My Journey editing an Indie Game Interview
So a few months back In I believe mid may, I did an interview with an indie visual novel developer named Soft Poison Studio on their game demo Scarlet Hunt.
Naturally I had a great time doing the interview and learning about the game and its creation, it was after that the struggle came...
You see my stupid ass took weeks preparing for the interview only to make 1 huge mistake.... I forgot to separate the audio tracks on OBS when setting up the discord browser source
You're probably thinking okay no biggie, and maybe it wasn't but the goal in what I do is to help creators. I'm a creator with little to no visible community so the fact that anyone even gives me the time of day let alone set aside time for me is huge.
That being said throwing a video in an editor and making a trim with little to no additional edits and calling it done felt wrong. Realistically It probably isn't wrong, its probably the normal way of doing things but logic is beside me and I always feel as is isn't good enough and like I need to do more.
Thus began my 3 month long audio editing journey .....
The low storage alert I got before I even started editing should have been a sign that this would not be a simple journey lol
Low storage is truly the ultimate enemy of all mankind!
My logic at the start was that I just needed to go in and separate the audio lines into two separate tracks.
Track 1: My Dialogue
Track 2 Discord guest Dialogue
A tedious chore sure but completely doable. The most difficult part was supposed to be recording an intro that kept the viewers attention and didn't make viewers immediately click off the video upon hearing my voice.
No...that became the second hardest thing.
My viewer stats...are honestly kind of pathetic and I can only assume it has something to do with my audio therefore I came to the conclusion that I needed to give my audio more attention cause I'm obviously not doing something right.
This meant any unnatural mouth sounds in dialogue were non negotiable to get rid of, and since I unfortunately have dry mouth I tend to make a lot of those....
That being said this naturally meant that it would be necessary to go in and edit out any pops, clicks, bzzz's, and microphone knocks and since I knew little to nothing about audio editing I figured the only way was to zoom into the timeline and manipulate the dots and remove the mouth sounds manually... One... By... One
That time around when the interview happened when I said I'd have the video on YouTube by the end of the week,
....Yeah that wasn't happening.
This...was the true beginning of the past 3 months.
Isolated mouth noises were pretty straight forward in how to get rid of them. You just trim the small section and delete it or make it silent. Its when they're mixed in to words that things become tricky.
I thought that while time consuming and tedious the solution was simply to go in and trim out the sounds and use the dots to smooth vocals out. Smooth waves mean smooth voices right?
WRONG!
But I didn't realize that until AFTER I went through the entire thing. By that point I didn't remember what I even changed and had no idea how to un-change it so in the end after trying a couple things that just made it worse, I couldn't figure out a solution so I just started over from the beginning....
~~~
De-esser, Normalize, EQ, Noise Reduction, Compression, Frequency
I didn't understand any of these terms aside from compression in which I understood the idea but not how it worked or when you're supposed to use it nor the effects the intensity level could have on certain vocal levels compared to other spoken vocal levels. I learned this lesson especially hard when I later first tried to use EQ as a solve all solution and then once more even later when I thought Dialogue Leveler could actually be used as a solve all solution.
Based on this it sounded like it should have but it didn't. They fed me a lie!
Okay maybe not exactly but you know, hopes dreams shattered, all that jazz.
As far as I'm currently aware there is no solve all solution just a 'Solve Some' solution with additional steps.
Round 2
So as we found out earlier, moving the dots has consequences.
Cutting an audio clip in the wrong place can also have consequences.
PAUSE!!!
Completely unrelated but I literally just found out that Ctrl + Z undoes your last action here in tumblr just like it does in Audacity and Davinci Resolve!!! I accidently deleted some text and I literally did it on instinct!!!
Anyway back to the blog!~
I kind of understood decibels since OBS uses them to show the audio levels but that was about the usable piece of knowledge I had and understood going into this.
In order to smooth out the audio I eventually figured out I would probably need to use audio filters.
The controls on the audio filters adjusted based on frequency, while I knew what it was I had no idea how frequency worked.
