Broadcast Audio levels for television.

this is strange. I used to use PPMs for audio never going over -6db. Since I moved into advertizing the level for an ad in the UK is -10db. I just spoke to a seasoned pro in the sound world who has never outputted at -10db - never and provided me with a soundbed peaking at -6db. This was a different sound sweet to who we regularly use - they provide -10db soundbeds and always have done. I don't get this. Can anybody ellaborate, exstrapolate and confere? The US is differnet I'm sure. Cheers Phil UK

Well, it is a well known fact that commercials are WAY LOUDER than the actual programming....so I think that -6 is too low. I think they might be closer to 0, or even up to +3!!

Kidding...but they are MUCH louder.


www.shanerosseditor.com

Listen to THE EDIT BAY Podcast on iTunes
[itunes.apple.com]

Cheers Shane, yep they are louder but -10db is what commercials in the UK go out as. It is the same for say concert films, music promos - it is baffling that a TV veteran should tell me otherwise and be so adamant. The place I do my mastering archives alot of BBC and Channel Four television and they do -10db.

If you're on PPMs, we're not talking dB.

"6" on a BBC PPM is Permitted Maximum Level (actually 6.3, but that's hard to see). Very shortly thereafter, the transmission system goes into overload, clips, and other nasty things.

Digital Full Scale meters are very different. They do use dB with zero at the top. Music and voice peaks occur 6 to 8 dB over the reference level which in Europe is -18, America is -20 and in DV -12. This is one reason going between DV and Broadcast is so darn much fun. The assumptions of what "natural level" is are very different.

<<<t is a well known fact that commercials are WAY LOUDER than the actual programming>>>

Yes, but they don't measure that way. That would be illegal. They do it with careful tailoring of the sound energy density and frequency response.

Koz

Isn't one of the problems that there is no real standard for monitoring sound (well... what about decibels). FCP's audio level monitor needs to be translated into what your output medium needs.

For an upcoming project to be aired on PBS, we're told the rules are -12 (FCP scale) for all audio, -27 for music under voices.

TVC's are subjectively much louder because the mix is highly compressed. An easy way to evaluate that is to look at the waveforms. You'll see that their tracks (as are radio broadcasts) are full up to the max at every frequency level thus making them seem louder although still within broadcasters legal limits.

Clay

Phil, I think you're mixing up PPM values with dBs. PPM4 is -18dB, PPM5 is -14dB and PPM6 is -10dB.

Either way, the bottom line is that commercial levels should be quite a bit lower than the peak level on regular programmes because the audio is always heavily compressed. I know that a few years ago the BBC ruled that promos would peak at PPM5 precisely because viewers complained about the relative sound levels.

In the ITV commercial guidelines it actually says you should aim for peaking at PPM4! I have no idea whether anyone takes any notice or how strict they are about enforcing it.

Martin Baker
[www.digital-heaven.co.uk]
Unique plug-ins and tools for Apple Pro Apps

<<<sn't one of the problems that there is no real standard for monitoring sound (well... what about decibels).>>>

This is me raising my bloody hand in praise of the "Virtually Useless" VU meter. Proper ANSI C16.5 VU meters respond to roughly the loudness of a show. I have test sounds here somewhere that sound roughly the same to your ear and read the same on a VU meter, but drive a digital meter bats. Unfortunately, digital is all we have left.

It may be determined by careful reading of the BBC specifications that the 1-7 PPM is considered a part of the sound quality control. Quality control is done on a high quality sound system (speakers) in a quiet room. Without that, you're only measuring one small characteristic of the performance.

There is no "FCP Sound Meters." All the digital meters read the same. We import DigiBeta tapes all the time and sure enough, the sound is 8dB low in Final Cut just like it is on the tape.

Like I've written before, to lay off a standard DV show to DigiBeta, you can either reduce the whole timeline 8dB, or you can throw the tape machine calibration off and do it that way. A straight transfer from a standard DV show will not give you a "legal " tape.

[www.kozco.com]

A less desirable way to do it is lay it off DV standard, -12 tone and all, and warn the people receiving it what you did.

Many people found that they can put a -12 tone on the tape and then peak the show at -12. That's not "legal." That's a broadcast show with a DV tone at the top.

