To record close-up speech, I tried using an Audio Technica Lavalier Condenser microphone plugged directly into the camera's audio input. First of all interference noise made it unusable. I found out that the Audio Technica ATR-3350 is very susceptible to interference from other equipment. The only other gear was the HDMI monitor screen. I plugged out the monitor and recorded again and the interference was gone. However on the subsequent recordings there were high levels of hiss. Further research showed that this is specifically a Canon EOS issue. The audio stage of the camera features an Automatic Gain Control (AGC) so during any silences the camera turns up the gain causing the hiss. I have read some interesting/smart/messy/expensive workarounds on this to fool the AGC but none of them actually do what Canon should have done, which is provide a configuration switch to disable AGC. In the end, it is easier and of course better quality, just to record the audio separately and synchronise it back together in the computer - hence the clapper board to give you a synchronisation mark. You can use the audio in the camera just to verify the synchronisation. There are a number of excellent tutorials on how to do this on YouTube.
So finally I am not really treating music, speech or general sounds distinctly different. As with any recording the mic'ing recording should be done with the best method for the job.
- Speech: Record directly into Logic Pro
- Live Music: I use a Boss BR-1600 with 2 inputs from the mixing desk and an ambient boundary mic. (I'd like more inputs but that may sacrifice monitor outputs needed by the musicians on stage)
- This leaves the issue of general mobile recording (not music - and presumably without a power socket being available). Most people sing the praises of the Zoom H4n and I will be investing in one before long. In the meantime, I can use my MacBook Pro which is more cumbersome but with an audio interface will give me the best sound quality and a battery life of a few hours.
After the filming/recording, I mix whatever stems there are first in Logic 9 to as near a finished product as I can, EQ, gain, effects if needed and then bounce down to 48K sample rate (because this is film - normal audio is 44.1K). Theoretically I may have to re-mix, re-edit post production but if this version is mastered quality, in most cases this will be the finished audio.
The audio should be then sync'd with your video using the clapper marking point or some other distinct moment that can be identified in the wave form shown on screen, and once sync'd, locked together.
Now the video (with its audio) can be cut, sliced, etc and possibly additional audio samples added in the video editing suite.
After the video edits are done, perhaps it all sounds good at this stage. If so, the job's done already. If not you can now bounce the video back down again as a WAV file; bring it back into your DAW and re-mix/re-edit using the original sound files. If you know the material, you should not find it too difficult to re-align the original source material to the video edited version. Obviously do not change any of the timing in the video's audio file. Using the Video's WAV as a timeline, any changes made to the finalised audio are then easy to re-align back into the video.
