-
Mastering to DigiBeta – audio levels question
-
[Michael Gissing] “Forget VU meters and use the FCP digital metering or external meters on your digi or HDCam deck.”
I use both usually.
you get a more precise reading on the VU meters on a BetaSP deck in my case.
Ideally the peaks will hit at 10 or below on the digibeta levels
and peak up in the red without ever hitting the top most VU bar on the BetaSP deck.I find the one in final cut too small to be completely accurate.
thats just me.
-
[Chris Borjis] “Ideally the peaks will hit at 10 or below on the digibeta levels”
I presume you mean -10dbfs. Broadcast levels are speced on digital levels. A VU meter however, is designed to model analog tape saturation and will have different ballistics depending on frequency. They might be familiar and experienced users can get a feel for apparent loudness with them but they should not be a reference for broadcast levels.
It took me years to get the tech check people off VUs in broadcast stations in Australia. Eventually I made up a digi beta tape that had illegal and legal signal and showed that on a VU they called the illegal levels legal and the legal levels illegal. When compared to the digital levels they had to admit that they were wrong and now use digital metering.
-
I had to take myself out of this whole argument. Apparently I can’t understand the specs and where it says ” In program, audio should average -15dBfs and not exceed -10dBfs” what it really means is that the average levels should stay between 10-15 but there is no problem with levels actually going higher all the way up to zero. That, based on two other specs: “In bars, Test Tone audio levels should be -20dBfs” and “Headroom should achieve 20dB above normal”. I’m seeing the levels being all the way up there most of the time but that’s the of this argument.
I’d love to be wrong and not pay attention to it anymore. Let’s hope I am 🙂
-
[Michael Niemcewicz] “Apparently I can’t understand the specs and where it says ” In program, audio should average -15dBfs and not exceed -10dBfs” what it really means is that the average levels should stay between 10-15 but there is no problem with levels actually going higher all the way up to zero”
I can tell you that that is total bollocks. When it says peaks cannot exceed -10dbfs it means just that. All levels above -10dbfs are illegal. I have delivered over 800 broadcast doco sound tracks and SMPTE levels state that -10dbfs is the maximum peak signal. EBU specs (for PAL countries) have maximum peak of -9dbfs and the BBC is -10dbfs. I always peak limit a mix to -10dbfs. Any higher peaks than this and tech rejection is expected.
-
Apparently the word “peak” is not used in the specs. And my understanding of the word “average” is incorrect as well .
-
The email I received from the network:
“Here’s further clarification on the audio specs from the NOC. I hope this clears everything up!
Program audio will go above and below the reference level as program content varies. The “average” should be close to your reference level (tone). Headroom is the maximum allowable signal before clipping – it’s a level which allows for intermittent peaks which are much higher than your normal program level.
It’s not that the audio should never have anything higher than -10 but that the audio mix should be close to your reference level and the amount of dynamic range (difference between loud and quiet material) should take those levels into account. Brief excursions higher than -10 (gunshots, explosions, etc.) are normal but average program levels should be close to reference level”So they pretty much say it’s OK to go above -10. I suppose you can interpret it as a licensee to stay above?
Whatever it all means it’s out of my hands and the tapes are being shipped out. All 18 of them. -
That particular network may accept if they don’t specify a maximum peak level then, but I can assure you, any tape with peak signals above -10dbfs will be tech rejected by all international distributors. It is so simple to put a high quality peak limiter on a show and be internationally compliant.
-
[Michael Gissing] ” they should not be a reference for broadcast levels.”
In my experience they work very well.
They measure the dynamics more accurately than digital.
As long as the VU “OVER” (top led bar) doesn’t light up, the levels are perfect
and never exceeding -10. -
Is no one implementing ATSC A/85 yet? That has requirements for average perceptual loudness (-24LKFS) and allows a peak level of of -2dBTP (-2dBFS on a true-peak meter).
Simon Pegg
-
[Simon Pegg] “Is no one implementing ATSC A/85 yet?”
This is a well thought out recommendation that some broadcasters are implementing into their specs. What we do is provide content to the broadcasters based on their specs. If it’s in the spec, we’ll hit it. If not, we’ll hit whatever levels they want in their specs.
[Simon Pegg] “That has requirements for average perceptual loudness (-24LKFS) and allows a peak level of of -2dBTP (-2dBFS on a true-peak meter).”
Mixers are so far reluctant to take this on without being told specifically to do so. Reason being that they’re going to have to adjust their dynamics processing severely to hit these targets. In my cursory testing in my own facility, most mixes coming out of the rooms average around -19LKFs. That’s going to be a drastic change when it’s implemented. Until the requirement is there (by the law and subsequently the broadcaster’s spec), nobody is backing down from “loud” mixing. Some broadcasters are already implementing it as test run and those mixes who are above -24LKFs by more than 2dB are being forcefully corrected at the broadcast end, generally by TC or Dolby loudness analyzers/corrective devices.
Reply to this Discussion! Login or Sign Up