I’ve updated this with a lot more info over on my Fortuitous Blog: Everything I’ve learned about Podcasting
A lot of people ask me how I do the MetaFilter Podcast (warning: the podcast makes no sense to anyone outside of MetaFilter uberfans). I know they don’t mean “how do you do it man, you’re making magic over there every week!” but rather “what software and hardware does it take to make a decent sounding podcast?” After almost a year of regular podcasts and trying out different software and equipment, I’ve gotten the workflow down cold and I wanted to share the my way of making a good sounding podcast on the cheap. This works perfectly well for me being in Oregon and talking to my friend Jessamyn in Vermont over Skype, recording at both ends, then tossing it all into Garageband to complete the podcast. I read a lot of podcast how-tos when I set out to do my own, and almost all of them are mired in technical details about microphone quality and USB vs. mixer board audio wankery. Most every tutorial about doing a podcast interview focuses way too much on studio-like sound quality achieved through your equipment instead of through software and a bit of clever thinking. So without further ado: How to record a good podcast between two mac users on the cheap Software required:
- Skype (free)
- Call Recorder for Skype ($15/each or 3 for $30, plus free 30 day demo)
- Garageband (included with most Macs, or $79 for iLife)
Hardware required:
- Logitech USB headset microphone ($35 from Amazon)
Though you might have heard bad phone interview podcasts with Skype before, having Call Recorder running on both sides of your interview will mean your interview partner will have a crystal clear recording just like yours. The cheap headset microphones are brain-dead simple to use on a Mac (plug-in, change audio prefs to use the headset for input and output, adjust the recording level) and produce perfectly good vocal recordings. I’ve used $250 higher-end microphones and had little audio quality improvement. This process assumes two people, each running Skype, Call Recorder, and having a USB headset microphone. The Interview Recording:
- Start a Skype chat between you and your partner
- Both parties hit the record button on their Call Recorder (I record on high quality, low compression AAC)
- Conduct your interview normally
- When interview is complete, end call, stop recording
- Call Recorder includes a directory of mini-apps called Movie Tools. Have your partner locate their recording file and tell them to drag it over the “Split Movie Tracks” application
- Have partner upload Track 1 of the split movie files to a server you can download the file from
Assembling the podcast in Garageband:
- Drag your copy of the interview recording over Split Movie Tracks to turn your recording into one file for each side of the Skype conversation
- Drag each resulting .mov file over another Movie Tools app “Convert to AIFF”
- Drag your partner’s half interview (that you downloaded from them) .mov file over Convert to AIFF
- Open Garageband, start a new podcast
- Duplicate one of the vocal tracks (my partner is female so I duplicate the default female track
- Drag your own Track 1 AIFF track into a Garageband track (my goes into the default Male Voice)
- Drag your own Track 2 AIFF track into Garageband, perfectly aligned with our Track 1 (this ensures the timings are exact for each side of your own interview recording)
- Drag your partner’s Track 1 AIFF track into the duplicated track in Garageband
- Garageband quickly analyzes each track and makes visual soundwaves to go with each track. “Line up” your Track 2 and your partner’s Track 1 audio files. The peaks and flat quiet area should look really similar (click screenshot below, view notes on the image itself)
- Once your partner’s vocal track is lined up (press play to hear all three tracks and your partner should sound like an almost perfect echo from their two tracks), delete your own Track 2 track. You now have two high quality recordings from each respective source, ready for continued editing into your podcast (you can level out the volume if one person was louder, clip out pauses and coughs together, etc)
How does it sound? To give you an idea of how it sounds, consider the following three sample recordings. The first is the worst possible: recorded Skype conversation where I dialed out to a phone and recorded the entire thing on my end (mp3 sample 1 96kbps) Second, here is what a standard recorded Skype call sounds like, where I recorded both sides of the conversation on my end, so my partner was recorded through Skype and even on my high bandwidth fiber connection, it does have artifacts (mp3 sample 2 96kbps) Third, here is the same interview segment as the second part, but with my partner’s local recording track thrown in and my recording of her track thrown out. Much better and to me, sounds like we could be in the same room, even though we are 3,000 miles apart. (mp3 sample 3 128kbps) Conclusion The basic premise of this approach is you can record a Skype interview without actually needing/using Skype. You are actually recording audio on each end independent of Skype, so you won’t suffer any sound quality problems due to Skype transport. So that’s it, for about $100 or so, you can have a pretty damn good podcast that sounds like two people sat in a room together talking and recording, even if they’re on opposite sides of a country.
Radio stations do a variation on this all the time. When they can’t use an ISDN line for a live studio interview, they’ll just use a regular phone line, record in each studio, and then ship the audio over to one studio to sync in post.
LikeLike
Yeah, I’ve done a NPR interview that way before — I drove to my nearest public radio studio and we conducted it live from his NPR studio to the one I was sitting in, then they pieced it together later for air.
The beauty of the approach I describe is that you can get an output file that sounds almost as good as an expensive thing like maintaining two professional studios cross-country for about $100 with just a couple laptops and some software.
LikeLike
I do this for my own podcast too and it’s been working great. Mine is a bit more complicated since my cohost and I co-interview a third party, all remotely, over Skype, so I need to throw Soundflower in there somewhere, but it’s essentially the same setup and has worked like a charm so far.
LikeLike
This has been known in radio circles for decades as a “double-ender,” and is a great way to get high-quality audio from all your participants. You can also do it for more than two people if you have a conference call or use Skype or iChat, or as a backup if you’re recording a Skype or iChat call anyway.
Paul Garay and I used to record our joint segments for Inside Home Recording this way, until we started using the Source-Connect plugin for Pro Tools (now also a VST for other software) from Source Elements in the U.K. — it’s pricey ($400) and requires an iLok dongle, but it’s also what many professional studios and voiceover actors use. It uses a high-quality, low-latency AAC codec and sounds great, very nearly as good as sending uncompressed AIFF or WAV files back and forth.
Source-Connect speeds up editing and post-production considerably because one party can record the other over the Web directly into a Pro Tools track and then start work as soon as the conversation is done. No need to send files hither and yon. So if you’re more sensitive to time than price, doing that is a good option. Cheaper variations include iChat/GarageBand and Skype/Audio Hijack, but I find the sound quality not as good that way.
Doug Kaye has posted some good stuff on his blog and at the Conversations Network on getting some of these setups going as well.
LikeLike
Why not just use iChat and Garageband? Garageband will automatically detect an audio chat and drop each person on an individual track.
LikeLike
This is so awesome Matt. Thank you very, very much!
LikeLike
Wait, but… if you’re both recording locally, then how come you need the Skype Recorder app at all? Or does Skype not like to share the audio device?
LikeLike
While I understand that Garageband crashing would drive you away from it, you don’t have to have Gband running on both ends to record an iChat conference on separate tracks. Garageband running on one machine will capture all of the audio to individual tracks.
LikeLike
Right Tom, but that defeats the whole purpose of this how-to, which is how to avoid the network transport artifacts from your audio recordings. If I recorded both sides of the conversation on my own computer, it would sound like the second sample above, not like the third.
LikeLike
If you don’t want to shell out for Call Recorder for Skype, Gizmo Project has recording functionality built-in.
LikeLike
Sounds great. I’ve seen video podcasts using a similar approach.
The main problem with this system in audio-only podcasts is that if the talent has have similar voices (John Gruber’s “The Talk Show” comes to mind), it can get pretty hard to distinguish between speakers.
Individual filtering and/or stereo placement might solve that, of course.
LikeLike
Great info!
LikeLike
This is an excellent trick. I teach how to use podcasts in education, and I’ll be using this trick the next time I give a workshop. Thanks.
LikeLike
’tis called a simul-rec, in radio circles…
LikeLike
Adam, you could do it that way… but then how do you know what the other person is saying?
In theory, you can do your interview/etc via chat, and then both separately “record your lines” but it won’t sound natural, plus you would have to do tons of audio editing to make all the sounds line up.
For instance, what if you say something, it takes me 17 seconds to answer, then you say something again.
But if you’re recording your lines without hearing me, you might only wait 14 seconds between your question, and responding to me.
