Audition CC: Loudness Normalization …

*** UPDATE: Please note this post was written in 2014. The current version of Adobe Audition CC has been greatly enhanced, specifically in regards to the Match Loudness Module. It is now possible to define a True Peak Maximum, as well as Integrated/Program Loudness targets. It is also possible to customize Loudness Normalization Tolerence.

Adobe Audition CC has a handy Match Volume Processor with various options including Match To/ITU-R BS.1770-2 Loudness. The problem with this option is the Processor will not allow the operator to define a True Peak Ceiling. And so depending on various aspects of the input file, it’s possible the processed audio may not comply due to an unsuitable Peak Ceiling.

For example if you need to target -16.0 LUFS Program Loudness for internet/mobile distribution, the Match Volume Processor may need to increase gain in order to meet this target. Any time a gain increase is applied, you run the risk of pushing the Peak Ceiling to elevated levels.

The ITU Loudness processing option does supply a basic Limiting option. However – it’s sort of predefined. My tests revelaled Peak Ceilings as high as -0.1dBFS. This will result in insufficient headroom for both True Peak compliance and preparation for MP3 encoding.

The Audition Match Volume Processor also features a Match To/True Peak Amplitude option with a user defined True Peak Ceiling (referred to as Peak Volume). This is essentially a True Peak Limiter that is independent of the ITU Loudness Processor. For Program Loudness and True Peak compliance, it may be necessary to run both processing stages sequentially.

processor

There are a few caveats …

[– If the Match Volume Processor (Match To/ITU-R BS.1770-2 Loudness) applies limiting that results in a Peak Ceiling close to full scale, any subsequent limiting (Match To/True Peak Amplitude) has the potential to reduce the existing Program Loudness.

[– If a Match Volume process (Match To/ITU-R BS.1770-2 Loudness) yields a compliant True Peak Ceiling right out of the box, there is no need to run any subsequent processing.

Conclusion

If you are going to use these processing options, my suggestion would be to make sure the measured Program Loudness of your input file is reasonably close to the Program Loudness that you are targeting. Also, make sure the input file has sufficient headroom, with existing True Peaks well below 0dBFS.

If you are finding it difficult to achieve acceptable results, I suggest you apply the concepts described in this video tutorial that I produced. I demonstrate a sort of manual “off-line” Loudness Normalization process. If you prefer to handle this in real time (on-line), refer to my article “Podcast Loudness Processing Workflow.”

-paul.

Podcasting System featuring the Allen & Heath XB-10 Console …

I continue to look around for a Broadcast Console that would be suitable to replace my trusty Mackie Onyx 1220i FW mixer. I was always aware of the XB-10 by Allen & Heath, although I did not pay much attention to it due to it’s use of pot-styled channel faders as opposed to sliding (long-throw) faders.

ah-mixer-480

Last evening I skimmed through the manual for the XB-10. Looking past the pot-styled fader issue this $799 console is packed with features that make it highly attractive. And it’s smaller than my Mackie, checking in at 13.2 inches wide x 10 inches deep. Allen & Heath also offers the XB-14-2 Console. It checks in at 15.2 inches wide x 18.3 inches deep with ample surface space for long-throw sliding faders. Bottom line is it’s larger than my Mackie and the size just doesn’t work for me.

XB-10: The Basics

Besides all the useful routing options, the XB-10 has a dedicated Mix-Minus channel that can be switched to receive the output of a Telephone Hybrid or the output of the bi-directional USB bus. In this case it would be easy to receive a Skype guest from a computer.

The console has latching On/Off switches on all input channels, supports pre-fader listening, and has built-in Compressors on channels 1-3. The manual states ” … the Compressor is optimized to reduce the dynamic range of the presenter microphone(s). Low signal levels are given a 10dB gain boost. Soft Knee compression activates at -20dBu, and higher level signals are limited.” Personally I would use a dedicated voice processor for the main presenter. However having the dynamics processing on-board is a useful feature, especially when adding additional presenters to the program mix.

The XB-10 is also equipped with an Output Limiter that can be used to ensure that the final mix does not exceed a predefined level. There is an activation switch located on the back panel of the device with a trim pot control to set the limiting threshold. If the Limiter is active and functioning, a front panel LED illuminates.

