I’ve been reminded that I’ve been remiss about writing about my experiences getting the Fusion Patrol podcast up and running. I’d like to wait until I’m sure that everything is working right, but that day may never come!
We’ve now put out 16 episodes spanning three continents and there have been some hard won lessons learned. I think, in fact, that I’ve forgotten much of the frustration, but I think I can remember what I need to impart.
The fundamental flaw we’ve had has been in sound quality. I suppose that goes without saying for an audio-only podcast.
Perhaps it will be easier to explain if I describe our setup. Ben and I conduct the podcast over Skype from our respective homes. Ben uses a PC, which plays little or no further part in this story. I do the recording on my end on my Macbook Pro.
Skype does not natively record phone calls and I purchased a piece of software called Call Recorder, which plugs into Skype and can record all Skype audio. Unfortunately, there is no distinction between callers. Skype turns both ends of the call into one audio stream.
In our first episode, Ben’s audio was of poor quality and mine sounded good. This was really surprising because, while in the conversation, it sounded great to us. It wasn’t until we listened to the playback that the problems were obvious. A little (read: a lot) post-production magic improved things, but it was clear Ben needed a better microphone/headset. I gave him mine and I bought a new, USB headset. I chose USB because… I don’t really know why. It seemed like the thing to do at the time.
For the next few podcasts, my audio sounded clean, but was extremely low in volume, week after week I was forced to spend hours attempting to balance the audio, manually raising and lowering the volume of each sentence spoken. Weirder still, my voice seemed to get quieter and quieter as each podcast progressed.
I decided I didn’t like my USB headset and preferred the quality of the original mic that I’d given Ben, so I bought another one, which has separate input/output plugs. My audio seemed to get even quieter.
It turns out the new microphone (and possibly the original that I gave Ben) is not powered, but a Macbook Pro doesn’t have a microphone in, it only has line in, which won’t work with an unpowered source. What was happening, it seems, was that I was actually recording though the built-in microphone on the Macbook Pro. As the podcast recording session would progress, I would begin to fidget and get further and further away from the Macbook, thinking that I had the microphone suspended in front of my face. The Macbook’s built-in noise-cancelling microphone is really good for telephony chatting but not what you need when podcasting.
When I discovered the problem I had a dilemma. I was in Taiwan and my options for buying and testing equipment were limited. I discovered, though trial and error, that the Macbook Pro’s headphone jack is actually the same audio in and out jack that is on an iPhone, and as I had my iPhone headset, I used that for one podcast. The microphone quality wasn’t as good, but it was definitely not going through the laptop’s microphone.
Having finally solved my problems, I was devastated to find that still my audio faded away as the podcast progressed. What could be causing it?!
About this time I also discovered a piece of software called “The Levelator” which automatically does something similar to what I was doing manually by raising and lowering the volume. It does a remarkably good job, but not good enough. I found myself using the Levelator and then tweaking the final results, which did make for a lot less work for me.
Perhaps belatedly, I began searching for other possibilities as to what the problem and I came across a possibility. It seems that the Mac version of Skype differs from the PC version of Skype inasmuch as the Mac version always has auto-level control enabled and does not have user-accessible option for turning it off. With that knowledge in hand I conducted some tests and, sure enough, if I would make a slightly louder sound, the Mac’s audio input control would be lowered by Skype. Oddly enough, the so-called auto-leviling never raised the volume when it got too low. In the essence, it simply kept lowering the audio and never raising it back up. By the end of a podcast, my audio input was down to nothing.
With a little checking online, there are some hidden control XML files for Mac Skype that you can edit, adding in some commands that will disable the level control.
Thinking I’d solved that, my next task was to try to find a plug adaptor that would take the in/out plugs from my headphone and neatly combine them into a standard iPhone style jack. Despite the millions of iPhones out there, I couldn’t find such a device, save for one company that makes equipment for court stenographers. Being such a vertical, captive market, their prices were not realistic to my budget. As a fallback, I purchased an iMic, which is a audio in/out to USB converter, effectively turning my new headset into a USB headset. (You might think I’d just go back to the USB headset, but it really does have a poor microphone.)
With no auto-leveling and proper microphone placement, things seemed perfect, just in time for our (extended) discussion about Doctor Who with our guest Simon from the UK, who also joined us via Skype over our first ever three-way Skype conversation.
That was quite a day. We’d expected to talk around 90 minutes but ended up talking for over 4 hours. It wasn’t easy scheduling a time when everyone was available and just 30 minutes before the appointed time, my neighbors began mowing their lawn. I was worried that the lawnmower would interfere with the audio. It did, but only indirectly. They were done long before the podcast started but the gunk tossed up in the air began to slowly, almost imperceptibly to me, clog up my sinuses. My breathing became somewhat more labored than normal. This is fairly common for me and I rarely notice it. Little did I know how horrible it would be!
With my audio now solid and uninterrupted, I should have realized Sod’s Law would take full effect and Ben and Simon’s audio would be very low. Once again, for the four-part podcast that it turned out to be, I was forced to raise and lower the volume on a sentence by sentence basis. It would have worked, except… for my sinuses which continued to labor away while they were talking. When I’d raise their audio, so my belabored breathing got raised up as well.
Once or twice during the long podcast, I no doubt bumped my headset, lowering the microphone closer to by breathing, making it worse in some parts of podcast.
It’s all horribly, horribly embarrassing.
While our most recent Podcast #16 wasn’t perfect, I did not have to extensively tweak the audio, although I did let the Levelator work on it, which seems a great improvement to me.
What’s really needed is multi-track recording, but that seemed impossible with Skype. It seemed the only solution would be to find a way to bring my audio and Skype’s audio into a mixer before being recorded. Ideally we’d bring each participant into a different input to the mixer and then I’d be able to adjust each audio source independently in GarageBand in post production. That’s an extensive hardware solution that I was just not willing to invest in.
But just yesterday I read an article about a piece of software called Wiretap Anywhere from Ambrosia Software that sounds like it may be the very solution I’ve been looking for.
Wiretap Anywhere allows you to create virtual multi-channel audio devices from any number of hardware and software sources. For example, you could put the audio input from the computer’s microphone, iTunes, a USB microphone and Skype output as four stereo inputs into one virtual device. That virtual device can then be used as an 8 channel input device into GarageBand (or any other input capable destination). GarageBand can only use one input device, but it can record each of the channels independently (either individually as mono sources or as stereo pairs). In simple terms, it allows GarageBand to do multi-channel recording from practically anything that generates sound on your computer.
They have a free 30-day demo and after a bit of trial and error, I was able to successfully record my end of a Skype conversation and the remote end as separate tracks: exactly what I’ve been wanting. The proof will be in using it in actual podcast conditions, but it looks like this might be the solution to one of our major difficulties.
At $129 it’s a bit on the expensive side for someone like me that’s just podcasting with no major ambitions towards world domination, but on the other hand, if it keeps my blood pressure down, it’s worth its weight in gold.
I’ll be testing the demo on the next podcast. Keep your fingers crossed.