Sony Vegas 13 CALM Act How Do I…?

[Graham Bernard]

[Norm Kaiser] “First question: Is this roughly true?”

When I do Broadcast, I send my work to a 3rd Party to Q&A, but there are at least another 2 members who deal with Broadcast: Steve Roden and Rodger Bansemer. If they read this then they would be able to respond nicely to your request.

[Norm Kaiser] “Second question: How do I do it? Just reduce volume on each audio track and keep generating the log until I get it?”

I’ve yet to use the LOG option, sounds a neat way to ascertain just WHERE the peaks are. And yes, I “ride” the Faders use the Vol. Latch option to generate the Vol Envelop. Maybe there is a slicker way to do this?

Grazie

Video Content Creator and Potter
PC 7 64-bit 16gb * Intel® Core™i7-2600k Quad Core 3.40GHz * 2GB NVIDIA GEFORCE GTX 560 Ti
Cameras: Canon XF300 + PowerShot SX50HS Bridge

[Erik Davis]

Norm,

I just wanted to say that I have to deliver spots and programming for broadcast and I use the exact method you describe. Didn’t realize I was a po folk though. 🙂
So far, I haven’t received any complaints.

Erik

[John Rofrano]

[Graham Bernard] “but there are at least another 2 members who deal with Broadcast: Steve Roden and Rodger Bansemer. If they read this then they would be able to respond nicely to your request.”

I was Roger’s finishing editor for 3 seasons of his PBS series “Painting and Travel with Roger and Sarah Bansemer” and I handled all of the broadcast audio finishing for his show that he now uses.

The easy way is to just buy iZotope RX Loudness Control. This is a brand new product which wasn’t available when we started. If you know nothing about audio and you don’t want to know anything about audio, this is the plug-in for you. Just set it and forget it.

If you want to do it yourself the old-school way, I would start by placing Wave Hammer Surround on the master audio bus and set it to hard limit at whatever the peak value is (i.e., -10dBFS). Start with the “[Sys] Master for 16-bit” preset and adjust the Output Level on the Volume Maximizer to -10dB or whatever their peaks need to stay under. Then use a SMPTE 1KHz test tone @ -20dB to adjust the Threshold until the peak meters read -20 dB again to compensate for the limiting so that your volume is still accurate. This is the same test tone that you would use with “bars & tone” and the beeps for your slate count down if they want those.

Then I would set the Track Compressors to kick in at -24dB with a 2:1 or 3:1 ratio. You can start with the “[Sys] 1.5:1 compression starting at -24 dB” preset and adjust the Amount to 2.x and see if that hits your targets. Increase it to 2.5x or 3.0x if the audio fluctuates too much. You don’t want to set it too high because it will squash your dynamic range but broadcast audio has a limited dynamic range so don’t be afraid to use it as needed. Finally use the Loudness meters (Ctrl+Alt+8) in Vegas Pro to 13.0 to see what the averages are and adjust your volumes accordingly. You should set the Meters Loudness Scale to “Absolute (-23 LUFS)”. Remember that this is average volume not peak so you must listed to a section of audio over time to see what the average is so it’s a balancing act when adjusting the volume. You may need to use a volume envelope if the volume isn’t consistent throughout the program but you should be able to control this with the compressors.

If it sounds complicate that’s because it is. Broadcast has very strict audio tolerances. That’s why iZotope can charge $349 for their plus-in. If you don’t know what you’re doing and you can’t afford to hire a finishing editor to make your program broadcast safe then the plug-in is the next best thing.

~jr

http://www.johnrofrano.com
http://www.vasst.com

[Norm Kaiser]

>> If you want to do it yourself the old-school way, I would start by placing Wave Hammer Surround on the master audio bus and set it to hard limit at whatever the peak value is (i.e., -10dBFS). Start with the “[Sys] Master for 16-bit” preset and adjust the Output Level on the Volume Maximizer to -10dB or whatever their peaks need to stay under…

John, you’ll probably shudder when I say this, but what I’ve recently tried (since downloading and installing the patch that fixes the Wavehammer beeps) is place the Wavehammer on the master bus and the set the volume maximizer at -24db. Then I check the loudness log and find that my integrated LUFS is -24 and conclude that I’m good to go.

That’s probably extremely crude.

How does it sound?

Well I guess that’s one of the MAJOR points I’d like to discuss in this thread — the difference between audio on a computer and audio in broadcast.

