Thursday, 22 October 2009


I'll skip the usual apologies for the lack of posts... take it as read from now on that I'd love to be posting more often but both work and my personal life are pretty busy these days so unfortunately this site gets left behind. My hope is that I'll upload more frequent, short posts rather than the rambling ones from earlier this year.

As some of you might know, I produce a podcast called PhotoLegal which neatly combines my interest in photography, technology, audio work and legal/civil liberties. It's hosted by the fantastic team of Darren Hector (professional wildlife photographer), James Barisic (our resident lawyer), Kate Day (Communities Editor at the Telegraph) and Phill Price (amateur photographer and IT guru).

The show managed to rack up 10,000 downloads over it's first series and we're now halfway through our second - this time around we're broadcasting live while we record so it's a bit more fraught for the hosts and my recording setups but it gives us a level of interactivity that other podcasts don't have.


The Early Days
I got involved early in the first series after hearing Phill sound like he was recording from a cave - since then we've had our fair share of problems ranging from noisy recordings, poor mic placement, uneven levels and more recently some bad audio dropout over Skype which actually caused us to abandon the recording, although we continued for the benefit of our live listeners.
Darren and Phill took inspiration from the audio setup posted by PhotoNetcast when they originally started recording the show - it basically consists of a Skype conference call routed through a mixer and an additional USB line-in to allow the recording to capture both the local vocals and the remote guests. It actually works very well, especially at the start when we were able to have Darren and Phill in the same room with just James and perhaps a guest coming in via Skype.

I streamlined this a little by removing the unnecessary digital-analog conversions that using the mixer required - my trick is to use something called a Virtual Audio Cable which is a small driver which mimics an output on your computer and allows you to route it where-ever you like. Using the VAC and it's accompanying Audio Repeater software I'm able to output our Skype call along with any local vocals to a recording package, to our live streaming software and back to my speakers/headphones as well!

It's a trick I used to be able to do with one of my previous soundcards (M-Audio Delta66) but since my desktop computer broke I've not been able to use it's features and have been reliant on a laptop instead. I've got a sneaking suspicion that Mac users might be able to do this natively but I'm a lifelong PC user so I have to resort to hacks like this.

If you're interested, I record using Cubase but this is overkill for a simple live recording - it was very useful when I had multiple local vocals to mix into the call on the early setup but once it was all being done using the virtual cables this was so much of a problem. Something like Audacity is a good, free, alternative. I know Darren favours Soundbooth on his Mac.
Our live broadcast is handled by LiveStream which is designed for video streaming - we block off the camera inputs but use it's video abilities to put up graphics and notes.
The problems came when we tried to avoid us all driving around to physically be together during the recordings - Darren, Phill, James and a guest would dial into a call hosted by me with my recording gear ready to go. Add another guest and the recruitment of Kate as co-host and suddenly we had 7 people on the call and discovered that Skype struggles a little!
General interference, noise, pops, varying volume levels, random audio dropout and finally completely dropping users from the call meant that our chosen method of communication was starting to look problematic.

The Solution
Our solution came from Rob Wright who suggested an alternative VoIP system called Ventrilo - it requires a host server to be running somewhere but when we trialled it for our latest recording the results were spectacular.

Call quality is a huge improvement over Skype which I suspect is because we have a handful of people using a single server rather than Skype's thousands of users. It's also got some nice features such as Push-To-Talk which avoids unwanted background noise and can be set to an automatic trigger or a keyboard button. Each user also has more control over the level of their audio input and output.

The big benefit for me is it's ability to record a call itself - I can set it running and not worry about it, plus it even records the local user unlike Skype's output. It records in a proprietary format but this isn't the problem it might seem since it actually logs and records each user's individual audio, allowing me to export each on separately into .wav files and then back into Cubase for some proper multi-track editing. I tend to apply noise reduction as required, EQ each channel and normalise the levels, then it's run through compressor and gate VST plugins before mixing down to a single .wav file again and another export to MP3.

It's really great to have a noise-free VoIP service since it makes the editing job much easier and the recording and export abilities mean mixing the final track is a dream compared to our earlier attempts.

Hopefully this has been useful to some of you - I do realise it's quite a narrow field of interest but there's not a lot of comprehensive information about this sort of setup as a lot of podcasts only have one host and possibly a guest; we're a bit unusual with our 6 or 7-way discussions!

The image at the start of this post is from Colleen AF Venable's Flickr photostream under the Creative Commons Share Alike licence.