Broadcast Audio - Bars and Tones?

Jay M wrote on 9/18/2007, 4:43 PM
I know what the bars are, and I know a good tone when I hear one ;)

Can someone explain how stations and duplicators use them?

How is the level of the tone related to the level of the audio in your project?

What is the comon standard in the US for tone level?

How do bars relate to the video in your project?

For the past year we have just been added generated bars and tone at the front of each broadcast. I kind of think they should be more meaningful.

~Jay

FYI- We use FCP to edit video, audio is mixed in vegas, then a stereo mixdown goes back to FCP to print to tape.

Comments

Former user wrote on 9/18/2007, 5:17 PM
Tone and bars are vital to the broadcast of the video and audio, but only if they actually relate to the setup of the video and audio.

Tone, for broadcast analog, should be 0db analog. In FC I think this equates to -12db digital. In Avid, it is -20db digital. You need an external VU meter to be sure. This is set upon output to tape. When tone is set at 0db, the audio should not go over 0db in level. This setting allows for the small overhead you have in the analog world to help prevent distortion in the audio.

If you look at BARS on your scopes within FC or Vegas, you will see that each color lines up with a point on the Vectoscope. In addition there should be White levels at 100 and black levels at 7.5 (or 0, depending upon video format._). These colors should reflect the white and black levels of your video, as well as the color of your video. TV stations and duplication facilities will set the playback of their vtrs to similar videoscopes and VU audio meters, with the assumption that these reflect the proper settings for the video contained in the program.

Unfortunately, we are finding at our facility that many people using FC and Avids and such do not understand the relationship and we usually end up either having to reject the video because it is outside the video standards of white/black levels or color saturation or we will adjust the levels ourselves to fit within standards. This can usually have an adverse effect on the final look of the video on broadcast. Usually, picture will be a little less saturated than you expect or will just look different.

Until the broadcasters and editors are all digital, this will be an issue. Contact your stations that you normally feed tapes to to determine what they expect as far as levels.

Dave T2
GlennChan wrote on 9/18/2007, 9:00 PM
1- It helps to mention units so people have a better idea as to what EXACTLY you are talking about.

dBFS = decibels full scale. This id kind of like decibels in digital systems. 0 dBFS is always where digital clipping happens.
VU = volume as measured on a (analog) VU meter.

For professional digital formats, tone should be at -20dBFS for NTSC countries and I believe -18 for PAL countries.
Consumer formats - tone at 12dBFS.

Peaks shouldn't exceed -10dBFS ever. kdm has a better post on this...

-10dbFS max peak is typical for US broadcast (-20dbFS reference). EBU is -18dbFS ref, and -9dbFS peak. Then of course some countries have varying standards for what they will accept - some will reject any peak over the max, some will dial the mix down - some may reject if peaks are too far below the max - e.g. just what Glenn stated with meter response - most digital meters are sample accurate, so a sample peak of -10dbFS isn't the same as a -10dbFS peak for 10ms, for example.

2- Analog VU metering is different. The needle should be ballistically(sp?) weighted to respond slower to transients. Some cheesy meters don't do that. Don't peak over +4 VU.

3- Levels don't tell the whole story. You can compress the crap out of your audio to make it sound louder. (This is especially especially prevalent for audio CDs + pop music... lots of CDs peak at 0dBFS and are heavily compressed.)

The broadcaster might have their own compressor to even out the audio levels (so there isn't huge variation in perceived volume when you flip from program to a commercial). So there's an interesting game there. But that's only if you are trying to cheat the system to make your audio sound louder.

4- The way I would get -20dBFS tone from FCP or Vegas is this:

Add bars&tone onto the timeline. Adjust the level of tone -8dB to -20dBFS.

On the dbeta (or whatever) deck, make sure that the deck is not adjusting the levels. On a dbeta deck you push the pots in.

Double-check the meters, and use full/fine to see the levels in more detail if necessary.

(You can figure out how to do the same thing in Vegas.)
GlennChan wrote on 9/18/2007, 9:19 PM
As far as video levels go, just make sure that the digital levels in Vegas are correct. If your target format uses Y'CbCr video space, there is only one correct set of video levels... black at Y' = 16, Y' = 235. Setup refers to the digital->analog conversion process.

More details here:
http://glennchan.info/articles/vegas/colorspaces/colorspaces.html

And 7.5 IRE setup demystified:
http://glennchan.info/articles/technical/setup/75IREsetup.html

*If you make broadcast masters, you only need to make sure that the composite IRE values aren't too high.
Jay M wrote on 9/19/2007, 2:15 AM
thanks for the replies.

Why would you want 18-20db of headroom? that seems like a huge waste of bits.

for audio mastering I normally go from 0dbfs down. My software VU meters in Wavelab are about -8 to -15, while dbfs is frequently slamming 0. Music CDs are about about -5 to -10 on the VU

So should I master like normal, then just lower the master fader to -18?

Should the tone = the peak DBfs in my audio?

When we print to tape we are using DVCpro50 via firewire.

Once i digest the audio, I'll try to understand what you guys are telling me about video.

this is a huge help! Please keep it up :)

Even though we are a relatively low budget oporation I want to have sound and video as good as we can possibly make it. There is no excuse for our sound to be compromised. Video is harder to have have perfect lighting, and cameras.

I find it fascinating to compare Jay Leno on HD vs SD. The audio in HD has an incredibly wide dynamic range. My understanding is that the AC3 is untouched by anyone between my house and the NBC studio. SD on the other hand is heavily compressed like most broadcast television.

When I mix our show I make it sound more like the SD channel. Perhaps I should leave more dynamic range knowing that who ever touches my audio such as the duplication people, and individual stations, will likely be adding a bit more compression.

~Jay

Former user wrote on 9/19/2007, 6:15 AM
"Why would you want 18-20db of headroom? "

In Digital, you do have this headroom, but in Analog you do not. And most broadcasting is still using analog.

I believe Spot has provided more detailed information at the Vaast website on this very topic.
"Deadly Analog to Digital Levels" is the name of the article.

Dave T2
Spot|DSE wrote on 9/19/2007, 6:43 AM
Yup...there is more there.
There is also a HUGE misconception about acquisition levels vs mastering/output levels.
One fairly famous author wrote in his HDV book that you should record with -20dB as your peak. Doing so would leave you with virtually unusable audio. You want to record hot, master not.
In other words, record as hot a signal as you can safely get away with, and control it in post, rather than recording quietly and attempting to boost it in post.
ATSC has specified that -20dBFS =0dBVU. Pretty straight forward.
Printing bars n' tone from Vegas will give you a -20dBFS tone.
rs170a wrote on 9/19/2007, 6:48 AM
Spot's article, The Deadly Analog to DV Audio Level, can be found here, test tones included.

Mike