I capture my audio and make sure I never clip. My waveform pattern in Vegas is very pretty — i.e., lots of pretty peaks and valleys with the peaks filling the waveform space on top and bottom without ever clipping. When I listen back with my computer’s audio slider bar set about halfway up, it sounds fantastic — rich, full, clean, very little noise, etc.

But I render to meet the CALM Act guidelines, the audio sounds so, well, quiet. I have to turn my PC’s audio all the way up for it to sound good. And then invariably I get the little dingdong chime from Facebook that I just got a new message and my eardrums get pierced.

Is this correct? I’ve read elsewhere that people report this outcome (audio sounds so quiet) and very experienced users report that it’s correct, but is it?

Does the broadcast audio system somehow amplify the sound once it gets to the viewer’s TV? Or do computers play back sound poorly in that they need a really hot signal to sound right?

[Norman Black]

[Norm Kaiser] “But I render to meet the CALM Act guidelines, the audio sounds so, well, quiet. I have to turn my PC’s audio all the way up for it to sound good. And then invariably I get the little dingdong chime from Facebook that I just got a new message and my eardrums get pierced.”

Normal PC UI sound effects are typically coded near 0dB. Full modulated volume. Movies modulate average dialog level to say around -27dB, and this allows for major dynamic range so that explosion is really frickin loud compared to the dialog. TV specs seem to limit dynamic range with peaks much lower than 0dB for whatever reason but the average dialog level is still modulated to similarly low level.

So yes, amplifier gain must be turned to a value to get average dialog to a suitable loudness. Not a problem on TV and movies, but on a PC with UI sound effects modulated to a completely different “standard”, then Yikes! the beep can be frickin load just like that explosion in the movie theater compared to dialog.

edit: I would speculate TV limits peaks because TV speaker systems can be pretty cheap and once amp gain is set for good dialog volume the speaker/amp is not capable of extreme dynamic range and the sound would crap out. This is what THX did for theaters. It guaranteed that a theater sound system was capable of a certain dynamic range and certain volume levels. No such guarantee with TV speakers.

[John Rofrano]

[Norm Kaiser] “John, you’ll probably shudder when I say this, but what I’ve recently tried (since downloading and installing the patch that fixes the Wavehammer beeps) is place the Wavehammer on the master bus and the set the volume maximizer at -24db. Then I check the loudness log and find that my integrated LUFS is -24 and conclude that I’m good to go. That’s probably extremely crude.”

Well… do you care about how your audio sounds or do you just care about passing the test? If all you want to do is pass the test than what you are doing is fine. If you care about how your audio sounds then you might want to reconsider using a brick wall limited as a loudness attenuator. (hint: brick wall limiters are design for peak attenuation not loudness)

[Norm Kaiser] “But I render to meet the CALM Act guidelines, the audio sounds so, well, quiet. I have to turn my PC’s audio all the way up for it to sound good.”

Let’s understand something… you are NOT rendering to meet the CALM Act guidelines. You are way below the guidelines. Your audio sounds low because you won’t allow anything louder than -24dBFS when actually you can have dynamics that peak at -10dBFS as long as the average loudness is -24dBFS. That’s a big difference.

Also, are your speakers correctly calibrated with a Sound Pressure Level (SPL) meter? If they are not, then you can’t trust them because you have no idea how loud anything is anyway.

[Norm Kaiser] “And then invariably I get the little dingdong chime from Facebook that I just got a new message and my eardrums get pierced.”

That’s because computer audio is probably normalized for 0dBFS which is much louder than broadcast audio. I don’t have this problem because the sound card and speakers that my broadcast audio use are different from the speakers that my computer uses for it’s silly beeps. That’s the proper way to configure an audio workstation.

[Norm Kaiser] “Is this correct? I’ve read elsewhere that people report this outcome (audio sounds so quiet) and very experienced users report that it’s correct, but is it?”

Yes it is correct and no, nothing sounds low on my system because my audio monitors are calibrated properly. You wouldn’t color correct on a video monitor that wasn’t properly calibrated, why would you mix audio on speakers that weren’t properly calibrated? If you had a separate sound card with separate speakers for working with audio you wouldn’t have this problem. You are also overcompensating by limiting your audio to -24dBFS peaks so you audio actually is too low on top of that.

[Norm Kaiser] “Does the broadcast audio system somehow amplify the sound once it gets to the viewer’s TV? Or do computers play back sound poorly in that they need a really hot signal to sound right?”

