- Get a microphone that has a low noise floor, directional pickup area and a decent wide band. It needn't cost more than $80 unless you have a penchant for gold plated lead over black plastic.
- Place this microphone far enough away so that no one can hear your tongue move or your lips touch. IE if you must use a headset, use a good one. And do consider a pop-guard if you're the animated speaker type (good for you). Do not produce a screencast like this one. Lip smacking happens 6 times in 30 seconds, and there's evidence the lead in was edited, so why not the rest?
- If you have pauses in your audio, and who doesn't, go back in and silence those pauses with a waveform editor like Audacity. It helps if you used a Compressor/Gate/Limiter (specifically the gate part) to stop minimal noises that aren't speech from registering. This is like what a guitarist would do so that you won't hear his fingers sliding to a new position between notes.
- Practice your screencast or podcast. Go ahead and see if you can put some out of casting-area prompts up, like with http://www.cueprompter.com/ perhaps?
- Do not do a full-screen cast since no one will watch it at full-screen. Exception: slide presentations that are 36pt font large or bigger.
- Once again, edit your audio to remove inhales, exhales, ums and aah. The truck outside your window probably means a retake of that section, and your neighbor carrying his bike up the stairs means you should have yelled "CUT CUT CUT" and then actually cut it. A practiced person can recognize a yelled "CUT CUT CUT's" waveform shape, and eventually a "CONTINUE IN 4 3 2 ..." waveform.
- If you have more than one speaker try to level the audio for both. Give Levelator a whirl; or work it out in your editor. If you're recording Skype consider getting the mono for your voice and the other party separately. The Tapur plug-in does this in the left vs right channels; it's both easy to level and easy to merge down; also you can de-emphasize when one person tried to talk over the other and failed to take over the conversational thread.
- DO NOT brand your production with more than a 1 second intro or outro. Remember that many web video players will display the first or last frames statically for a long time, but might obscure the center or corners. Even Microsoft shows some modesty by keeping the start button's windows logo relatively small. But it's always there... don't try ever-present marketing; don't overlay a "channel logo" style corner; you have real-estate outside the video frame for that.
- DO transcribe, out of band, URLs, product names, and anything that needs to be spelled correctly for the listener to use it.
- Summarize, out of band, all the content. Don't just archive a teaser, assume that someone will remember a great part of a video or podcast and will want to find only that part again two years from now, without recalling the surrounding main-topics; especially without knowing the date or current events of the time.
- If you must credit, do it like a reporter by introducing names and organizations as they're discussed. Otherwise for production credits put them where it's safe to stop the playback without missing anything else... namely the very absolute end.
- Assume nobody has time for what you just recorded. Aggressively edit it down to the minimal required information. 1 hour maximum; 2 minutes optimally.
BONUS: 5555. Never use a saxophone clip in any part of the production. Similarly, please strongly analyze the purpose of and need for any other reed instrument you wish to include and put it up front in your summary that it is present. E.G. "Warning, a clarinet plays 3 notes before the credits at time-code: 04:55.008. Thank you for tolerating my poor taste and timbral faux pas for 0.4s". I'm only kind of joking here. This guy plays the sax and still notes it's overuse to general agreement.
This has been my attempt to not blog about blogging and similarly not (screen/pod)cast about (screen/pod)casting. How's my driving?
No comments:
Post a Comment