Now you have to shift all of your audio past that point over by 3 seconds… and you have to do this every time you don’t sync.
Using skype lets you have the conversation in real-time, naturally. The double recording avoids any static or poor sound that skype would cause.
LikeLike
Matt, thanks for sharing. I’m going to try the same thing with Call Recorder or WireTap on the remote end and Übercaster on my end, since I should be able to align the remote captured tracks there (and I may well have more than 2 people on the call).
This won’t help the crappy sound from our callers on TalkShoe, but that’s unavoidable. 😦
–Mike Rose, tuaw.com
LikeLike
couldn’t each party just record the mic input to their pc? that way you can combine them both afterwards by just mixing the tracks together?
LikeLike
Great ideas, Matt, and very timely, as a friend and I are in the planning stages of a podcast. I have mics and mixers in my home office, but it’s not always convenient for him to come over here. This is a great (!) alternative.
LikeLike
Is this meant as a joke? If each person uses a microphone and records directly into Garageband while on the phone this would work. I don’t understand Palmer’s overcomplicated thinking… there’s no “syncing”… once both ends are continuously recording they are on the same timeline…
LikeLike
Bob,
Garageband takes up a lot of processor power. This could potentially get in the way of Skype’s performance, and vice versa. Call Recorder is much “lighter” in its processor power.
On a more powerful Mac, using Garageband would work, but the premise of this piece is to record a podcast cheaply. Having to get a more powerful computer defeats that idea.
LikeLike
I’ve done this a couple times myself – using Skylook to record (it has a 24 hour rental option at 99c per time as well as 30 day, annual rates).
I use a PC so I do the editing using Audacity.
LikeLike
Has anyone tried to use two Macs whilst recording with Audio Hijacker? I wonder if the same thing can be achieved that way.
LikeLike
Jurie, I wrote this with two hosts in two locations in mind. So both hosts would have Call Recorder and Skype, yes. When I do interviews, the interviewer usually uses a phone or skype and I record as normal on my side only.
This guide is primarily written to talk about a two-person podcast recorded between two people far away from each other, to make them sound like they are in the same room.
LikeLike
Just for my understanding: this requires the interviewee to have Skype, CallRecorder and a Mac, right? So the interviewee pays $15 for being interviewed? Or it costs the interviewer $15 to do the interview, and you need to get the license over somehow?
LikeLike
Matt, do you use Skype voicemail for listener mail? I know you used to use that Netzero service, but that appears to have have shut down in February.
LikeLike
Matt this is great stuff…I am a podcaster novice but I want to get started doing interviews with one other person in the same room. Can I just use 2 usb headsets and plug into the same computer to produce an audio file? Or do I need some kind of cable splitter or mixer?
LikeLike
This is going to be really helpful, thanks. I have a problem, though. While I’ll be one party in the conversation, and I use a Mac, the other party will be using a PC. Is there a PC-friendly Call Recorder out there? I’ve used Audacity before but it may be too complicated for the other user…any advice?
LikeLike
Hi Matt, Great tips. my problem is that the person that i am going to be doing a podcast with is not tech savy at all and has a PC. How do i do like example two if he is on Google talk and i am using iChat to talk with him.
LikeLike
Matt-have you figured out a way to add a “caller” into the mix efficiently on Skype? You could use your original method and layer your 2nd host and the caller (which are on the same “line” in Skype). This would lead to a thick outout of the caller and a masked version of the 2nd host. Both could be compressed in post to probably sound ok. The layer of the caller might even make that part sound “thicker” and stronger since it is a true Skype phone call. Any thoughts?
LikeLike
Has anyone tried this two host model doing basic video chat? I would imagine that in theory it could work the same way…
LikeLike
This is great stuff. Like Sara, I have dozens of people around the world I want to record but they’re usually on PC with a USB headset. Is there any easy PC application to record just the Mic end of their conversation? I’ve Googled it death and can’t seem to find something as elegantly simple as Call Recorder.
LikeLike
Nice, concise, though detailed explanation of the double-ender method. I have heard Leo Laporte mention this and I understand this is how “You Look Nice Today” is done as well.
Great intro and primer!
LikeLike