It really doesn’t matter what the level is as long as it is consistent. Once you adjust the volume on your TV, you expect everything to sounds equally as loud. Having a broadcast standard whether it’s -24dBFS, -18dBFS, or whatever is just so that all of the audio is consistent relative to all other audio in the program. Most of this comes from the Analog world where loud audio would saturate an analog tape and distort so you want to be careful not to do that. -18dBFS digital is about 0VU analog so -24dBFS average is like -3dBu average in the analog world.

Here is a chart that shows the relative digital and analog measurements:

~jr

http://www.johnrofrano.com
http://www.vasst.com

[Norm Kaiser]

OK, bam, I’m following you *mostly* now. Thank goodness you are here. I realize now that what I was doing with the Wave Hammer is dumb. And you were politely telling me so.

I get it. So to meet CALM my overall average must be -24dBFS but I can occasionally peak at -10dBFS, so long as my average is still -24. I gotcha. And what I’m doing now with the Wave Hammer is “smashing” the peaks that I should be allowing and effectively crushing sound quality.

OK, so I want to back up and do it the right way. the iZotope tool is very tempting, but if I can save $350, I’d obviously prefer to do that.

So here’s what you suggested in your earlier post:

If you want to do it yourself the old-school way, I would start by placing Wave Hammer Surround on the master audio bus and set it to hard limit at whatever the peak value is (i.e., -10dBFS). Start with the “[Sys] Master for 16-bit” preset and adjust the Output Level on the Volume Maximizer to -10dB or whatever their peaks need to stay under. Then use a SMPTE 1KHz test tone @ -20dB to adjust the Threshold until the peak meters read -20 dB again to compensate for the limiting so that your volume is still accurate. This is the same test tone that you would use with “bars & tone” and the beeps for your slate count down if they want those.

Let me digest this:

Start with the “[Sys] Master for 16-bit” preset and adjust the Output Level on the Volume Maximizer to -10dB

Got it. I know how to do this.

Then use a SMPTE 1KHz test tone @ -20dB…

OK, I need help. Where do I get the test tone and where do I put it? On the timeline?

[John Rofrano]

[Norm Kaiser] “OK, I need help. Where do I get the test tone and where do I put it? On the timeline?”

As I stated above, normally this is the “tone” under your standard “bars & tone” that you lay down before your slate at 00:58:29;29 (NTSC) when your broadcast starts at 01:00:00;00. You can make one with Sound Forge (which is what I did) or find one on the Internet. These are required for analog (e.g., HDCAM tape, etc.) submissions but may be optional for digital submissions depending on the broadcaster. They tell the broadcaster what -20dB should actually be according to your show.

Typically a show starts at 00:58:29;29 with 1 minute of bars and tone, 30 seconds of slate, the last 10 second of which are a count down, the last 1 second is broadcast black and your show starts at 01:00:00;00. If you’re not laying down bars & tone and slate before your show then just use the [Sys] Master for 16-bit preset, go to the Volume Maximizer tab and set the Threshold to -8.1 dB and the Output level to -10.0 dB and that should compensate for the compression. You can drop a -20dB test tone into an empty project to check that it’s set correctly. The VU meters should still read -20dB with the Wave Hammer preset applied.

If you want to calibrate your speakers, use a Pink Noise test tone at -20dBFS. Drop it into an empty project, create loop region around it and hit play. Then get a Sound Pressure Level Meter, set the weighting to “C Weighted” and the Response to “Slow” and then adjust your speaker volume until the SPL reads 75dB at your listening position with the mic pointed at the speakers. (technical note: -20dBFS is actually 85dB but 85dB is too loud for near field monitors and should only be used when calibrating for theatre) Then your speakers will be at the proper volume for you to judge loudness with your ears. WARNING! 0dBFS is 105 decibels which is about the volume of a power lawn mower so your windows beeps will be incredibly loud unless you use a separate speaker system (but you’ve already experienced this). The calibration assumes that your speakers are placed properly, They should be as far away from each other as they are from you. That is to say, if you are sitting 3 feet from the speakers then they should be 3 feet from each other to form a triangle with your ears. It’s also important to have them level with or slightly above you ears. I have mind on pedestals to get them to the proper height. I also use Auralex Mopads to isolate them.