Koz

Quote
Many people found that they can put a -12 tone on the tape and then peak the show at -12. That's not "legal." That's a broadcast show with a DV tone at the top.

Koz, could you explain that further? If I were to prep for broadcast I thought I would have to make sure the show didn't peak above a defined level, as indicated by my bars and tone at the head of the tape. Are you saying this is not the case? Or is this a consideration of going DV>DigiBeta?

JK

_______________________________________
SCQT! Self-contained QuickTime ? pass it on!

<<<I would have to make sure the show didn't peak above a defined level, as indicated by my bars and tone at the head of the tape.>>>

Of course not. That would be too easy.

I should warn you to keep doing whatever it is you're doing if the client doesn't complain--too loudly.

Digital systems are a lot more robust than the older analog systems are and if you do slightly oddball things, nobody will care all that much. The problem is interchangability.

If you do everything exactly right--like in the demo videos--then you have 100% interchange between all the formats whether analog or digital. No exceptions.


To make analog services (which don't use peak meters) look right, the show has to peak 6 to 8 dB over the reference tone. This is no matter which digital machine or timeline you start from. If you start out life in US Digital, reference tone is -20 and the show peaks at -14 to -12.

If you start out life in DV. The reference tone (from the Effects Tab) is -12 and the show is expected to peak at -6 to -4. Just like the video demo shows.

People who never experience analog television figured out that if you put DV tone on the timeline (-12) but broadcast level show (also around -12), most people would accept it.

My guess is the Injest people at the station/network just ignore the tone. That's what I'd be doing. During the tone playback, close your eyes and say "LALALALALALA" really loud and only start to pay attention when the actual show comes up.


If you were to select tone at any television station Master Control, you won't get -12. You'll get -20 (in the US).

So to bring it around. Don't change what you're doing if the client is happy, but to be surgically correct, you in particular should be reducing the tone level (and only the tone) to -20. Our DigiBeta machines have -20 marked right on the meter face (see demo video). That's where they're expecting the tone to be.

Koz

>My guess is the Injest people at the station/network just ignore the tone.

well they darn well shouldn't be! they should be using the tome to calibrate the ingest. if the reference level is -12 then they should adjust the playback/rec level to bring it down to -20, and then the following show levels, if its been mastered correctly, should line up correctly.

and its been said before over and over again but its worth repeating ... folks working in broadcast should always check and adhere to the technical specs for program delivery as provided by the specific network/broadcaster.

<<<well they darn well shouldn't be! they should be using the tome to calibrate the ingest. if the reference level is -12 then they should adjust the playback/rec level to bring it down to -20, and then the following show levels, if its been mastered correctly, should line up correctly. >>>

But a lot of them aren't mastered correctly. See: John K above. I believe I've heard a number of editors who do that -12 thing posting here. In those cases, if Injest pays strict attention to the tone, the air show will be 8dB too low-noticeably quiet.


I wouldn't want to be Injest. They get three different shows; full-on DV on DigiBeta, that's easy drag everything down 8dB and go with it, full-on broadcast, do nothing, and the oddball split where the tone is DV, but goodness knows where the show is......

Judging by my off-air digital viewing/listening, most Injest services are unsupervised.

Koz

Hey Koz, I was refering to FCP audio meters, not PPM's sorry if there was any confusion. Cheers Phil UK.

Me again, people do question the regulation of broadcast rules. The BBC rules the roost but have come under scrutiny for their practices. What got me was having a sincere sound proffessional with years of experience saying he was unaware of content being trafficked at a peak of -10db...

But it depends what he means by -10dB. It's possible that audio peaking at PPM6 will go above -10dB on a digital meter and it would be perfectly legal for UK broadcast.

Martin Baker
[www.digital-heaven.co.uk]
Unique plug-ins and tools for Apple Pro Apps

Phil,
What i don't get.
You're into commercials. You got decent meters. You've made broadcastmasters before that probably were not rejected.
Now you are going to change a prooven workflow based on one guys statement?

