2 voices 2 channels headaches

[Nick Mirro]

Hi. Using two different Canon camcorders (XF400, XC15), I've been recording one person's voice to each of the two channels on each. When editing, I duplicate the audio track, convert both tracks to mono and then reverse one of the track's channels. This centers both voices.

But always have huge problems with audio levels and also balancing. Using this approach, cameras need attenuation (~20 dB) for some reason. Canon warns that (in certain scenarios) clipping is possible without attenuation, and that the meters won't always show clipping. Seems correct, since it's easy enough to hear in the headphones.

Then when editing, one channel frequently comes back significantly quieter, even though they seemed equal in the headphones.

Both camcorders have the ability to mix the two voices to the center, but only for purposes of headphone monitoring. Tracks always end up one voice left one voice right.

I'm trying to avoid bringing the field monitor, but have been struggling with this approach. Is there something inherently problematic with how I've set it up?

[Paul R Johnson]

Some steps are just unnecessary but you are recording voices? Speaking, not yelling? I can’t imagine why you need attenuation at all. A moderately quiet source shouldn’t clip and it should, if anything be too quiet? In your scenario, I’d bring the two camera clips onto premiere. Each would have stereo audio and be hard left or right, so all I would do is convert each one to a mono track. Keep in in mind that if you duplicate, flip and add, then I’d anything exists in the ‘empty’ channel then if the level of the clip is sitting just a bit below max, then merging them will go over.

All I do is just convert each half empty track to a mono track. I pan them almost mono, but with a small left right shift to reflect where they are. I’m talking maybe 11 and 1 o’clock no more, perhaps even 11.30, just not 12 and 12. On speakers unnoticed, just detectable on headphones with eyes closed. Never heard distortion. When you listen to the one legged audio, is that clean and undistorted? If so maybe it creeps in with your copy, flip and add process? If you reduce the clean tracks by 3dB then fo the flip and add, that should result in distortion free audio, but does your editor not have a simple convert to mono function?

[Patrick Tracy]

That sounds like a software limitation. In Vegas Pro you simply right-click the waveform, select Channels, select Left only or Right only. Then it's easy enough to deal with levels. What software are you using? Video NLEs are sometimes a bit limited when it comes to audio.

Doubling up the same audio will yield a 6.02dB increase in level. You'll need to account for that.

[Rick Reineke]

"<i>"Doubling up the same audio will yield a 6.02dB</i>"
> Yes, but OTOH, two mics can affect the output level of one or both when they are summed to a mono bus.

[Patrick Tracy]

Originally Posted by Rick Reineke

"<i>"Doubling up the same audio will yield a 6.02dB</i>"
> Yes, but OTOH, two mics can affect the output level of one or both when they are summed to a mono bus.

I'm not quite sure what you're saying there. If there's some kind of dynamics processing (AGC, compression), the level of one mic could be affected by the other. If there's no dynamics processing, the second signal shouldn't change the actual level of the first, though masking can change perceived levels. Generally, when you double unlike signals you get a signal increase of approximately 3dB, but that's less consistent than doubling identical signals.

[Nick Mirro]

Thanks much for the help.

Here's where this is coming from. Both camera manuals say the following...

"If the input level is too high, audio may become distorted even if the audio level indicator shows an appropriate level. If the sound is distorted even though the audio level indicator shows an appropriate level, activate the microphone attenuator."

I've never had this issue with a line level field mixer. Is Canon audio processing optimized for a narrow range of standard gain mics?

Editor is Davinci Resolve. The doubling tracks up (+6dB) issue in the editor does not account for unexpected clipping that originates in the camera. Still I was not aware of this. I can see now in DR that it does not create clipping originating in the camera but it can worsen it.

With this setup, unattenuated or untrimmed audio always creates distortion. I'm starting to think it's the mics, which not only have high sensitivity, but are also worn just to the side of the wearer's mouth. We are also always outdoors and speaking less than 5 feet apart.

This is why I bought them. The sound is much more bright and natural than with lavs. The Countryman E6s as configured have sensitivity of -43 dB.

So the mixer must have given me greater control. Both cameras have the same clumsy manual audio adjustment dials. They don't encourage fine tuning like the mixer. That and Canon's limited audio processing...

[Paul R Johnson]

You need to do some proper tests to establish if the cameras are capable or not. With camera without manual gain, distortion is always likely, usually with wind noise. Why not just buy a zoom, feed your two mics into that and just use camera audio for sync. One camera has better audio than the other? I don’t think one had XLR? Could you try recording two mics to that one and just using it’s audio. Xlrs usually mean a little more care put into the audio? Or are they both bad?

The audio on my JVCs isnt bad, but it always seems to be slightly automatic, even switched to manual and levels adjusted. As if a compressor is always lurking.

When you look at the audio waveform, is it over level peak distortion, or just distortion created earlier in the chain and not related to the level? If the audio has distorted in the preamps because it’s too loud, it’s too late to cure it by adjusting levels later?

[Christopher Young]

Talk back radio has had the problems you are having for years. I've had the same problems producing Financial Planning videos for the FPA for their training videos every month over nine years. Three and four way round table interviews and discussions. Which can get quite lively at times with audio levels all over the place. Back to the talk back radio problem and one of their solutions that I've used now for all manner of productions from entertainment shows to sport.

The quickest and one of the simplest ways and a massive time saver which will give you consistently good results is to render out each individual interviewee's mono audio track as a .WAV file and then run each of them THE LEVELATOR and create new WAV files for each camera track. After running all tracks through Levelator I just sync them up with the original audio tracks for each speaker and mute the original tracks. With most NLEs you can lock the the new audio tracks to their appropriate video track.

