audio calibration via NTSC -20db tones...
I'm a video editor dropped into a new environment where I feel out of place regarding audio conventions here.
I had a conversation with the younger post production supervisor about audio mixing and he's got an entire staff of editors to mix to something like -6 on our Digibeta and SX decks.
They still print tones to -20 digital which is what I'm familiar with.
But the mix is kinda wishy washy. He says no one ever mixes to the little notches on decks anymore (!) because there is so much headroom on the digital decks.
By his definition, bars are now used to indicate the middle range of content, and it's OK to go as high as you think you need to go as long as it doesn't go 'near' the top.
I'm sure you can imagine my reaction, but is there any validity to this audio strategy?
In my old skool days (yes, we're talking analog which Sony has kindly pandered to me by adding those little ticks on the digital VU meters) you set bars to hit those ticks (aka 0db analogue or -20db digital). Then when you do your mix you try to keep the meters from going TOO MUCH above those points of reference. Some guys even go so far as to say you NEVER go over bars but I've never quite done that.
Lately I've been getting a lot of mixes that use an incredible amount of compression (ie it comes onto the timeline as a solid black waveform!) which seems to be the trend. (Sidebar read this excellent article "why no one listens to music anymore":
...but even when I use these heavily compressed and expanded sources I still try to keep the overall levels BELOW old skool 0 or new skool -20.
So what's your feeling on this (who's right here) and how are you dealing with the new audio frontier... or is this young kipper off his rocker! (c:
COW Leader Discreet Edit*ors
I'm also "old skool", but my understanding of "standards" (there seem to be some variations, but at least you have to have a standard to vary from), is as follows:
For analog, tone at 0 db, program material averages around 0 db with the peaks going slightly above (Maybe +3db or so)
For digital, tone at -20 db, dialog/narration hovers around -14db with peaks going up to -10 or -9 or so.
If you're keeping everything below the level of the tone, you would be safe, but your levels would end up being somewhat low when used by somebody else on their system (broadcast, duplication, etc.).
Now this is for video. In the music buisiness, peaking at as close to 0dbU as possible seems to be the norm. That's where you get the solid black box waveform you're talking about, because they have compressed the crap (and many times the life) out of the music! But that's a whole 'nother subject...
Hope this helps. Have fun!
Your explanation is excellent thanks. It confirms what I've been seeing and has encouraged me to be more generous with my peaks. I've always been afraid of going too far above 0db (analogue) and/or -20db (digital) and I'm finding the extra headroom tempting and something I don't fully understand.
Now if someone could explain a little of this to our Podcast producing friends that would be GREAT! How many times my ears have been blown out.. but that's another post!
I think hes talking PANTS
I only hope i never have the misfortune of hiring him as a mixer.
I think Id find his theory riddled with holes within the first minute of a job interview.
Are you in the UK or not.
The rules are in the UK
1k tone @ -18dbu=PPM 4. Audio peaks at +8db above reference so PPM 6 which is -10 on a digi beta meter (not that youd EVER attempt to use the front of a machine as the meter.
Its all about creating a relationshi between the tones and the audio that follows it. In theory the actual tone to audio levels are un important as long as the tape tech log clearly showed what the relationship actually is, but in practice the rules are there to be followed and in UK broadcasting the numbers above are GOD.
In the uSA i belive things are lined up a little differently so im not expert on their statndards, but I think -20 is what they use.
My understanding is that reference tone was meant to be a way to allow another engineer in a different facility to set a level when playing your tape and not have to listen to the entire program tracking the peaks to make sure nothing clips (distorts). With analog tape proper practice was to never exceed +3 over 0VU.
I guess the same theory holds today. If the tone on a tape you receive is at -20 or -14 and you setup your input to play those tones at -20 or -14, no peak should go to a level that would clip the audio (ie. exceed 0 on the digital scale).
If you are mixing a film with dynamic audio content the dialog might be at -30 to -20 with explosions going close to clip level (0 digital). If the program is a commercial, it's more likely it will be highly compressed to raise its average level to the max so the commercial sounds "loud." I would expect that the peaks of this kind of program would be near 0 digital.
When I mix audio for a DVD for an industrial video, I typically use mediun compression and let peaks go to -6 on the digital scale. This way, the DVD will have plenty of punch if it needs to play in a loud envirnment. As long as you don't clip the signal, it's all relative. If your audio on a DVD has peaks going to -20 then audio will have to be cranked up at playback. But may still be too "soft" in certain situations.
Dual 2.5 G5 4 gigs RAM OS 10.4.3 QT7.0.3
Dual Cinema 23 Radeon 9800
FCP Studio 5 (FCP5.0.4, DVDSP4.0.2, Comp2.0.1, STP1.0.2)
Huge U-320R 1TB Raid 3 firmware ENG15.BIN
ATTO UL4D driver 3.50
AJA IO driver 2.1 firmware v23-28
SonicStudio HD DAW, Yamaha DM1000, Genelec Monitors