Audio Meter says ok, but clipping is still happening?

[Arthur Kay]

Ok, so what you mean is gain staging mismatch. Now I get it, thanks!

Still, I don't subscribe to that theory. For correct gain staging, you would need to do a calibration of your system first, no? One amplifier's 10 o' clock might be too little, while on another one it might already cause distortion...

So most of you guys use consumer stereo amps? Sorry, if this sounds snobby, really not my intention, but I come from the studio world, so I am not familiar with what the average videographer uses.

Cheers
Arthur

[Alessandro Machi]

Originally Posted by Arthur Kay

Ok, so what you mean is gain staging mismatch. Now I get it, thanks!

Still, I don't subscribe to that theory. For correct gain staging, you would need to do a calibration of your system first, no? One amplifier's 10 o' clock might be too little, while on another one it might already cause distortion...

So most of you guys use consumer stereo amps? Sorry, if this sounds snobby, really not my intention, but I come from the studio world, so I am not familiar with what the average videographer uses.

Cheers
Arthur

I'm refering to audio clipping when one is doing camera acquisition work. In my studio my amp is actually set to 2 out of 10 for mono, and 3 out of 10 for stereo, so clearly the 10:30 to 3:00 rule does not apply in an edit room environment for purposes of speaker playback settings.

If the CAMERA default position for the audio dial is 12:00, default meaning the middle point, I've found the 10:30-3:00 rule works very well. The rule can possibly be stretched to 10:00 rather than 10:30. Once the sweep spot of 10:30-3:00 is missed, either loud sounds can distort (when the setting is over 3:00), or lower sounds can become inaudible (when the setting is lower than 10:00)

[Bill Davis]

I think there's a major language problem happening here.

"Clipping" is a term used in ANALOG audio recording. There really is no similar phenomenon in digital audio recording. In analog, when you pushed a signal over the "optimal range" and into the red zone on a VU meter, the result was something like compression - up to the point that "s" sounds got splashy and things weren't pleasing to hear.

In the digital realm, if you exceed the input level to the point where you run out of bit depth to map the audio into, your audio CATESTROPHICALLY fails. To the point where you can't understand what someone's saying.

So, if you have clean audio at any stage after the recording, there's no actual audio "clipping" happening.

If the "client" is hearing something that sounds like old analog clipping, then it's probably either over-compression making the sibilent sounds splashy - or perhaps they have blown tweeters on their monitors - or it's one of other things that can "sound" like old style "clipping" - but again, if you're recording digitally, its certainly not "CLIPPING" in any traditional sense.

(just trying to keep the language straight in this analog to digital transition world!)

Hope this helps.

[Arthur Kay]

Bill, as far as I know, clipping is also used in the digital realm, i.e. if the amplitude of the waveform tries to go over 0 dBfs and thus the waveform is 'clipped' or chopped off, since indeed the amplitude cannot go over 0 dBfs. So you get a trapezoidal waveform instead of the normal sinusoidal waveform, hence the term clipped.

BTW, bit depth has nothing to do with the amplitude or strength of your signal, but with the resolution of the sampling process.

Just thought I should add this.

Cheers
Arthur

[Douglas Spotted Eagle]

"Clipping" is, and will continue to be used in the digital realm. Whether you're talking about truncating, brickwalling, squaring, digital clipping, whatever....once the signal reaches 111111111111111111111 (sixteen ones), no greater level may be achieved and it's cut off, clipped, truncated, smashed, whatever term you want to use may be.
Many, if not most, still refer to this as "clipping" or truncating.

I don't know that it matters what word is used, so long as we know when we're discussing digital vs analog.
Sony, Apple, Digidesign, Avid, Cakewalk, Echo, M-Audio, etc all refer to the term "clipping" in their digital product user guides.

However, it is true that the nomenclature from the analog to dig world can get pretty weird when crossing terms.

[Steve House]

Originally Posted by Arthur Kay

Ok, so what you mean is gain staging mismatch. Now I get it, thanks!

Still, I don't subscribe to that theory. For correct gain staging, you would need to do a calibration of your system first, no? One amplifier's 10 o' clock might be too little, while on another one it might already cause distortion...

So most of you guys use consumer stereo amps? Sorry, if this sounds snobby, really not my intention, but I come from the studio world, so I am not familiar with what the average videographer uses.