Regardless of the audio levels of each individual speaker rising and falling throughout an interview or discussion you will end up with surprisingly even well matched tracks between each speaker and achieve this in minutes. Not hours of playing around with volume envelopes, normalization and chasing down peaks etc.

Read up on Levelator and see if it will work for you. It's a free app. An extract from their site:

"What is The Levelator®?

Do you believe in magic? You will after using The Levelator® to enhance your podcast. And you'll be amazed that it's free, now even for commercial use.

So what is The Levelator®? It's software that runs on Windows, OS X (universal binary), or Linux (Ubuntu) that adjusts the audio levels within your podcast or other audio file for variations from one speaker to the next, for example. It's not a compressor, normalizer or limiter although it contains all three. It's much more than those tools, and it's much simpler to use. The UI is dirt-simple: Drag-and-drop any WAV or AIFF file onto The Leveler's application window, and a few moments later you'll find a new version which just sounds better.

Have you ever recorded an interview in which you and your guest ended up at different volumes? How about a panel discussion where some people were close to microphones and others were not? These are the problems the post-production engineers of Team ITC here at The Conversations Network solve every day, and it used to take them hours of painstaking work with expensive and complex tools like SoundTrack Pro, Audacity, Sound Forge or Audition to solve them. Now it takes mere seconds. Seriously. The Levelator® is unlike any other audio tool you've ever seen, heard or used. It's magic. And it's free.

When we developed the IT Conversations component-based show-assembly system, we realized all the components had to be of the same loudness or the results would sound awful. We limped along for many months using the RMS normalization functions in various applications, but the results weren't satisfactory and it required tools and skillsets that some of our post-production audio engineers didn't have. One of our best engineers, Bruce Sharpe, offered to write a standalone software RMS normalization utility, which we've been using as part of our production system CNUploader since 2005."

Check it out. You have nothing to lose and maybe heaps to gain... pun intended:)

Chris Young

The Levelator® from The Conversations Network

[Nick Mirro]

I'm seeing DR has a good audio normalizer. So I just have to ensure the audio being recorded to both cameras (they have identical xlr audio hardware) is set to the correct level (i.e. no clipping).

Seems obvious enough but it's pretty tricky as per Christopher. Need to listen carefully to the peaks of the loudest voice talking. DR has no problem getting voices to the same level, even embedded to the same audio track. All else is well if there is no clipping.

[Andrew Smith]

See, this is one of those things that makes DVinfo just so dang cool. There's every chance you'd never stumble across the Levelator in your lifetime if it weren't for this.

Andrew

[Christopher Young]

True, True. Andrew, I think you are in Oz, aren't you? Do you ever remember youth TV series like The Rock Eistedford and Wakakirri Story Dance? They were major national performance music and dance shows. "Wakkakiri Story Dance" (An Australian Aboriginal name) had everything from loud booming show presenters to young kids being interviewed along with everything from quiet ballet to thumping rock. An absolute audio post nightmare. Audio post time was cut down from a couple of days per show to literally an hour or so with The Levelator.

We would edit the entire one-hour shows purely based on vision and graphic content and then run each individual audio track, commentary, atmos, music, and sound FX through The Levelator and then sync all Levelator output tracks with the master vision We would then adjust and duck levels for each track for the correct mix and balance across all audio tracks. NOTE: The Levelator will adjust each individual audio track output to -3dB which is the radio broadcast standard. For TV broadcast in Oz -10db Is your max deliverable. The final mix was then adjusted for a peak level of no more than -10dB to meet those broadcast standards. Then it was render out time to a master for network delivery.

A lot of people bagged The Levelator. But most who tried it were pretty amazed. There is no substitute for a pure audio sweetening and mastering job but for day-to-day TV The Levelator sped up production amazingly. Speed is money!

EDIT: Forgot to mention why The Levelator was such a bonus for us. Each year we would end up with shows from all over the country. About 70 shows of about 3 hours each with 4 cameras per show. That's from the finalists to go into the final edits. About 840 hours of footage. On average for all the heats throughout the year, the total roll time was about 3,000 hours, yes 3,000 hours of camera footage that was all post switched from TC locked camera ISOs. Each run of DVDs was about 5-7000 DVDs! Phew! Glad that's all behind me noe.

Chris Young

[Rick Reineke]

In US TV broadcast, -10db peaks was the max deliverable peak level as well (along with a 1kHz -20dBFS reference tone) .But that was years ago. Practically all US TV networks now use the loudness scale (typically LU), -24 LUFS (integrated , plus or minus two). True peak, is not a huge factor, usually a -1.9 dBTP max

[Paul R Johnson]

I tried to download it for my MacBook but it's too old - I'll try windows later.

[Christopher Young]

Originally Posted by Rick Reineke

In US TV broadcast, -10db peaks was the max deliverable peak level as well (along with a 1kHz -20dBFS reference tone) .But that was years ago. Practically all US TV networks now use the loudness scale (typically LU), -24 LUFS (integrated , plus or minus two). True peak, is not a huge factor, usually a -1.9 dBTP max

Yes true, we are basically the same, -24LKFS, but didn't want to drive people crazy with the tech levels. If anyone in Oz is interested the standards here they are covered by OP48 and OP59.

https://www.freetv.com.au/wp-content...-July-2016.pdf

https://www.freetv.com.au/wp-content...tober-2018.pdf

Plus we had to add a compliance line to our program slates stating that the program audio did indeed comply with OP48&59.

Chris Young

[Rick Reineke]

Thanks Chris!

" but didn't want to drive people crazy with the tech levels
> The US ATSC A-85 document takes 70+ pages.. to state the same.