Music Video Production
My wife is the singer in the family. Always has been. From her time in 3 different choirs in high school to getting her Vocal Performance degree from UTSA, singing has been in her blood.
So when she came to me and told me about this idea she had to create singing videos on YouTube, it made perfect sense.
What was less clear was exactly what those videos would look like and how much time and effort they would take to produce.
A lot. The time and effort would be a lot.
One of our projects in Garage Band with 17 audio tracks.
(At least that’s now many are visible on screen.)
An example of a final video with multiple vocal parts.
My wife wanted to create her own arrangements of popular songs, creating multiple harmony parts, recording them – both audio and video – individually, then combining them together into a single video.
You’ve likely seen similar videos on the internet, and they are quite wonderful.
While I had done some music production in the past and some small video editing projects, it quickly became apparent that this was going to stretch my skills far beyond anything I had done before.
Step 2 - Get the Music and Practice
Before recording anything, we had to acquire the music to whatever song she wanted to perform.
In some cases there were legitimate websites where the music only portion of the song could be purchased. For some of her more obscure tastes, it was simply a matter of searching YouTube to find the best sounding knock-off karaoke version we could find and ripping the audio from there.
I would then take that audio and create a project for her in Garage Band, where should record her own melody part, then use that to start crafting each of the harmony parts she wanted to create.
That process would take her anywhere from 3 days to 3 weeks depending on the song and how complex she made it.
At the end of that process she would have an audio rough cut in Garage Band of what she wanted the final song to sound like complete with all the different vocal parts.
Step 1 - Create a Recording Studio
The first thing I did was build a make-shift video studio of sorts right in our bedroom where we could do all the video and audio recording. I put up a black backdrop from my photography gear and strung some lights across it that would be nicely out of focus in the background adding a sense of depth to the videos.
Of course, through the magic of framing, you would never know that we were recording in our bedroom…although we weren’t very shy about sharing the reality of our production environment, knowing there were many creators out there struggling with their own limitations and assuming everyone else had a “perfect” setup.
I already owned some of the gear we would need, including the backdrop, a RODE NT 1-A microphone we could use, and I recorded all the video on my Nikon D750 full frame camera at 1080p resolution.
What I didn’t own were video lights. Our bedroom has some nice large windows to let in natural light, but when recording video you never want to be dependent on natural light if possible. Natural light changes dramatically throughout the day, and we were going to be recording for several hours a day across multiple days, so we needed to control the environment as much as possible so all the videos would look the same when we combined them in the final project.
So I purchased 2 Fovitec 600XB video lights using one as the key light and the other as a backlight to separate Val from the background.
Our “fancy” recording studio in our bedroom.
A bit about the process…
The recording process required two parts, both taking place simultaneously.
Val using in-ear monitors to hear the music while we recorded just her voice singing onto a new track in Garage Band.
Recording the video of her singing on the Nikon D750, ensuring the lighting was correct, she was in focus, and the framing was right so you didn’t see all the clutter just out of frame.
Later in Final Cut, I would take her clean audio recording from Garage Band and the video from the camera and sync them up.
The real challenge was that, unlike audio only recordings you listen to on Spotify and Apple Music, we were recording video of her singing the parts LIVE. That meant if she messed up on any one part of the entire 3 or 4 minute song, we had to RE-RECORD THAT ENTIRE PART OF THE SONG!
This was essentially a live performance. Yes, we could do multiple takes of the entire song, or an entire harmony part, but any mistakes could only be fixed by doing the entire part again. We couldn’t just punch in a line here or a word there like artists do in audio-only recordings.
The other thing that most people don’t realize is that when you watch a professional music video, you’re watching the artist lip-sync to their perfectly produced audio recording they did in a sound studio weeks or months earlier.
In our setup, we weren’t doing that. In our videos you are watching Valarie sing every single part of the song live, not just mouthing the words to some perfectly produced recording she had done earlier.
This was nuts!
But this is exactly what Valarie wanted. She wanted to showcase her ability to sing live, and her ability to sing multi-part harmonies live.
That’s really hard to do, but that’s what she wanted, so that’s exactly how we did it.
7 Parts! And somehow she convinced me to play guitar.
Step 3 - The Live Recording
We always recorded the melody part first.
