In a comment I've made a little bit ago, I mentioned that I was tasking myself to discover music that was played on an old radio program that I listened to from the 2000s. And it is looking to be a lot more prolonged and tedious than I had thought. I've been able to find a program that has done an amazing job at removing the host's voices to where, I can't tell where they start or stop talking, I get hints that there were points of voices being there, but it's non-existent.
I've tried before in the past to use Audacity, but being that all recordings were done in Mono and not Stereo, no matter what I tried, the voices would still remain. So now that hurdle is done with, the next task is to go through all 139 episodes and all episodes average 1 hour to 2 hours. That's a long time if you're doing radio or podcasting, it's a lot of talking to do. Then it's a matter of listening back and forth at one points certain songs begin and end, marking times to point them out with.
I might pick out some standout favorites, episodes that contained the most songs that I would have wanted the most from them. Then once all of that is figured, the next course of action is to clean up the sample audio, because most of these episodes were recorded in Mono so there's going to be a lot of distortion and muddiness.
Then once all of that is done, the next challenging task is, actually finding someone who'll be able to identify what is played. I don't know electronic/techno music too well, I'm not entirely familiar with artists outside Daft Punk, Celldweller, 3Teeth and Pendulum to name a few. The only thing that sortof helps narrow things down is that they were all played on DI.FM at the time, so it may or may not help.
From there, it's just hoping I find them out there online.
It's a big project, but I've listened to these episodes for 18 years now and what kept me coming back to them besides nostalgic purposes, was the music played in them that never got identified.
This sounds so fun. And overwhelming. Truly the hallmarks of a major personal project!
There might be a way to use technology to help with some of the song ID.
MusicBrainz Picard is a free service with apps and an API you can call to identify a sonic fingerprint of a song. It might be able to help you, but I don’t know if the audio artifacts you’re cleaning up would interfere with the matching.
There’s also apps like SoundHound and Pandora that were all the rage. They are half decent at identifying songs even with a lot of background noise, so they might have some extra processing that does a better job.
But that all assumes you’ve already separated all the recordings into individual tracks…
Too bad DI.FM didn’t publish their track lists. That would have made life easier!
I think MusicBrainz Picard is more for local tracks you already have on your drives. Like it's an advanced version of what Windows Media Player would do with downloading metadata when you're using it to rip CDs.
With EDM-type of music, there's a wide array of artists, remixes and versions of songs that it can be difficult for even some of these programs to narrow down. So the samples I have to make, have to be clear enough to be heard to be identified and even then, there's a low chance.
I haven't even started separating tracks because there's show segments I have to identify and remove, there's figuring out when a song starts and when it ends. It's a process, for sure. Not to mention, the time available to me to even do all of this. It takes over 45 - 50 minutes to have Ultimate Vocal Remover 5 to go through each file. Then it'll probably be like another hour and a half to figure things out.
I don't think the responsibility lies on DI.FM than it did on the hosts of the show, I mean, I tried looking up archived show notes when they were on Rant Radio and nothing could be found. It wouldn't have hurt to credit what music you're playing but, this was at the time, a one of a kind show that was very makeshift and took its time to find its identity. So I think the hosts and everyone working on the show, haven't had a thought of letting listeners in on what was being played.
I mean, maybe they did at the time? But I'm not sure because I was not an active listener of the period of which I'm targeting the music being played the show is on. I discovered them in 2008 and the shows I'm doing are 2003 - 2005.
So, it is what it is, I'm doing what I can with what I have available to me.
Also there's AudD, that's pretty neat for that purpose.
Definitely will be trying that, sounds and seems very interesting.
This is awesome!
I love that they position themselves as identifying music to catch “pirates”. Seems only fitting that it can also be used to help pirate music.