Honestly I straight up didn't understand how voices worked (still don't completely understand them) but as we found out dots were not the ultimate answer and that left me with the conclusion that it was time to experiment with audio filters and make a massive amount of google searches.
Our voices are different so this means we would each need different things, the real question became "What do each of us need" and "How do I use it"
Oh and of course "What does (insert word) mean?"
While manipulating the dots does help in some situations, it doesn't help in all situations.
A frustrating lesson, but a lesson none the less and a valuable learning experience.
I'll spare you the step by step experience but I ended up experimenting in my video editor with Noise Reduction, De-esser, Compression, Dialogue Processor, Vocal Channel, Frequency Analyzer, Normalization, Dialogue Leveler, Equalization, and Dynamics.
While I can't say I understand enough to give you a technical detailed explanation I certainly know a hell of a lot more than when I began.
I'm a bit more confident in my understanding of the general idea in how these things are used and in what situation I should use them for but there's still a lot I don't understand.
EQ was the thing my experiments focused on the most since it was the solution to a majority of what I was looking to smooth out.
The biggest factor in how I learned to use this even remotely effectively lays in these images.
A lot of trial and error during all this but still an incredibly useful tool, even more so if you actually know how to use it.
The biggest game changer came when I fucked up the audio because I used EQ individually on a bunch of separate clips but when I played back as a whole ended up sounding unnatural and uneven.
I learned the one thing that if known before could have allowed me to get this done in one month....
YOU CAN EXPORT THE AUDIO ON ITS OWN!!! And it doesn't take forever to export or take up Gigabytes of storage like exporting a video does.
Once I learned this I was able to throw the audio tracks into Audacity and clean them up in a dedicated audio editor.
I did accidently export a track with an unatural amount of compression in one area that I forgot about and didn't know how to remove anyway so that area I ended up redoing the cleaning work in audacity a second time which was frustrating.
But it was during that second time I had to clean that piece of audio that I made my second biggest game changing discovery!
Introducing my lord and savior, the spectrogram!
I'm a visual learner you see so if I can see it I can understand it so much better than if it were given to me in a worded explaination.
(The bed ended up being a surprisingly productive place to work)
Mouth click have a certain look and the spectrogram editing makes it easier to identify and remove them. Once I learned of its existence and how to use it I got more done in a week then I did in multiple months of audio editing experimentation!
While mouth clicks have a particular look to them some of them are more difficult to find and get rid of than others plus there are still other sounds in the audio that I didn't understand how to identify so it wasn't all blue skyies and smooth seas, I still had to learn how to remove and experiment with these other sounds.
It was here that the third and biggest game changer was discovered.
The Audacity declicker pugin! πΌπ½β¨
If using the spectrogram got me more done in a week than in a few months then the declicker plugin allowed me to achieve more in two days than I did in multiple months!
If I had known what I know now I might have actually been able to get the video done in a week like I originally said, if not that then at least in a month.
Once again its not a solve all solution but it was a solution to the thing I was struggling with the most and the most self conscious about. Of course the audio still needed some filters, compression and EQ applied to it which took some time and more experimentation but the overall editing time was dramatically cut!
The video is still in no way perfect and probably still has some flaws but the undesirable sounds that I believe to be the main cause of my low viewer rating are I believe for the most part gone and that in itself is a massive achievement and thing to be worth celebrating.
I do have slight fear that I listened to the audio so much that it affected my judgement on how the audio actually sounds when I watched it before exporting it.
Oh man...new fear unlocked!π°π©
While exhausting and time consuming I know more now than I did before and that in itself made it worth the struggle.
I have hope now for improved quality in future editing projects and the idea of editing audio feels much less intimidating than it did before.
I feel like I might actually have a chance now at getting things done faster and not feeling like I let someone down!
That feeling, that hope, is worth everything!!
Anyway if you want to check out the interview video you can find it in the link below! If you haven't checked the game demo out you should! You can find it on Steam and Itch.io
