A FAT WAVEFORM or thin is in?

B

bwest

Well-known member
I’m a little confused about my audio waveform on my NLE, and this could be the dumbest question anybody’s heard here yet! I’m a little afraid to ask, but here it goes! My friend and I have the same version of Vegas for our NLE’s. My friend’s camera is a Panasonic 1/6th 3ccd camera…a little more consumer than mine. I think it’s called the GS series and I think it’s one of the newer models since Panasonic started these smaller 3 ccd cameras. When he captures his video to the NLE his audio’s waveform looks very fat, meaning there is very little white/blank space left in the audio track. Since we both lack studio monitors and are playing our videos over little computer speakers I really can’t tell which one’s audio is better than the others. I know it’s dumb, but speakers are to be upgraded soon He uses his camera’s onboard mic or the Rode Videomic.
Now, I just purchased a 302 and calibrated it to my dvx100a…tonight! I’m using the Oktava on channel one and a Lectrosonics 400 series with the m150 mic. on the other. Upon my first test everything looked good and I did a capture to see if I actually stopped clipping my work. My NLE’s audiometers are indicating I’m just a hair under clipping [0.6] in my peaks when I was screaming at the mic. Trust me I’m happy…it appears I’ve calibrated correctly, that .06 is the limiters kicking in, and I’m recording as hot as I possibly can! Attenuating the master output did the trick and I now have headroom.
Here’s the dumb part…my waveform only gets fat when I’m making lots of vocal noise, screaming, singing etc. When I pause, between words or screams, my waveform turns into a little hair of a line representing momentary silence…I suppose. I use to think the fatter waveform [my friends] with little or no variation in its size [just big] was representative of what a waveform is suppose to look like…but now, I’m thinking, his fat waveform is representative of a much poorer signal to noise ratio than mine. Am I correct? [told you this sounded dumb…we both need speakers]
 
Your friend's waveform is representative of automatic gain control. The camera is constantly adjusting the gain so that whatever sound is out there reaches a specific level. If you listen, you'll hear background noise rising to fill the silence, quiet dialog boosted to match screamed dialog, etc. The DVX waveform is representative of the natural variation of sounds, and has what we like to call "dynamic range".. that is, quiet parts, loud parts, everything in between. When noone's talking, the background sound doesn't rise get louder, etc.

This undefeatable AGC is one of the things that really separates professional from consumer camcorders...
 
If the sound on you friend's Panny isn't horribly distorted (meaning the input level was up way too loud) then it sounds like his camera as either automatic gain control (AGC) turned on or a built-in compressor. As the sound gets quieter, the camera boosts the audio, resulting in the "fatter" waveform. Note that the more consumer you get, cameras have squat for audio controls, meaning you may not be able to disable the AGC.

To answer your question, silence should always be a flat (thin) line. Dialog, unless you're zoomed way out, should have visible peaks. What you're describing on your newly calibrated system (congrats by the way!) sounds correct.

In music, the final mix is often compressed so that hangs out in the upper -6dB or so (very little white space). The only time I can think that would happen with video is for commercials, which is why they always seem louder than the show. But you don't really want your recordings looking like that; you loose all your dynamics.

edit: gah, ullanta beat me to the punch.
 
You might also be talking about the difference in 'appearance' of mic level waveforms (coming from your friends camera) versus the line-level waveforms coming from your pro set-up. If I recall correctly, mic level waveforms will always look more BIG (mountains), while line-level waveforms look smaller (hills), even when the meter displays the same overall db levels. Bottom line= worry about your meters more than the waveforms.

Others may have more technical advice for you.

-Vic
 
vicrattlehead said:
You might also be talking about the difference in 'appearance' of mic level waveforms (coming from your friends camera) versus the line-level waveforms coming from your pro set-up. If I recall correctly, mic level waveforms will always look more BIG (mountains), while line-level waveforms look smaller (hills), even when the meter displays the same overall db levels. Bottom line= worry about your meters more than the waveforms.

-Vic
Click to expand...

No offense, Vic, but - no!
 
What are you saying 'no' to? My recollection that line level waveforms aren't as big as mic level, or when I suggested worrying more about meters? I could be wrong about the waveform thing, but I seem to remeber reading that somewhere.
 
Mic level is indeed, by definition, much lower than line level. Mic level likely wouldn't show up at all on a typical waveform scale. However, the consumer cam is pre-amping that level; and the "professional" cams also accept mic or line level, and amplify them if necessary to get to line level. Once sound is digitally recorded, the waveform has no relation to mic or line level - it's a series of digital quantities. The readings on the meters (if we ignore mixing/panning/etc) are instantaneous snapshots of the waveform - or to look at it the other way, the waveform is just a record of the meter levels over time. The meter does not do different things for different types of signals.

Therefore, you're right that the magnitude of mic-level signals is much smaller than that of line-level signals, but that's not the issue here. But one rarely if ever sees a mic-level waveform... meters are almost always at a line-level stage of the signal chain... and the digital sound recorded by consumer or pro-cameras are on the same scale.

If you were to scale the mic-level signal, it would be exactly the same shape as the line-level signal - it wouldn't change the standard deviation from the average level. The original question is comparing a signal with large dynamic range (lots of deviation from average) to a signal with very limited dynamic range (little deviation from average); and this is something that does not vary with signal level.


Also, you are a bit reversed - if you line them up on the same scale, mic-level waveforms are tiny (grassland) compared to line-level waveforms (mountains)...


Does that make sense?
 
Last edited:
Yep, I get it...everything gets boosted to line level at somepoint in the chain, no matter what the chain is! Mixer, Mic Level, consumer or pro cam!
Ullanta, What is 'signal to noise'...is it?
1. subject's signal vs his background/ ambient signals...cars, planes, buses
or
2. electronic noise from wires, camera, etc. picked up in the chain vs your subjects signal...may your subject be birds, music, or dialog?
 
Either - basically, "Signal" is what you want, while "Noise" is what you don't want.

So it's context dependent... one might talk about the "signal to noise ratio" of a State of The Union address, or of data from a scientific experiment.

But in talking about specs of a piece of audio equipment, it usually has to do with noise generated by that particular piece of equipment.

Make sense?
 
You must log in or register to reply here.
Back
Top