Comments welcome on improving my Audio (example given)

[IanB]

As you can see in my signature, I am 'trying' to get this audio thing correct. The right equipment, with the right technique. I couldn't have hung blankets or sound protection in this space, but I did make an attempt to put in music, ambient noise, and background noise, as well as applied NR, and Compression on the Dialogue tracks.

What are you all thinking here? Should I have had the hyper hooked up, rather than the lav? Should I have positioned the lav somewhere closer to the mouth? Was my post processing 'to much' or poorly executed? Individual files are available for those who wish to get them.


http://vimeo.com/26727744

 

[BRH]

I think it's fine, except you out VO doesn't match well, and the audio is out of sync. If you use too much processing the audio should be re-rendered or printed, bounced....... and brought in again so you can re-sync it.

 

[IanB]

I think it's fine, except you out VO doesn't match well, and the audio is out of sync. If you use too much processing the audio should be re-rendered or printed, bounced....... and brought in again so you can re-sync it.

BRH, thanks for the comments, I haven't heard of outputting my audio and re-syncing it after the editing process is complete. I am just using a 'clap' (one or two claps of my hands to be exact) to sync my external audio and my raw video footage. Then I begin editing when everything is synced to the 'claps' at the beginning of the clip.

Also the sync'd clip was only 3 minutes long at most. Should I really need to worry about drift or un-syncing over that period of time? Camera is 48k audio and SD702 is 48k with 24 bit, I am using a GH13, and the SD702, so I would be surprised if my audio can't hold the frame correctly for more than the 3 minutes.

I also checked my output/rendor and project settings. Sony Vegas is set to 48kz 16 bit with 'stretch quality' = good. (not sure what that means). Render settings are the same. So the only difference I see is going from 24 bit to 16 bit.

 

[DrDave]

I really liked the video. I think most people would not notice some tiny volume issues and so on, but if you like they are easy to balance. Apparent loudness is a minor issue, as opposed to measured loudness. If you skip around the video, you can hear that in the beginning, for example, the voice is louder. However, most ppl don't skip around when viewing, the transitions are more important. To fix this, I use splits, which is to say I cut the audio in a few places, add a cross fade and tweak the gain. I do this instead of normalizing, just by ear.
I like the warm voice sound you get. It is inviting without being too tube-like, and the sibilants are well controlled. I like that you can hear some birds and so on. I might bring the music down a hair more in a few places or bring the voice up if you are going to match gain with the beginning. Around a minute (1:06) into the vid there is a strong change in the basic voice sound, you can use a room simulator to even that out like the free "SIR" plug-in. You can also add some EQ and gain, I'm assuming that the sound change is that you are off axis to the boom mic. So boost the volume a bit and tweak the highs to compensate for high freq loss as you move away from the mic. Be careful not to add more sibilance at this point.
Freeway noise--it is in a few places of course. Is this added noise or is it present in one of the tracks? If it is added, I would just cut back a bit in the exterior scene. If it is present in the source material, there are a couple of approaches for that, but a little freeway noise is not a bad thing, as you need some ambient sound. In the roll up to the interior scene, there is not much going on in the bass, either in the voice or the guitar track, so I would roll off the bass starting around 120 Hz and see if that cleans up the freeway sound. Sounds like you have some roll off already, so maybe that is as far as you can go with that.Although freeway sound covers a wide spectrum, there is a lot of energy buried in the low frequencies, and roll off even up to 160 or 180 Hz will get rid of the apparent loudness. However, the roll off has to be smooth, and not too much, or you will have an artificial background. I'm assuming it is all on separate tracks so you can even EQ the "Freeway" track. Even if it isn't, you can safely EQ the master channel to 120 as you don't have any key bass frequencies. More heavy NR will affect the sound, you don't really need it but you can do a noise print for the freeway sound if you like. A step up from that is Renovator or Cedar.
I myself would add a little sparkle to the guitar track, the high end sounds a little covered, but this is really maybe being too picky. A touch of EQ combined with a track specific very short reverb (1.2-1.4 sec tail, 10-12 percent wet/dry) will do the trick, or you can even experiment with dynamic compression where you split off the compression into bands.
You do a real nice thing with the into and outro using the same text, and keeping the bird sound into the credits. I would tweak the gain so the those two audio "scenes" match in basic sound. I divide my audio into scenes, then check each scene against each other. Easy to do with splits in the audio track, and most of the time you can do this right in the Video editor.
For the transition to the outro VO, I would use a longer cross fade, and cut in half a second later if possible.
I see you mention that you used NR, you might see if adjusting the NR to EQ in the interior gives you a little more realism. The interior is nice and quiet, so it doesn't need a lot of NR.
As BRH mentioned, there are a few sync issues. If you have already matched waveforms with the Cam audio, the cam audio could have drifted. If not, line up the audio wavs. If you don't have a clap or slate for sync, look for a place, like where you clunk the bottle on the table, to line everything up.

Hope that isn't TMI.