This involved running in-ear monitors to Valarie so she could hear the music while we recorded just her voice singing the melody on a new track in Garage Band – all while I recorded it live on my DSLR.
Things rarely ever took just one take to meet Val’s exacting standards, so we would end up recording several takes of just the melody.
Once Valarie was happy with that, she would do a quick wardrobe change, then she would pick which harmony part she wanted to record. We would then create a new track in the Garage Band project and start recording that one harmony part.
This process would continue for as long as we could until it was time to pick the kids up from school, and we would pick it up the following day.
Doing just the recording of all the parts could take from 2 days to a full week depending on how many parts Valarie had written and how quickly she could sing them completely without any mistakes.
Step 4 - Mixing the Audio
At the completion of the recording sessions, I had a Garage Band file stuffed full of all the different vocal parts to the song, plus an SD card from the DSLR full of all the video recordings.
The first step was mixing and mastering the audio. Although Valarie was doing everything live, mixing the vocals was critical to get all the vocal parts to sound good together.
I always assumed I would have to jump from Garage Band to Logic Studio at some point to get the editing tools I needed, but I was pleasantly surprised to find the editing capabilities in Garage Band to be extremely robust.
The audio plugin library was extensive and just about any audio manipulation tool I needed was available directly in Garage Band, from full EQ to compression to autotune plugins to fix specific notes here and there.
I was both impressed and relieved I didn’t have to add “learning Logic Studio” to my already full plate of tasks.
Garage Band even had a Graillon plugin for the occasional pitch correction.
Step 5 - Putting it All in Final Cut
Once the audio was mixed, I would export it as a single audio file of the entire song.
I would then take that song file, along with all the videos from the DSLR, and import them into a new project in Final Cut.
The first step in the video editing process was to lay the audio track down as the baseline, then one-by-one layer on each of the video files matching them up to the timeline precisely so they would sync up with the audio file.
Basically, if any one of the video files wasn’t aligned correctly, Valarie’s video mouth wouldn’t match up with her audio voice. Not cool.
Final Cut actually has an automated tool to help sync up video and audio files. Using the audio waveform from the video file it tries to match that waveform precisely to the separate audio file waveform. I found this tool to be pretty accurate, but I honestly felt I got better result matching up the audio and video files manually.
Perhaps I’m just a bit neurotic like that.
A project in Final Cut with multiple videos synced to precisely match the audio.
Step 6 - The Animations
Now that all the videos were where they needed to be, it was time to plan out the animations and resize the videos.
At this point all the videos are full-screen, meaning I could only see one video at a time, whichever was the top-most layer.
Now I had to listen to the song and find out when there were multiple voices and how I wanted to display those corresponding videos on the screen. Did I want the additional videos to slowly animate on the screen? Did I want the main melody video to animate to a smaller portion of the screen to accommodate the other videos? Or did I just want to abruptly switch the sizing of the videos and have the other voices instantly appear as their vocal part rang in?
There were a lot of creative choices at this stage of the project and I had a lot of fun figuring out the best way to build from a single voice/video up to an array of 6, 7 or even 8 voices/videos on screen at the same time.
Ultimately, Valarie had final creative sign-off on anything I did, so while she often loved the choices I made, she occasionally had some notes and changes she wanted implemented.
Once those changes were done, it was time to export the final video and upload it to YouTube.
But wait! There’s more.
Of course, uploading the official video to YouTube was just the beginning.
Specific shorter cuts of the video had to be created for Facebook and Instagram and other social channels to promote the new video. Each video had its own formatting rules and length requirements and you had to create a post and a story and a reel and it was best to schedule them out over the next 24 hours to get the most engagement.
But I suppose that whole process would justify another post entirely.
Epilogue
Over a period of about 9 months we produced and published 11 videos. It was A LOT of work, and extremely gratifying to work on.
But, ultimately, after 9 months of hard work, Valarie had some other projects she wanted to pursue.
I was kind of sad to see the effort end, but I know it was a ton of work for Valarie to not only find the songs, but then memorize them, write multiple custom harmony parts, practice those parts for days on end, then ultimately perform them live so we could record them.
The technical work I did was truly minimal compared to her efforts.
In the end, I’m glad we were able to do this together, each of us using our specific skills to do something that neither of us would have been able to do on our own.
As they say, teamwork makes the dream work.