Mastered audio files clipping in Vegas 16 and 17. Can't determine why.

Norrin_Radd22 wrote on 7/29/2020, 5:00 AM

Hi gang,

When I master audio files in Cubase, to peak at -0.3dB, they display as being 0.6dB loud in Vegas all the time. I have no idea why. I am sure there is some default setting that Vegas has that video industry users are all used to. As an audio guy, however, I find it weird that professional audio would be clipping in Vegas when nothing has been touched. -0.3dB should be the same no matter what is playing it.

Even regular music from bands clips when I drag it in, so it can't just be something I do with my files.

Help?

Windows 10 64bit, Vegas Pro 16 and 17

-Matt-

Comments

Marty74 wrote on 7/29/2020, 5:02 AM

Same issue as you. (running cubase, musician , vegas Pro 16 blabla, no clipping in cubase and after mastering, and clipping in Vegas).

adis-a3097 wrote on 7/29/2020, 5:33 AM

Are you guys using mp3 or other lossy audio formats?

Norrin_Radd22 wrote on 7/29/2020, 7:58 AM

Both. And in both cases, it registers as the exact same level in Cubase. In Vegas I can't audibly hear any clipping either, which makes me feel there is almost a fake volume ceiling, or some other sneaky little thing. ?

adis-a3097 wrote on 7/29/2020, 8:07 AM

Try rendering to .wav file in Cubase, import in Vegas and see what happens. Don't change the sampling rate in Vegas...

Musicvid wrote on 7/29/2020, 8:29 AM

Vegas now uses true-peak (intersample) meters, so it can monitor broadcast levels. I don't know what Cubase does, but the scales are obviously different. That's only one reason recording best practices are still <= -1.0 dBFS peak. Best not to shave your margins to be "competitive."

rraud wrote on 7/29/2020, 9:20 AM

I suspect different meters and scales are being used. Are they on VP's final output bus? I have never experienced clipping after rendering PCM files based on the peak dBFS master bus meter. Occasional 'overs' may occur encoding lossy formats though if levels are close to 0.0dBFS.

Norrin_Radd22 wrote on 7/29/2020, 3:37 PM

If you guys have any, just drag any modern music files in to Vegas, the louder the better. Tell me if they clip or not. I tried Meshuggah's "I" album in mp3 format. I did not rip this, so it never touched my machine before testing. In Cubase it goes to 0dB and never over (still extremely loud for a commercial release).

In Vegas it goes to exactly 1dB, which seems to indicate that Vegas is always reading 1dB hotter than CD quality albums. This is accurate to what I have experienced with my own files as well.

rraud wrote on 7/29/2020, 5:38 PM

Vegas uses the internal floating point 32 bit audio which can exceed the meter's 0.0dBFS scale and show on the + side. When rendering to a fixed bit depth though, clipping will occur when the master bus meter exceeds 0.0dBFS.

Many 'loudness war' CD's have inter sample peak clipping.

Musicvid wrote on 7/29/2020, 6:38 PM

What version of Cubase do you have?

Norrin_Radd22 wrote on 7/29/2020, 8:33 PM

What version of Cubase do you have?


9.5
As for loudness wars, this isn't a matter of trying to ramp up the perceived loudness without going over. This is literally Vegas saying everything is 1dB louder than it actually is.

Musicvid wrote on 7/29/2020, 9:29 PM

9.5 doesn't say which version you have. Can I assume it is Cubase Elements? Maybe show us your splash screen?

Norrin_Radd22 wrote on 7/30/2020, 2:40 AM

Oh, no haha. Cubase Pro. I am a professional video game composer.

But I am pretty sure it is not related to my Cubase, as even just dragging in random MP3s, they clip in Vegas.

adis-a3097 wrote on 7/30/2020, 3:08 AM

Your project sample rate and media sample rate, do they match?

Musicvid wrote on 7/30/2020, 11:21 AM

And which version and build number of Vegas do you have?

rraud wrote on 7/30/2020, 11:44 AM

, select as even just dragging in random MP3s, they clip in Vegas.

Try right-clicking the event, select 'Properties' and remove the 'Normalize' check-mark, Reapply the check mark and click 'Recalculate'.

