Why We Started This Journey
The Talk Shows
We started with talkshow podcasts Talkily and Trekily. Talkily was fun but despite our best efforts even some of the more general topics could become dated after a few years. Some episodes like pens vs pencils may still be relevant, but an episode we did on Streaming Services for example, which featured us talking about the soon to release Disney Plus would be immediately dated for anyone listening to it today.
Trekily allowed the Star Trek fans in our group to talk Trek. We were fairly excited about it when we started developing it. As a new Trek show had just released, our first episodes revolved around reviewing episodes. The problem with review shows is that they have to be recorded, edited and then released right away to stay relevant.
We started looking at retooling both shows but as a number of us had creative writing projects or story based projects we opted to try to create a fiction based podcast.
Inspiration
Wolf 359 was a podcast that Cody had introduced to our friend group sometime before our Podcast adventures. They had done such a great job with their show that it inspired us to try to spearhead a fiction story ourselves. From a production standpoint, it was our hope that a story based podcast would have a lot more longevity even years after it had been completed and could be enjoyed regardless of when you listed to it the first time.
This is also a fun creative project that the majority of us have essentially volunteered for. Being able to devote our time and effort into a labour of love and then release it so that it could be enjoyed with a beginning, middle and end was more appealing then getting into the grid of having to pump out endless content.
That doesn’t mean that we’re not hungry to do another season of Don’t Ghost Me or more Storily Tales, but being able to tell more stories, explore different themes, characters and stories is something we’re excited about.
STORY DEVELOPMENT
Podcast from a Hat
The project initially started as a “Podcast from a hat” idea. We sent out a Google form asking for scenarios, character ideas, themes and objects. We also queried how involved everyone wanted to be (are you willing to be voice a main character, how involved do you want to be in that character’s development, do you just want to do lines, only a cameo, etc). That’s partly how the idea of a spy, magic, ghosts, a blind date gone wrong, an island, the Catacombs, the amulet, the fear of water, lost memories and a book of doom all ended up together.
Cody, Syd, Brenna and Alvin then started meeting over Zoom calls to start piecing together and building our world, cast of characters, themes and backstories.
Online Workshops and Community Writing
We did about three or so different versions of our story before getting to the workshop stage. In the workshop process we essentially opened up our story to our larger group so that we could flesh out ideas and work on dialogue.
As it was in the middle of Covid, we did it mainly on weekly Zoom calls (later Google Meets).
Opening up the world you’ve built to a larger group can be really intimidating, especially when your ideas, characters and themes are challenged and scrutinized. Going through that process though really allowed us to focus on what was important and cut out what wasn’t working so that we could create a more balanced story.
We went through a few different iterations of our story and fine tuned, tweaked or altered characters significantly based on feedback from the overall group. While quite a bit changed between the start of the workshops and the end, the overall story arc and beats remained consistent.
PRODCUTION
Recording Setup
Alvin built two little recording booths in his basement in the area where we previously recorded Talkily along an inner wall. The walls were built with standard 2 by 4s and stuffed with acoustic fibreglass insulation. It was a bit cramped and could get hot after about an hour but it got the job done. We used some commercial grade professional acoustic foam for the walls to help with sound dampening which was supplemented with cheaper 2 inch foam from Amazon. There was a window between the two booths and then smaller ones on the outside walls so with the idea being that Alvin could see who was recording their vocals.
When Alvin ended up moving, the booths were rebuilt with a proper acoustic conscious ventilation system to keep it more comfortable. We were also able to expand to 3 and a half booths (with Alvin’s direction station being used for Midnight lines). To help improve sound dampening the window was removed and monitors with webcams were installed. Because of the lighting we really couldn’t see each other through the windows anyway and providing a mouse with touchpad allowed the folks recording the ability to see the script and update any newly ad-libbed or updated lines as we went (we originally had cheap tablets).
Music, SFX, Equipment and Tools
For Mics we started off with three sE V7s but then later upgraded to three SM7b with sE Dynamite’s to increase the gain and just one sE V7 for the direction station. We used a Scarlett 18I20M3 as our audio Interface. Headphones are all Audio-Technica.
Google Drive, Sheets, Docs and Meets were used for cloud storage scripting, collaboration and tracking.
One of the hardest things to figure out initially was where to get the music and SFX as well as what DAW to use. There were a lot of open-source or free for commercial use places but the licensing aspect where ever we looked always seemed to be super confusing or too intimidating. During our Talkily production we came across Art-List and really liked having a one-stop shop for music and sound effects with a simple licensing structure.
Our podcast is hosted on Spreaker which allows us to manage ads and distribute to all the major platforms.
Alvin’s been using the Adobe Suite for more than a decade so he stuck with Audition for editing the podcast. There are other options (Kyle Masters the episodes in Pro-Tools) but it’s the beast that he knows.