You can pick up an SPL meter at Radio Shack or if you have an iPhone, I can highly recommend the SPL Meter app from Studio Six Digital. It’s extremely sensitive and accurate and.comes from a company that builds test equipment; a worthwhile 0.99 cent investment.

All of this assumes that your speakers and sound card are giving you a flat response and not coloring the sound. If you are using the built-in sound chip on your computer’s motherboard, chances are it is designed to make DVD’s and other movies sounds great so it’s altering the sound that you hear and you can’t trust them for doing audio work. It’s like trying to color correct video on a TV set. Every TV looks different! You really need studio monitors to perform proper mixing but if your program material isn’t demanding then mix them on anything you want. Just be aware that your sound chip might be artificially boosting the base or treble to make it sound better so you aren’t going to get what you hear when you listen on another device.

~jr

http://www.johnrofrano.com
http://www.vasst.com

[Norm Kaiser]

Well, I still haven’t had much success.

Here’s what I did:

– Placed a sample piece of material on the timeline

– On that sample’s audio track, I added the Wave Hammer plugin

– On the Volume Maximizer, I set the Threshold to -8.1dB and the Output level to -10dB.

– I then added the Track Compressor.

– On the Track Compressor I have the Threshold set to -24dB and the Amount ratio set to 3:1.

I then generate the loudness log, but my Integrated LUFS is still at -15.29.

I’m still obviously missing something. At this point, the only way I can get it down to -24 LUFS is by decreasing the volume on the track bit by bit until I get there.

But decreasing the volume on the track seems so counterintuitive to me. Because when I render the piece and if I place that rendered piece back on the timeline, the waveform is so tiny!

And that’s where I’m lost. Is that “tiny waveform signal” going to be “right” when broadcast? It seems to me that signal would have to be amplified just to be heard. I get that my PC is likely optimized for 0dB, but doesn’t reducing the volume in Vegas smash the signal such that when it’s broadcast bot the desired audio AND the noise floor get amplified?

[John Rofrano]

[Norm Kaiser] “- On that sample’s audio track, I added the Wave Hammer plugin”

Sorry but you need to go back and read my original post. Wave Hammer goes on the Master Audio Bus. Track Compressor is used on the Tracks. Here are my exact words:

“If you want to do it yourself the old-school way, I would start by placing Wave Hammer Surround on the master audio bus and set it to hard limit at whatever the peak value is”

So be sure to use Wave Hammer on the Master Audio Bus and not the Track.

[Norm Kaiser] “- I then added the Track Compressor.”

The Track Compressor should come BEFORE Wave Hammer. This is why you place Wave Hammer on the master audio bus so that it is last after everything else. The position of audio plugs in the chain is extremely important because the output of one affects the input of the other. Since Wave Hammer Volume Maximizer is being used as a Brick Wall Limiter, you want to place it last in the chain.

[Norm Kaiser] “I’m still obviously missing something. At this point, the only way I can get it down to -24 LUFS is by decreasing the volume on the track bit by bit until I get there.”

Yes, you need to adjust the volume appropriately. The compressors are there to catch any transients so that you don’t have to ride the volume with an envelope but you still need to set the volume to a reasonable level to start with. You should get the audio mix in the ball park and then use the track compressor to smooth it out.

[Norm Kaiser] “But decreasing the volume on the track seems so counterintuitive to me. Because when I render the piece and if I place that rendered piece back on the timeline, the waveform is so tiny!”

Tiny is relative. 😉 When you play it back, see if the loudness meter is registering around -24 dBFS. If it is, then it’s at the right volume. if you have any way of recording a TV program, like from a DVR, and getting it onto your computer, you’ll see that the waveform from TV audio is quite low and massively compressed.

[Norm Kaiser] “And that’s where I’m lost. Is that “tiny waveform signal” going to be “right” when broadcast? It seems to me that signal would have to be amplified just to be heard.”

Did you calibrate your speakers like I explained? if you did not, then you have no idea how loud anything is. You are just guessing.

[Norm Kaiser] “I get that my PC is likely optimized for 0dB, but doesn’t reducing the volume in Vegas smash the signal such that when it’s broadcast bot the desired audio AND the noise floor get amplified?”

Which noise floor? When you reduce the volume on a clip you are reducing the noise floor with it. If you having problems with the noise floor generated by your computer then you need a better audio device. Motherboard chips are quite noisy and they are not intended for professional audio creation.

~jr

http://www.johnrofrano.com
http://www.vasst.com