Musicvid wrote on 7/30/2020, 11:59 AM

So you have A85 / R128 metering available. I think these streaming best practices will interest you, since Youtube and others have begun throttling audio levels that exceed their loudness guidelines, with predictably bad results apparently.

I have downloaded some of your samples from Youtube.

Broadcast is ~6dB softer than Apple.

 

Musicvid wrote on 7/30/2020, 1:12 PM

Matt, it turns out there is a simple fix for you, I believe.

You are creating audio for local delivery (not online games, right?).

Go ahead and push the ceiling all you want. -0.1 TP, -10 ILU, -6LRA, have at it. Whatever the kiddos want to make their hearing like mine someday, with their headphones turned up 93dB.

Streaming delivery is a different matter. Without rehashing points made above, your audio is too loud.

When encoding your captures for streaming video delivery, turn down the gain by -6dB. That will put your LUFS at Apple levels, and that will put them in line with other streaming services, as well. No need to re-encode for the 'net unless you want to. Issue solved, no hair-splitting necessary. As for your 1dB discrepancy, another thing to check is that the same audio drivers are selected in both applications.

I know the temptation to want to have your viewers experience a theatrical moment when they play your online videos. The movie producers want us to do so, as well. But the ceiling is lower, and the future is with those who choose the light side of the force. Welcome to the forums.

rraud wrote on 7/30/2020, 3:47 PM

In addition to MV's ( @Musicvid ) comments. The on-line 'Loudness Penalty' analyzer, can give one an idea how soft or loud a project is compared to the 'recommended' streaming levels for music YT, iTunes, Spotify, ect. (typically around -14 LUFS integrated), not to be confused with peak dBFS

"Whatever the kiddos want to make their hearing like mine someday"

+11, same here

Norrin_Radd22 wrote on 7/31/2020, 7:16 AM

I appreciate the posts guys. The loudness wars are a whole other issue. I personally don't like to squash my music, or even make music that is squashable a lot of the time. The track you saw was one of the few that is meant to be slammed.

But the overall number of how loud a track is didn't end up being my issue here. My issue is that Vegas's meter and Cubase's meter are not the same.

I did a test where I put a limiter on a track to make it EXACTLY -6dB, and Vegas tells me it is actually -5.2dB. I am not even talking about clipping. I simply mean, why does it read -6dB as -5.2dB?

I use Vegas Pro 16, abut I also recently got Vegas Pro 17, and the issue remained, lock step. That's how I knew there was something wrong with Vegas. An entirely new install should have fixed the problem if I did something to my old version.

Even when I normalize a track inside Vegas Pro 17, this -6dB track for example, becomes 0.6dB. I don't usually use Normalize in Vegas since I use Cubase to treat all my audio, so this is actually news to me. For the record, the Normalize peak level (dB) is set to -0.1

Could it be my external soundcard somehow? I have a UR242. But I have no idea how it would be affecting a track inside Vegas after normalization, but before the final meter...

Guys, what the hell is going on on my machine?

Marco. wrote on 7/31/2020, 8:30 AM

What about checking standardized test tones first?

Norrin_Radd22 wrote on 7/31/2020, 8:56 AM

I'm not sure that would do it. If I normalize within a fresh install of Vegas and it displays as higher than what I have set inside Vegas as the normalize peak, then where is the problem? How do I troubleshoot this?

Vegas gets the normalize instruction, and "normalizes" it to the set peak, then it also says its basically .7dB higher than what I set as the max.

Where is the failure point?

Had this exact same issue on a Vegas Pro 16 I was using prior, so it must be something on my machine... But where in my machine would it allow me to make the file louder after normalization, but also before the meter?

Musicvid wrote on 7/31/2020, 1:22 PM

What about checking standardized test tones first?

I did, using the BBC -18dbfsstereotone.wav as the benchmark. In both native and normalized mode, it is within ±0.1 dBFS of expected.

Unable to duplicate your reported behavior. I have searched the term "meters" all the way back to VP14 and I don't see any similar reports. Do post back when you have found your solution; it may help others, and best of luck.

mp3 files often have an audible "click" on the first sample upon playback because of the metadata header placement. Is this what your peak meters are locking onto?

adis-a3097 wrote on 7/31/2020, 3:04 PM

Your project sample rate and media sample rate, do they match?

Bump.