How I Use Open Source and Standards to Create Podcasts

Summary

Tools used: 2 Kubuntu 7.10 PC's, Gizmo VOIP client, Gizmo VOIP conference feature, Icemat Audio headset and usb card, Audacity audio editor, Podsafe Music from http://music.podshow.com,

Process:

1. Use only SIP compliant VOIP clients (iChat, Gizmo, Ekiga) so that everyone can use the Gizmo VOIP conference call number.

2. Use two Gizmo clients on your end. One to participate in the call and one just for recording. The PC that is recording with Gizmo should have its speakers and mic turned off. Otherwise strange echoing or feedback will occur.

3. Everyone call in using the full phone number that you get from Gizmo (It should look like 1-222-xxx-xxxx). Using other methods doesn't seem as reliable.

4. Record the call. Everyone will hear a message that says that the call is being recorded. Make sure that at the end of the call everyone is off mute and you record about 10 seconds of “silence”. This is for later on when you use the noise filter feature in Audacity.

5. Hold your podcast session

6. When completed click on the record button again to stop recording. Then hangup. Do NOT forget to first stop recording! Immediately back up the recording file. Be default it saved it to the root of my home directory. In fact, I would think of cutting it to a CD just in case.

7. Find your intro/bumper music on the PodShow web site. Don't use that horrid big studio crap! Use music from artists that actually need your help that refuse to let their music get trapped by a big studio. 

8. Use Audacity to fix the sound. Use the Leveller, Compressor and Noise Removal effects in Audacity to smooth out the levels from different people and remove any hum or static that is there (remember that 10 seconds of silence? Use that as your sample for the Noise Removal effect). Don't be worried about totally messing up your podcast with Audacity. It has multi-level undo and you backed everything up right?

9. When you are done, export to MP3, Ogg, etc...

10. Make sure all the MP3 meta data is filled out. Artist, Album, Track, etc... and upload to your site for inclusion in your RSS feed. You are done!

The Details:

I don't claim to be some expert on podcasting. In fact, I'm a total newbie with only 2 podcasts under my belt (Well three really but I don't want to even discuss the test podcast for the Lords of Tyr, also known as “The Lost Episode”.)

I did notice that in my quest to seek out information about podcasting on Linux, that there was a great lack of information on using VOIP for podcasting. Mostly what I found were tutorials on capturing network packets on certain VOIP ports and dumping them to .au files. Of course I tried it, and it did capture lots of packets, but unfortunately the VOIP client I was using was using something different than the tutorial I found. I think it was just encrypted so I couldn't decode it once I captured it without more hacking around.

So I set out to find a way on my own. I wanted to be able to have any of the guys in my gaming group, regardless of their OS, to call in remotely and participate. There is no sense in driving all the way to my house just to talk for 60 minutes. We have Windows, Mac and Linux users in the group. Cross platform is a must.

I also wanted our tools to be standards compliant. That means no Skype! Skype is all proprietary. Any solution we used must be compiant with the SIP protocol.

The next problem (the biggest one really) was that I needed to record it easily. At first I tried to play around with Pulse audio, but even that still seemed a little complex. Then I recalled that the Gizmo VOIP client that I had played around with before had a record button!

I did some research and sure enough! Gizmo is a standards based SIP Voice Over IP client that can record conversations. Not only that, but they have this awesome conference call feature. It let's you setup a conference call number that let's everyone dial in via VOIP or even a land line number (the land line will cost you unless you have unlimited minutes on your landline or Vonage or something).

I tested the client with my friend Brian and found two problems. The first was that Gizmo's Mac client did not want to install at all on his Mac. This is where standards save us. Once I found the right server information, Brian was able to just configure iChat to dial into the conference. I love standards!

The second problem was that my voice locally was so much clearer on the recording than anyone else on the remote side. How do I solve this one? What I really need is a way to record where my voice sounds the same as everyone else. That's when it hit me. I'll just create a second Gizmo account and dial in on that just to record everything. Now everyone will be on the same playing field.

The test podcast that we did went really great from a dicsussion and organization side of things. However, I believe that the PC that was doing the recording briefly glitched when some power management kicked in. So I lost the recording. Man was I angry.

The lesson I learned from that was to set the PC's (Kubuntu Linux I my case) power management settings to Presentation. I made sure that Presentation had every little piece of power management and screen savers turned off. I also made sure that everything else was closed except for the Gizmo client.

The first “real” podcast went much better and the recording worked properly. At the end of the session I asked everyone to take off their mute settings so we could record 10 seconds of “silence”. This is used later for the Noise Removal effect in Audacity.

The first thing I did when the podcast was over as to backup the .wav file to my server so I would have some piece of mind.

Then I started working on editing the podcast with Audacity. I used the Levelling, Compression, and Noise Removal effects in Audacity to try and make everyone's voice levels more similar. Then I edited out the beginning & ending setup talk. Once I'm relatively happy (I'm never really happy with the quality and there is probably always more tweaking to do) with the quality, I add in the music and record an intro and add music again to the end to make a nice bookend to the podcast. Then I simply export it all to an MP3 file. I make sure that all the useful MP3 tags and meta data is used as well.

Any questions, suggestions or comments? Please register a user ID and leave me a comment! I already have a bunch of updates for this and it is the first version of it. I'm sure there will be many more to go as I learn more about podcasting myself.