Editing
Alvin would initially cut the episodes in Adobe Audition and then have Cody review them for any final notes. We’d then export them for Kyle to Master in Pro Tools.
Recording Phase
We produced about six test episodes with the current script so that we could get a feel for themes, our characters and get a sense of real world process of recording, editing and the time to complete an episode.
After completing them we took a step back and refined some ideas, characters and overall story beats. One of the biggest changes was the positioning of the flashback/memory episodes alongside the present day episodes.
After the test phase we recorded over the course of about two years. We’d make occasional edits to dialogue and altered some scenes as needed based on how we played off of each other. Editing started as soon as we had completed the dialogue for an episode and went concurrent to recording. Each recording session typically lasted about two hours.
HOW REAL A.I. HAS AFFECTED DON’T GHOST ME AT MIDNIGHT
A Message from DGMAM Co-Creator Alvin Cadonic
(and ironically, the voice of Midnight, a character who happens to be a sentient A.I.)
When we started working on Don’t Ghost Me At Midnight around 2020, A.I. tools like ChatGPT had not yet come out. That was the dream though! I can’t count how many times I thought to myself, wow, wouldn’t it be great to type something into a box and have the computer realistically read out the lines for me? Or for me to create custom music without having to worry about all the time it would take to develop it? Wouldn’t that be great?
Shut up Past Alvin.
Since we started building, recording, editing and now released Don’t Ghost Me At Midnight, A.I. usage has soared and those tools have become available. Adobe, Art-List and Google are tools we use that have gone all-in on A.I.
99% of what you’ve listened to in Don’t Ghost Me At Midnight was created by a real human. We developed it, wrote it, scripted it, recorded it, edited it and then released it ourselves. So… why not a 100 percent? Because we use Art-List for music and sound effects, we (and nobody really) can guarantee that 100 percent of what we used was created by a real human. When we started, I feel like I could be certain, but now, because of how convincing A.I. can be, it would be dishonest to claim that A.I. wasn’t used in some way, especially since Art-List has got all-in on A.I.
I use filters and effects in Adobe Audition and I’m not sure how much of that was A.I. driven in the backend? When we needed additional voice work done, going onto Freelance sites like Fivrr was a bit of a nightmare because I couldn’t tell if someone was a human or just a person typing our script into a prompt. We wanted a real person to record their voice or a variation of it. If we wanted to use an A.I. voice, we literally have Art-List that can do that for us. I’m fairly certain the folks we hired were real (and they all seemed genuine in communications) but because I haven’t met them in person I can’t say for certain that at least one of them wasn’t a real person.
Transcription services seem to all use A.I. and we have used the free service provided by Spreaker, though in our defence there, we use it to generate the SRT file and have to manually adjust the majority of the lines and [noted effects] anyway using our existing scripts, since it’s not the best at grabbing the correct words (likely due to the music and effects used).
Regarding artwork, we’re either using stock images that were labelled as authentic from Pixabay, creating them ourselves in Photoshop, or doing what we can to ensure that the artists that contribute to our project are creating it themselves instead of using a prompt.
We’ve made the choice to try to avoid the use of A.I. in this production because of the moral issues around its use (models learning off of creative work without permission, people losing their jobs to A.I.). Selfishly, a part of that decision was to avoid blow-back for purposefully using A.I. (and I’d argue, anyone stating they do a thing for 100% moral reasons isn’t being 100% honest because humans are imperfect beings).
The case could be made that if we really wanted to be A.I. free, we could cut out Art-List, Adobe, Google, use open-source resources that specifically don’t use A.I, go to extreme lengths to vet voice actors, etc. But I’ve committed to the production pipe-line we currently have and don’t have any plans to change anything at this time. I like Art-List as a service, I’m used to the Adobe ecosystem and don’t have the patience or time to learn another DAW, and I did my best to vet that the Fivrr voice actors are real people.
If you’re really Anti-AI and/or have been personally affected by it in a negative way, we can understand that our usage of certain apps and services may be disappointing as we’re still indirectly or directly supporting companies that are in, or are drivers in the A.I. space. If your pro-AI then hopefully you can understand why we’ve taken the steps we have to try to keep things as authentically human as possible.
Regardless, I wanted to be up front about our A.I. usage so at the very least, we can say that we’ve been open and honest about it.
IT’S BEEN MAGICAL
Bringing the world of Don’t Ghost Me At Midnight to life has been one of the most challenging and rewarding experiences in my life. It’s something that would not have become a reality if it was not for the group of talented, dedicated and wonderful human beings that lent their time and effort into making this something truley magical.
On behalf of the DGMAM team, thank you dear listener for giving us your time. We hope that you’ll enjoy / have enjoyed the adventure we’ve conjured up and that it will be something you’ll still be happily familiar with years from now.
At the end of the day, there is nothing more magical than the things we can dream up with our own imagination and the people we share our time with.
– Alvin Wayne Cadonic –