Cheers
Arthur

The "10-3" rule is more of a guideline than a way of setting levels, as least as I interpret it. If you need to turn the recording level control down under about 10 o'clock to prevent the meter from going too high you're probably sending too hot a signal and are in danger of clipping at the input on peaks while if you have to turn it up past 3 to get an adequate recording level on the meters you're presenting too low a signal at the inputs and are likely to suffer a reduced S/N ratio. Like Goldilocks and her porridge, we want to get the signal level presented to the camera's audio inputs to be just right, high enough for good S/N but low enough to avoid the risk of clipping.

I don't know how many use consumer amps but I doubt if many professional videographers do, certainly not many who are shooting for broadcast or theatrical distribution. Very similar considerations apply in a professional video studio as do in a music studio and while there are differences in the specs of some of the tools, such as the optimal monitor crossover points for dialog editing versus for music mastering, for example, the level of quality required for professional results in just as demanding.

[Arthur Kay]

OK, now I am really confused, feel like on the tower of babylon *grin*

So you are talking about the input levels (gain) and others are talking about volume levels (suggesting to me the output levels on the amp towards the speakers) Hence my reference to the consumer speakers.

If we're talking input levels, then I gotcha.

As a guideline, when recording digital, I mostly record around -10dBfs to -6 dBfs at 24bit bit depth and around -8 dBfs to -3 dBfs at 16 bit bit depth. So my 16 bit signals will be hotter, since I am more prone to loose detail at 16 bit with poorer s/n ratio (-98 dB) than at 24 bit (-144 dB).

Still, when recording audio on location, I make sure I got a good mixer with clean preamps that have a good headroom and as accurate metering as possible.
Like I stated before, I come from the audio world and video is new to me, so I joined to learn more about my new Canon XH A1, and while at it, I would like to share some of the knowledge I gathered from being 20 years professional in the music recording and producing business.

Cheers and have a smashing weekend y'all (hey, I got a lot of texan friends, and they taught me well *grin*)


Arthur

[Jeff Mack]

You didn't say what type of wireless. Folks also mentioning limiting. I got burned once using a wireless transmitter to send a signal to my receiver bodypack which I had plugged ito my Z-1. I watched the signal on the transmitter off the board and everything looked good. The problem I had was I had the auto limiter selected to the on position and whenever the signal got too hot, the signal got limited. On playback it sounded like a lot of pops when these spot happened.

If I would have done what Greg suggests byt listening to the signal with phones, I would have heard the problem and could have turned off the limiter.

I actually blamed it on the sound guys board limiter. When I got home, I had three body packs with no batteries in them and the 4th one that did (the one I used) had the limiter on.

Better to have it off and manage your signal.

Jeff

[Michael Nistler]

Geez, another practical Q&A thread devolving into theory - how can I resist jumping in? (apparently not)

Anyway, to begin with, I was a bit confused with the original post:

"We cranked the volume on the camera all the way. That made it loud, but the meters on the camera and dvrack did not show any clipping."

Something isn't quite right here, is it? Aside from the camera's amplifier certainly not operating in the linear "Class A" range, I didn't think adjusting the camera amp would affect the DV Rack audio via the firewire port. My laptop is in for repair this week so I can't test this out. Anyway, I seriously doubt the JVC camera mic amp will perform well running maximum gain. We don't know about Dennis' transceivers and cabling so it's unclear if there's a possible impedence mismatch, bad cable/connector, etc. Who knows, it might even be a mic problem - assumedly the listener who complained of distortion was listening to BOTH channels, not knowing the left-right used different mics (twice the opportunity for cockpit and equipment errors). Or perhaps Dennis only used on channel in Premiere and converted the best channel back to stereo in post - we can only guess. And who knows what the DVD sounds like in player/TV.

I also am suspicious along the lines mentioned by Kit and some others. For instance, on my Sennheiser G2s if I'm not careful I could easily overmodulate the transmitter by forgetting to adjust the sensitivity (say going from a dynamic to condensor mic) - the receiver setting could be be spot-on and the camera, DV Rack, MP3/WAV recorders all are happy with their VU meters seemingly perfect (garbage in - garbage out).

Anyway, aside from the troublesome JVC amp cranked all the way up, I'd put on a good set of headphones and carefully listen to both the right and left channels of the PP2 timeline sequence. And if that sounds peach, burn another DVD and crank it up.

Finally, be extra careful with the mix-n-match scenarios. Using a lav on one channel and shotgun on the other is great insurance, but if you're not certain both channels are pure, either could come up and bite you if something goes amiss (and in fact, may have here). Your client won't be happy if you didn't listen carefully to minute 33 and the wireless receiver gets a noise burst, the mic cable is loose, etc. In audio engineering, the more open mics the more opportunities for gotcha's. Gad, these theoretic discussions are fun! Okay, back to cutting In/Out points...

