Andrew Grumet asked in one of my comments what hardware and software I used to produce my audioblogs. Since I thought this might be of general interest and would be the kind of thing that people might like to find via search engines, I thought I’d actually write down the nuts and bolts of how I go about it.

I’ve used two different machines to do my audio recording: the first is my rather generic HP laptop (2 Ghz processor, 512M of memory, 30gb disk space), and the other is my HP multimedia PC (2.8ghz processor, 512M memory, 160gb disk). In both cases I use a cheap stereo headset with microphone such as you might use for audio conferencing via Skype. The other day I also purchased a $25 Plantronics headset: I have yet to use that for recording, so it remains to be seen whether it will result in better sound quality. Occasionally I do notice breath noises on the cheaper one, so there is room for improvement, but generally the audio quality is acceptable.

Both of my machines are currently running Windows XP (hiss!). There really is no reason why you couldn’t use Linux or FreeBSD (my personal favorite), but since ultimately the goal of these things is to interface well with the vast majority of consumer hardware, it seems to me that means Windows. I actually did some early tests using Audacity on FreeBSD, and it works quite well for recording. I also experimented with using gnupod on FreeBSD to make my own sort of “geek ipodder”, which worked very well without using any Microsoft products whatsoever, so it can be done.

The key bit of software I use is Audacity, which is a pretty nifty little sound recorder application you can download from sourceforge. It has versions for Windows, Linux, FreeBSD and even Mac OS X (so I’m told, I haven’t got a Mac). I set up a project for my audioblog that records new tracks at 22050Hz, 16 bits per sample, and has some default tags that are the same on every post I make so that I can just click a single icon and begin recording.

If you’ve been following Adam Curry’s explorations, you know that most of his problems has to do with wanting to record other “generated” sounds together with his microphone simultaneously. The reason you want to do this is so that you can incorporate other sources (iTunes, Quicktime and even Windows Media) into your audioblog just by playing them on your computer. This relies on a bit of cooperation from your soundcard. I’m not sure if all cards have this capability, but both of my machines do. The trick has to do with the mixer application. You want to examine the various recording inputs, and configure it so that you are recording not just from the microphone, but from the output of the stereo mixer.

Here is where it gets confusing: if you click on the mixer application in your taskbar, you can get access to all the playback and recording devices. First, look at the playback devices. If you are lucky, there is some kind of microphone input. If you unmute it, and crank up the levels (and perhaps turn on the Microphone Boost that you find under “Advanced”), you should hear your voice coming through the mixer when you talk into the microphone. If you set the gain too high, you might get the squeal of feedback. So far, so good. Trim the overall levels back a bit.

Now, go to the recording devices. If you are lucky, there is something called Stereo Mixer or some such. Set that to be your recording device, not the Microphone. Now, fire up Audacity, and try to record something. Try talking and playing a tune from iTunes or whatever. Both should be recorded. You may need to adjust levels to prevent clipping. I’m still experimenting with how to do that.

That’s really all there is to it. Except that my desktop PC is considerably more complicated.

You see, my desktop supports 7.1 sound and has all sorts of whacky recording and playback devices. To the PC it looks as if this single device was multiple sound cards, so Audacity needs to be configured to use the right ones for input and output (the defaults don’t work very well). To be honest, I’ve got a setting that works, but it was mostly trial and error and I’m still not sure that I understand how it works, so explaining it to you would probably be difficult. The general idea remains though: use the mixer application to configure your sound properly, and Audacity should do the right thing.

I use a couple of other features of Audacity that you might explore:

• Audacity allows you to record or input multiple tracks, rescale their amplitude, slip them in time to align them, and then mix them down to a single stereo track. This is a useful way to record intros and exits, although I usually just ride the volume controls like an amateur and record it live.
• You can select regions of the recording, play them, and delete them. I use this to delete complete asides and idiocy that is too great to share with others.
• You can do fade ins and fade outs. Very nice.
• You can pause and unpause during recording to gather your thoughts when flustered.

When I am done, I usually play back the recording to check it once again, and then export it as an MP3. Since I do 22050hz recordings, I usually do 64kbps MP3 encoding, which frankly sounds indistinguishable to me from the live encoding under virtually all circumstances. I then upload the recording using the PuTTY scp client PSCP to the /audio subdirectory on my webserver.

To get these in my weblog is actually just the crudest hack. I use WordPress as my weblog system, which has a feature called “custom fields”. These are name/value pairs that you can associate with any post. I add two fields, audiolength and audiourl, which point at the length and url for the uploaded mp3 file. I then modified the index.php and wp-rss2.php to look for these tags and generate links and enclosures.

That’s pretty much all there is to it. I’m actually going to write some better tools to help automate this process (I want to make a gadget that when i drop an audio file into a given directory, it mirrors it automatically onto my website and creates an unpublished article that references it so I can edit it later), but it isn’t enormously difficult to do audioblogs that sound fairly good with this basic level of sophistication.

If anyone has any questions, feel free to post questions or email them, and I’ll try to be helpful.