The vast majority of it wasn’t used in the project, but it also won’t be usable in anything else. I used all of the stuff I liked. This is not a project I’m likely to revisit. So do I need to keep all of the individual shots?
If this were a corporate video or I were shooting stock footage, absolutely. You never know when something will end up being usable again. But aside from a couple of clips that I want to keep just for me (I’m thinking the birthday cake shot is my new LinkedIn profile photo), what use will I ever have for this? It’s very difficult for me to delete originals – I face the same issue with my photo libraries – but that’s a lot of space for a lot of nothing.
My compromise was to make the Behind The Scenes video below. This allowed me to pull some of the fun moments that I wanted others to see, as well as get a feel for what was going on during the project. I feel like I’m not leaving anything important shoved in a folder somewhere, and can delete without guilt.
Have a look at some of the shenanigans that went on:
Needless to say, this project would have never happened without a lot of help, so my deep gratitude goes out to Mary (management, driver, videographer, costume tech, script review, assistant editor, psychiatrist), Ron (videography and monkey wrangling), Fawne, Julie, Jade, and Meghan (videography and monkey wrangling), friends Mary (adequate boy after demotion) and Brian (key grip and stunt double) and all of the family members who made this silly project an overwhelming success!
- MOTU Digital Performer
- FabFilter Suite
- Valhalla VintageVerb
- Final Cut Pro X
- Coremelt Slice-X
- Pixel Film Works FCPX Removal
- Sheffield Softworks Vintage
- Kingluma Granularity
- Canon EOS 90D
- iPhone 6s+