Friday, March 13, 2009

The Making of Just Plain Sense Podcasts

Some people have asked me for technical details of how the Just Plain Sense Podcasts are recorded, edited and made available, so I've decided to describe it here.

We'll look at each part of the production process in turn.


I have two different recording setups depending on the circumstances, so I'll describe each one separately.

Pocketrak 2G

For face to face interviews in the field I use a really tiny device called a "Pocketrak 2G" recorder, by Yamaha. The Pocketrak is pictured on the Just Plain Sense Podcast front page. It's about 4½ inches long by 1¼" wide and just half an inch thick – so it slips very comfortably into my handbag and can go anywhere.

Despite the tiny size, the recording quality is fantastic though. It has a tiltable stereo microphone setup on the top and records in a variety of formats. These days, after experimenting with different approaches, I record in uncompressed PCM stereo (CD quality). This means that the device can hold about two and a bit hours of recording on its' built in 2Gb Flash memory. If you record in 128Kbps MP3 then it can hold about 32 hours of material. The rechargeable battery will power the Pocketrak for about 4-5 hours and you recharge it by plugging into a USB port on a computer.

The USB feature is really neat, in fact. There's a slider on the back of the device. If you push it down then a USB connector emerges from the base and means you can plug the Pocketrak straight into a PC, where the recorded files can be accessed just like a normal pen drive.

Marantz PMD 670 plus mixer

My other recording option is for studio or events recording and features a Marantz PMD670 digital field recorder, which I usually now front-end with an Alesis six track analogue mixer and a set of professional vocal microphones. The mixer allows me to cover events with multiple sound sources, such as conferences or round table discussions. This is also what I use now for recording remote interviews with Skype.

The Marantz recorder does everything that the Pocketrak can do, but a lot more besides. For instance, it has a low and high pass filter. The filters allow me to cut out the background rumble from my household central heating. There are lots of option switches, neatly concealed under a screw plate (so they can't be accidentally altered).

My favourite feature is that the machine has a big obvious red record button, which just requires a finger-flick! A red LED indicator then confirms whether the machine is recording or paused, and levels can be set accurately on a backlit LCD display.

The Marantz records onto compact flash cards, so transferring recordings to the PC involves removing the card and slotting it into a USB Compact Flash card reader. As with the Pocketrak, I record in PCM stereo, so a 30 minute recording takes about 300Mb.

Editing and Mixing

Nowadays I produce shows with two separate software packages for editing and mixing.

Editing – Sony Sound Forge

The first stage of production is to edit the various recordings that are going to be combined into the show. Each show usually has an introduction by myself, the interview or feature piece and then an outro. Each of these is recorded on one or other of the setups described above and then I gather them all in a fresh folder on my laptop.

I like Sound Forge for editing because it has a simple no-fuss user interface. You can scale the waveform right down to individual samples to edit out 'pops' and you can change the saved file format any way you desire. For instance, I've sometimes wanted to insert 128Kb MP3 recordings into the underlying PCM master and Sound Forge allows you to do the necessary rescaling so everything fit together and sounds right.

When I first started Just Plain Sense I made the mistake of recording interviews in MP3. The problem with that approach is that the master starts out with limited quality because of the compression – and then gets worse as you inevitably save and reopen the file at each step. Nowadays I keep everything in PCM (WAV file format) until the very last step of production. That way there is no loss of quality through the process.

Mixing – Cubase AI4

The mixing stage is where the various recordings are combined with the signature tune and occasional sound effects or backgrounds to make the completed program.

Cubase wasn't a conscious choice for this. It came free with the Yamaha Pocketrak 2G and, since it does the job, I've not shopped around for anything else.

On screen, Cubase looks a little bit like Apple's Garage Band. You can set up as many tracks as you require. I normally have just one vocal track and one music track. You then just drag and drop the sound files where you want them to go. Fades are handled by dragging with the mouse to shape an automation line for each track's fader. Putting the components together like this and auditioning the fades is the fun bit of production in my view.

Once you are happy with the overall mix of the program, you drag an envelope round the whole sequence that you want to output and generate a new file. Again this is output as another WAV file in my case (because Cubase can't write MP3's itself).

Tagging and Upload

Having produced the final mix down of the programme, I pull it back into Sony Sound Forge once more just to add the IP3 tags for the MP3 file. E.g. The title of the episode, the names of the participants, a description and the copyright details. I then use Sound Forge to save the tagged file as a 128Kb MP3 file.

With the MP3 file produced, it only remains to upload the file to the Podcast Platform. I use but there are others. When a 30 minute programme is converted into 128Kb MP3 format it goes from being about 300Mb in size to just under 30Mb. Even so, the upload will generally take 12-15 minutes on a UK ADSL connection. I use that time to (a) listen to the completed episode end to end and (b) write the description that appears on the Podcast's web page.

I can't emphasise how valuable it is to listen to the whole recording in its final form before publishing. Although you'll probably be sick of hearing your production by the time you've been through the editing and mixing stages, it is always possible that you may have missed something. I once witnessed a radio programme going out with a string of expletives in the middle that an editor had missed!

How long does it all take?

The time for the whole production process depends on many factors, such as the number of pauses or verbal grunts the interviewee makes in the original recording. I like to present my interviewees in as good a light as possible, so it is not unknown for me to spend 1-2 hours carefully removing unwanted sections and matching the pace of the interviewee's speech so that (hopefully) you can't spot the joins).

However, don't forget that there is also time to be spent often in preparation before an interview too. Depending on the subject, I may spend an hour or two researching the interviewee and deciding what kind of questions to ask.

Mixing usually takes 20-30 minutes – depending on the number of sound elements and how the joins needs to be made. If one speech segment follows another, for instance, then I may decide to go and look for a short segment of music to bridge the join. On other occasions I've spent hours on the internet looking for a suitable sound effect to download for inclusion.

The tagging and uploading takes about fifteen minutes and the final publishing step (adding in the text that appears on the web page and the iTunes info box) adds another 10-15.

Overall, therefore, a typical 30 minute episode can require an hour or two in preparation and spend three hours in production.

No comments: