Last time we talked about what we wanted to accomplish with our podcasts: limited dynamic range, minimum file size and maximum quality. Today, we’ll talk about how we get there. Like most things, it’s a multi-step process to achieve the best results. But before we have anything to edit, compress or publish on the interwebs, we must first record something.
If you are working with a digital console and are using a virtual soundcheck system, you are already in great shape for recording the message. That’s how we do it at Coast Hills, using our RME MADIFace to run the audio directly after the A/D conversion into the MacBook Pro and on into Reaper. As I wrote recently in Automating Reaper (Again), I record a two-track board mix of all three services and a discreet track for our pastor on both Sunday AM services. I use the discreet track for the podcast; the two-track is backup only.
Since we’re recording speech (for this purpose anyway), I don’t worry about getting to 192 KHz or anything crazy. 48 or 44.1 KHz at 16-24 bits is just fine. Our system runs at 48 KHz, 24 bits, so that’s what we record as a series of WAV files. WAVs are uncompressed, so the quality is quite good (AIFF files would also be a good choice). You want to record uncompressed if at all possible.
If you are using an analog system, don’t fret. Use the direct outs — or in a pinch, the output of the insert jacks — to come directly out of the pastor’s mic channel to your recorder. You really want to pick off the output before EQ, compression, or other processing. The reason for this is that most times you are making EQ and compression adjustments on your console for the room, which is where most people are listening. However, those same settings may not work for the recording, and there are advantages to making some EQ and compression choices specifically for the podcast.
You can record to a CD (we do an archive CD of our 9 a.m. service each week for backup), but I really prefer to record straight to the computer since we’ll be editing and processing the file there anyway. Even an inexpensive USB interface like a Lexicon Alpha (about $65) will get you great sounding direct recording. Combine that with a laptop, Mac Mini or inexpensive PC and a copy of Reaper and you’re in business. If you’re stuck with a CD, rip it into DAW (Digital Audio Workstation) software for further editing and processing.
What works in the room may not work online. You will have to do some experimentation here, but I tend to high-pass our pastor fairly high (up around 130-140), and boost the upper mid’s by 1-2 dB. I’m trying to add a little bit of clarity to make it easier to listen to in loud environments. But, be careful here because you can easily make it annoyingly bright.
I suggest you try some settings, encode a section of the sermon and listen to it on several platforms to see how you did. EQ settings that work in your 7506s may not work on a cheap set of computer speakers, so check it out. And don’t forget many people will listen to the podcast with Apple’s cheap, white earbuds.
Since my pastor’s voice is recorded pre-EQ, I do some subtle changes to make him easier to listen to. Then I hit it with the compressors. Currently, I’m using R-Channel from Waves to do both EQ and compression. I will say the R-Channel compressor is one of the more transparent ones I’ve heard. I routinely have it hitting 12+ dB of gain reduction and it’s really tough to hear.
Before I had that plug-in, I achieved results almost as good with a combination of ReaEQ and ReaComp; both plug-ins come with Reaper. Most DAWs have basic EQ and compressors built-in, so play with those first before you go spending money on plug-ins. However, it would be worth it to sign up for Waves' mailing list. They often do super deals on individual plug-ins and you can pick up one or two that will rock pretty cheap.
The final step is to use a mastering limiter to really clamp down the dynamic range. I was using JS: LOSER MasterLimiter (included w/ Reaper) for quite a while along with ArdazMaximzer5. The MasterLimiter allows you to set a maximum level (I went with -.01 dB) that it will allow; it’s a brick-wall limiter so nothing gets over that. It will also do some compression to keep the signal level up. The Maximizer does some other magic to raise the overall level without driving it over the limit. That combination worked really well and sounded pretty good once we got it dialed in.
Then I picked up the WAVES L3 UltraMaximizer. And that was pretty much that. After setting a few sliders, I can pretty much crank the level like crazy and it sounds amazing. I showed this picture last time, but you can see how little variation in the waveform we have on the rendered file, indicating very little dynamic range. If this were music, I would be upset, but for a speech podcast, it’s about perfect.
Most DAWs can render out to an MP3 file. If you have the option to use the LAME encoder (which you have to do in Audacity or Reaper), use it. It’s a great encoder that produces better-sounding MP3s at lower bitrates than most other encoders. As I said last time, I use the 48 kbps CBR setting and render in mono. I haven’t had any problems with mono files anywhere, and given that it’s spoken word, stereo unnecessarily doubles the file size.
I got to these settings (and all the ones I didn’t tell you about) by doing a lot of experimenting. I intentionally didn’t show you all my EQ, compression and limiting settings because they don’t really matter; they are all specific for our pastor. If you spend a few hours working on a chunk of the message getting the processing settings right, then tweak your rendering settings, you’ll end up with great results. Then don’t forget to save those settings as presets so you can use them next week.
I’ve mentioned a ton of stuff in this article, but here is a recap of what I recommend for this process. If you have something else that works, by all means, keep using it. If you’re looking for a place to start, consider this list.
If you want to hear the results of this processing, you can check out the Coast HIlls media page. Listen to the MP3 files. Those all have processing similar to what we’re taking about here. Keep in mind, since we have most of this stuff in presets, it takes about 5-10 minutes to edit, process, render, upload and post our podcasts. Once we did the hard work, the weekly stuff is easy.
Image • ChurchTechArts.org