Grace & Steel Ep. 85: Production Notes

I encountered few problems getting Episode 85 of Grace & Steel shipshape and posted. When Grace and I co-host a show, as opposed to Grace appearing with a guest, things typically go smoother in post-production simply because I have much experience working with files recorded on Grace’s system and my own, of course. The only real hiccup with Ep. 85 was when something on Grace’s computer shut down his Audacity recording. This has happened a few times before, and we’re not why sure why. It’s no big deal. It simply means that he must stop talking, save his recording by exporting and restart the program. He then sends me two Audacity files instead of one, and I match these against the Total Recorder wav file which Grace had kept running even as he fiddled with Audacity. I edit out the gap.

For curious audio nerds, I’ll run through my editing process. It’s important to keep in mind that, with the exception of the mastering process, this is a bespoke process. For mastering, I paid a few bucks to take an online video course on the subject, worth every penny. As for the rest, there may be better ways of doing it. I’ve watched YouTube videos on producing podcasts that recommend various utilities to do this or that; for instance, one utility automatically cut out a substantial portion of dead air between phrases. While this would save time, it would result in a loss of control over the edit. And so, I prefer my laborious method. I stress, however, I learned this on the fly, and the process will likely change as I learn more.

Here is the process step by step:

1. As I download all files sent to me, I stick them in a working folder separate from the Audio folder used by Sonar. After each change I make to an audio file, I save it under a new name with some kind of indicator that will tell me instantly what I’ve done. For instance, if I treat a file with noise reduction, I’ll put “nr” in the file name. This helps me if I, heaven forbid, have to do any punch-ins or match files, as in the case, mentioned above, when Grace’s recording gets broken in two.
2. All files are then converted, using WaveLab, from various rates and formats to wav files with a sample rate of 44,100 Hz and 32-bit float (the Audacity defaults).
3. I then close WaveLab, and open each file in Audacity, and treat them with a noise reduction utility.
4. Then it’s back to WaveLab where I first check average volumes and, if need be, boost the files with a hard limiter to ensure sufficient audio to work with in the mixing program. An average volume of -25 to -20 dB will suffice.
5. I then go through each file individually, removing background noise. This takes a long time and is really boring. However, it should be noted that none of this editing changes the podcast length, which must remain approximately the same as the original file length to enable synching it up later in the multitrack program. Again, as mentioned above, all conversions and treatments are saved as unique file versions in the working folder.
6. Then everything is moved over to the multi-track program, Sonar, where I start laying out the show—music, guide track, intro, Grace’s voice, Steel’s voice and an effects track along the bottom if needed. (As I load up the various tracks, Sonar saves new versions of each file in its own Audio folder; this is important to remember because it means I don’t have to save the files in the working folder at the end of this process, merely the Audio folder.) Next I must make sure everything lines up correctly. Things can get out of sync. So I have to spot check the mix, slice up the mono files and move them a little this way and that to match the guide track: the stereo Total Recorder file.
7. Then comes the big mixdown. Usually it takes two to five tries to get this right, adjusting volumes mostly but sometimes adding a bit of compression and eq (equalization). As mentioned in a previous post, I split the stereo track with each person 30% left and right, as opposed to hard panning. The goal is for an average volume of -19 dB to -18 dB for the person who speaks the most. (Note: I save the mixdowns in a separate folder.)
8. Then it’s time for the final edit in WaveLab. This is where the chopping is done, where the podcast is reduced in length. As I do the final edit on my desktop, I also compile the links for the Show Notes on my laptop.
9. Once the final edit is complete, it must be mastered. This is the application of three utilities, always in this order; compression, eq, and a hard limiter. I use the utilities made by Steinberg, maker of WaveLab. For the final file, I’m shooting for an average volume of -16.5 dB to -16 dB. (Note: When we started the podcast, I had read somewhere that some Scandinavians had decided that podcasts sounded best with an average volume of -15 dB. So I aimed for that. About six months ago, I read somewhere else that everybody was now going for -16 dB, which suits me better.)
10. Once final mastering is completed, there is one final mixdown and compression to mp3, using a dithering utility along with the three mastering utilities. I then place the final files, one wav and one mp3, in a separate folder. The wav will be used to create the YouTube version of the podcast, and the mp3 will be posted to SoundCloud.
11. Because all important files have been saved elsewhere, I am free to delete the working audio file which is now massive in size. I also delete all the failed mixdowns.

At the top of this post, I mentioned the gap caused by the inexplicable shutdown of Audacity. For those who have read through the 11 steps above, I’ll explain that this cut should be made during the final edit. Bad experience has taught me not to do it in the multitrack.

There are a few other things that must be done before the podcast goes out to the world. For instance, as these final mixdowns are rendering, I begin writing the 2Kevins page for that episode with all the Show Notes.

After the mixdowns are done, then it’s back to the desktop and into Photoshop to create the cover art, which is then added to the mp3 file in a metatag program (along with a lot of gobbly gook file info). This mp3 is then posted to SoundCloud. Once posted, I add the SoundCloud file links to the 2Kevins page and go live with that. And then I start publicizing the show.

Next, I take a few screenshots of the episode’s SoundCloud and 2Kevins pages to be used in the YouTube version of the podcast. I do this in Sony Vegas, using the final mastered wav file, of course, not the mp3 file, and I change the picture every 15 minutes, which is recommended for YouTube podcasts. Then the file is rendered as a wmv file, and I upload that to YouTube.

At the end of this extended and complicated process I sit back and think, why am I doing this? It’s nuts. But then someone makes an encouraging comment, or makes a donation, and I think, alright then, let’s do it again. It’s encouragement like that that keeps me going. After all, when I think about it, what else do I do that makes any bloody difference in this sweet old world? Not much. And I still enjoy “kicking the pricks,” as we used to say in the Alberta Report newsroom.