Good luck, Michael

PS - Jeff Mack also gets my thumbs up with the possibility of a limiter compensation problem (ugg I hope that wasn't it)

[Steve House]

Originally Posted by Arthur Kay

....

If we're talking input levels, then I gotcha.

As a guideline, when recording digital, I mostly record around -10dBfs to -6 dBfs at 24bit bit depth and around -8 dBfs to -3 dBfs at 16 bit bit depth. So my 16 bit signals will be hotter, since I am more prone to loose detail at 16 bit with poorer s/n ratio (-98 dB) than at 24 bit (-144 dB).

Still, when recording audio on location, I make sure I got a good mixer with clean preamps that have a good headroom and as accurate metering as possible.
Like I stated before, I come from the audio world and video is new to me, so I joined to learn more about my new Canon XH A1, and while at it, I would like to share some of the knowledge I gathered from being 20 years professional in the music recording and producing business.

...

What sort of material are you recording and on what sort of equipment when you see those meter readings, also what kind of metering are you using? I ask because they seem a bit hot for in-camera recording levels for DV. Standard practice for gain-staging is to adjust the camera's recording levels so that with your mixer's output set to unity gain, a 0VU reference tone from the mixer reads -20dBFS (I see you're in Europe and the EBU calls for -18dBFS) on the camera's meters. The mixer output level is left to unity gain, the camera's level controls aren't touched after that, and level adjustments during the shot are made with the mixer's input faders. With the gain staging setup that way, source material adjusted to average around 0VU on the mixer's meters will show average levels around -20dBFS with peaks hitting around -10 to -8 dBFS on the camera's meters. But when you do your final master for release, nothing should ever peak over -10dBFS.

[Douglas Spotted Eagle]

Originally Posted by Steve House

But when you do your final master for release, nothing should ever peak over -10dBFS.

Steve, is that what you really meant to say? Peaks never beyond -10dBFS? That level would be rejected by most of the bureaus that we service.

[Bill Davis]

OK, I'll happily acknowledge that the term is used in all sorts of manuals - and is pretty popular. It's just that I still have trouble feeling comfortable with it relative to my years in radio and analog recording.

The point I was trying to make (ham handedly and obviously not well!) was that clipped audio in an analog relm left you with a usable signal. Heck, lots of analog circuit designers touted "soft clipping" as a desirable feature to maximize S/N in their machines.

Re-purposed into the digital relm, people are getting the term confused, as evidenced by the OP.

"Clipped" audio (or whatever other term you want to employ) in the digital relm is pretty easy so spot as it's totally unintelligible.

The fact that the OP described his original source as "clipped" - but was able to play it back and hear it correctly off the camera master - indicated to me that it could never have actually been "clipped" in the first place - and led me to think he was confusing "clipping" in the analog sense with it in the digital sense.

Of course, maybe I just read things wrong. (Wouldn't be the first time, nor the last!)

But the conversion from analog to digital terminology seems to be messing up a lot of folks like the OP who are trying their best to learn both the language and the techniques of good recording.

For what it's worth.

[Douglas Spotted Eagle]

I know what you meant, Bill...didn't mean to sound like I was picking on you. I think we all have our pet-peeve words or terms. "Over modulate" when it's not transmission related is one of mine, and then there are the odd-ball descriptives.
Then again...there are those here that wouldn't know a 7.5 ips from a 15ips either... :-) For many, the analog world is an alien concept

[Alessandro Machi]

Originally Posted by Steve House

The "10-3" rule is more of a guideline than a way of setting levels, as least as I interpret it. If you need to turn the recording level control down under about 10 o'clock to prevent the meter from going too high you're probably sending too hot a signal and are in danger of clipping at the input on peaks while if you have to turn it up past 3 to get an adequate recording level on the meters you're presenting too low a signal at the inputs and are likely to suffer a reduced S/N ratio. Like Goldilocks and her porridge, we want to get the signal level presented to the camera's audio inputs to be just right, high enough for good S/N but low enough to avoid the risk of clipping.

Thanks for explaining better than I said it.

[Steve House]

Originally Posted by Douglas Spotted Eagle

Steve, is that what you really meant to say? Peaks never beyond -10dBFS? That level would be rejected by most of the bureaus that we service.

Average program material at -20dBFS, absolute 'never exceed' peak of -10 is what I was given to understand to be the normal specs for both broadcast and theatrical release, although theatrical might push the max a bit higher while keeping the same average. Are you saying that's too high or too low?