Bouke

One thing i never understood was why audio has to peak 10 dBs below 0? Isn't it better to fully utilize the bitrate? Especially since we're in post production and accidental clipping shouldn't happen if the audio was calibrated properly and it's unlike a 'live' scenario...

bouke Wrote:
-------------------------------------------------------
> Phil,
> What i don't get.
> You're into commercials. You got decent meters.
> You've made broadcastmasters before that probably
> were not rejected.
> Now you are going to change a prooven workflow
> based on one guys statement?
>
> Bouke

Personally, I won't just be happy with masters that weren't rejected. I'd like to be able to get my commercials as "hot" as I can onto TV screens without sacrificing the necessary dynamic range required to carry the show. Ignoring the tone ? Not the best of ideas, but the tone isn't as useful as it was in the analog days- where you'll need the pitch to tell you if the tape is spinning at the proper speed.
Different broadcasters will expect different peak levels and sound floor levels. You'll need to get a copy of the technical publication from the station you're sending the tapes to. Although where i am, the last time i followed it (on paper it specified dialogue peaks at -10dbs), it got rejected, and after some "louder", "softer" prompting from the guys from the station who didn't seem to understand the digital full scale measurements and were still using the VU (virtually useless) meter to measure peaks, we worked our way back to my initial -6dB levels. Basically, if you have a team of professionals at your station, you can afford to peak where they want you to. And I'll go in as loud as I'm allowed to go, so that my commercials carry the impact when played on air next to the other commercials.

<<<I'd like to be able to get my commercials as "hot" as I can onto TV screens without sacrificing the necessary dynamic range required to carry the show>>>

Right, then. There in one go you have the problem. Step one of Advertising. "Get your attention." (the other two are "create a need" and "satisfy the need."

"Hey! Are you bleeding out your ears? Apply my cream!"

What better way to get someone's attention than yelling at them?

The -20 digital (-18 everywhere else) was designed to have more than enough theatrical headroom for any performance; guns shooting, canons booming. The problem comes when the whole :30 performance lives up where the artillery is supposed to be.

But don't worry, the process is self limiting. Ever wonder why older CDs sound softer than newer ones? People gave up on the Red Book -17 reference level and just record the performance as loud as possible--which turns out to be peaking at 0. I have one Trance Nation CD which never gets below -2. The audio waveform looks like I painted it with a magic marker--solid block of blue.

So do whatever you want that's accepted by the client. It won't create any serious damage like overmodulating the old FM sound carrier used to do.

But that process also promotes TiVo-ing past the commercials. Loud spots were an irritant in analog, in surround digital they terrify the dog and knock over the wine glass.

Koz

For the noise wars, truth is that there isn't too much of a way out of this- I love my pink floyd cds, and if they were mixed and mastered today, they would sound horrible (like a wimpy piece of loud noise). Then again, for commercial standards, without artillery in your commercials, if it was just straight up dialogue with background music, and you were to observe the headroom, and the guy broadcasting your commercial doesn't bring the volume back up, what happens is this: people watching stuff on TV will suddenly hear a slight drop in volume. The effect is that they probably just won't pay that much attention (as compared to a slightly louder commercial, ceteris paribus).

But what I usually do, is to keep the average sound floor up as loud as possible without sacrificing essential dynamic ranges as far as possible. IE. If the only spike I have is 1 shout that doesn't need to be that loud, I would compress that peak and bring up the average volume.

BUT... what i would advocate is that the loudest peak is recorded with maximum bitrate- as close to 0 as possible to maintain signal integrity as much as possible. Headroom is there to prevent distortion, but if it's not used, it's a waste of digital bits. Record an orchestra as close to 0dBs without clipping, then record the same orchestra keeping an excessive headroom of let's say -30dBs. Play both back, and see which is clearer. One would use maximum bitrate, the other is recorded effectively with less than 12 bits, since the rest of the bits are used only to "preserve" headroom, otherwise, it's not used at all.

On top of that, most homes do not have state of the art sound system that's dolby certified, and their speakers aren't capable of reproducing any form of reality, much less excessive dynamic ranges. But yea, on one hand you have the loudness wars, that results in a loud wimpy sound with almost no dynamic range... On the other hand, I constantly deal with recordists so afraid of clipping that they some even record with a 30 dB headroom (although they do not seem to understand that they can clip on their analog mixers without peaking above 0 on tape)