How it is done

Video:

Oldest videos were recorded on a Sony Digital-8 camcorder. The more recent videos were recorded with a Sony mini DV camcorder model DCR-HC32. This is a nice small camcorder with decent low-light performance. Not great though. I really should upgrade to a 3CCD camcorder -- maybe when a decent AVCD-HD model is released.

Audio: I used to use the built-in mic on the camcorder. It is decent for loud events (kirtans) but not very good for classes.

Audio:

• Marantz PMD-660 -- This is a great portable flash based digital recorder. I record at 48khz, 16bit, WAV onto a 4GB card. This gives me 6hrs of recording in stereo. I can extend that by recording in MP3, but since I always postprocess the audio I prefer raw WAV.
• For classes I use two mics. The Audio-Technica Pro70 is a lavalier style mic that I get as close to the speaker as possible. In the temple this might be on the mic stand. At programs, I may even ask the speaker to allow me to clip it on their kurta/shirt. This mic gets very clear audio with very little ambient sound (say from an air conditioner). I also use a boundary mic to attempt to pick up the audience and other ambient sound. There are quite a few out there, I use one that I got a freebie when I purchased the Marantz.
• For Bhajans I use a Stereo Sony ECM-MS957. I just got this, so don't have much experience with it. So far, it seems to do well in Kirtans and Bhajans. I used it today for recording the Jhulan Yatra kirtan and that sounded great.

Software:

• Scenalyzer is great for firewire video capture and tape indexing.
• Sony Sound Forge 8.0 to convert from Stereo to Mono tracks (when using two mics for lectures). This is probably overkill for just this application. But since I have and use it for other projects, why not. A free alternative would be to use Audacity instead.
• Sony Vegas Video for video editing and audio mixing. You could use other packages like Adobe Premiere. I found Vegas Video to be the easiest to use and quicker overall to use.
• Sonalksis Compressor, EQ, and Gate for Audio processing. Vegas Video comes with built-in equivalents, but I found these plugins to be much easier and more powerful to work with.
  

Process:

The tricky and time-consuming part. There is a lot of post-processing that needs to be done in order to get everything working together. This is because I'm not just using the audio from the camcorder which is already in sync with the video.

First, (for lectures) the audio track must be converted from one stereo track to two mono tracks. One for each mic. This is so I can mute the boundary mic for the entire video except for when the audience needs to be heard.

Pull in all video and audio assets into Vegas Video.

Group the two mic tracks together so that when you move or trim them, they are done together. Then sync up the audio track to the video track. It is best to find a spot where there is a loud nosie in both tracks. Then, whole zoomed out, roughly line up the two tracks by sliding one or the other. Once it is close, zoom in all the way so you can see the waveforms very clearly. Select a small section around the loud noise and playback in loop mode. You should be able to visually line up the two and then check that they are lined up perfectly by playing it back. You'll want to check the video preview as well to be sure that looks synced up.

At this point, you can mute the camcorder audio track. It won't be used anymore. Using the volume envelope, also mute the boundary mike for the whole track. Then find sections of the track where the audience is speaking (questions, etc) and for that section increase the volume using the envelope tool. Don't bother fine tuning the volume, just put it back to 0db.

Finally, the audio quality needs to be cleaned up. No matter how good the mics are, you'll have ambient noise that intrudes during quiet portions of the lecture. The audio will also vary quite a bit in loudness depending on how well the speaker stays on center to the mic.

On the master volume control, click on the Audio Plugin button and make sure the plugins are in the order: EQ, Compressor, Gate. This is different than the default, and it works great for me. Do all the work on a portion of the audio that has both the speaker and a significant silent part. You want to minimize the noise in the silent part without impacting the normal sound of the speaker's voice. In the EQ, you want to chop off the low frequencies and the high frequencies. They will be removed when converting to MP3 anyway, but you might as well do it here as well. Many ambient noises reside in the low and high frequencies as well. If you still hear a strong sound, use the spectral analyzer mode to see if there is a strong sound at a single frequency range. If so, you can adjust the EQ to remove that portion of the audio. Once the noise is minimized it is time to compress (or normalize) the volume. I usually just use the default settings for the compressor and just adjust the limiter to -5db. The last step is to remove as much noise from the silent portions. The noise gate does this. Start out with the "Vocal de-breathe" preset. Then find portions of the audio where the volume is very low but you don't want silent. This will be where the speaker is mumbling or when someone in the audience is asking a question from way in back. Adjust the threshold so that these portions are not reduced to 0 volume while still having "silent" but noisy sections reduced.

Ok, what about noise reduction? I sometimes use it. This works by sampling the noise in a "silent" section of the audio and then subtracting that noise from the entire track. I've found that to remove enough noise to make a difference results in the audio sounding "metalic." Usually, it is just isn't worth the trouble. I might use it over a particularly difficult section though.

Thats it. I then render the video and audio and publish.