One other feature that is worth mentioning is the Remote Connector interface located on the back of the device. This can be used to implement CD player remote triggering, ON AIR light illumination, and external metering options.

I decided to design a system using the XB-10 as the controller that is suitable for flexible Podcast Production and Recording. Bear in mind I don’t have any of these system components on hand except for older versions of the dbx Voice Processor and the Telos Phone Hybrid. I also have a rack-mounted Solid State Recorder by Marantz, similar to the Tascam. I’m confident that all displayed components would work well together yielding excellent results.

Also note there are many ways to integrate these components within the system in terms of connections and routing. This particular design is similar in concept to how I have my current system set up using the components that I currently own (Click to Enlarge).

AH-system-480

System Design Concepts and Selections

The mic of choice is the Shure SM7B. The was the first broadcast style mic that I bought back in 2004 and it’s one of my prized possessions. As far as I’m concerned it’s the most forgiving broadcast mic available, with one caveat – it requires a huge amount of clean gain to drive it. Common +60dB gain trims on audio mixers will not be suitable, especially when setting the gain near or at it’s highest level. This will with no doubt result in problematic noise.

In my current system I plug my dynamic mic(s) into my dbx 286a Voice Processor (mic input) and then route the processor’s line output to a line input on one of the Mic channels on my Mackie mixer. By doing so I pick up an additional +40dB of available gain to drive the mic. Of course this takes a bit of tweaking to get the right balance between the gain setting on the processor and the gain setting on the Mackie. The key is not to max out either of the gain stages.

I’ve recreated this chain in the new design using the updated dbx 286s. In doing so the primary presenter gets the voice processor on her channel. If there is the necessity to expand the system by introducing a second presenter, I’ve implemented the Cloudlifter CL-1 gain stage between the mic and the console’s mic input on channel 2. The CL-1 will provide up to +20dB of additional clean gain when using any passive microphone. Finally I point to the availability of the on-board dynamics processor and consider this perfectly suitable for a second presenter.

I mentioned the XB-10 has a dedicated telephone interface channel with a built in mix-minus. Once again I’ve selected the Hx1 Digital Telephone Hybrid by Telos Systems for use in this system. The telephone interface channel can be set to receive an incoming telephone caller or something like the Skype output coming in from a computer. I’ve taken this a step further by also implementing an analog Skype mix-minus using the Console’s Aux Send to feed the computer input. The computer output is routed back into the Console on an available channel(s).

As noted the USB interface on the Console is bi-directional. One use case scenario would be to use the computer USB output to send sound effects and audio assets into the program mix. (I am displaying QCart for Mac as a possible option).

The rest is pretty self explanatory. I’m using the Monitor output bus to feed the studio speakers. The Console’s Main outputs are routed to the Tascam recorder, and it’s outputs are routed to an available set of inputs on the Console.

Like I said I’m fairly confident this system design would be quite functional and well suited for flexible Podcast Production and Recording.

In closing beginning in 2004 besides designing sort of generic systems based on various levels of cost and complexity, it was common for an aspiring Podcast Producer to reach out to me and ask for technical assistance with the components they purchased. In this case I would build detailed diagrams for the producer much the same as the example included in this post. A visual representation of system routing and configuration is a great way to expidite setup when and if the producer who purchased the gear is overwhelmed.

Note:

At one time I was providing a service where two individual participants were simultaneously calling into my studio for interview session recording. Since I had two dedicated phone lines and corresponding telephone hybrids, the participants were able two converse with each other using 2 Aux buses, in essence by creating two individual mix-minuses.

Here is the original diagram that I built in October 2006 that displays the routing of the callers via Aux sends:

dual-mm-480

Even though the XB-10 console contains a single Aux bus, a similar configuration may still be possible where an incoming caller from the telephone hybrid would be able to converse with a Skype guest, minus themselves. I need to read into this further before I am able to make a determination on whether this is supported.

Components:

[– Shure SM7B Broadcast Dynamic Microphone
[– Cloudlifter CL-1 Gain Stage
[– Allen & Heath XB-10 Broadcast Console
[– dbx 286s Voice Processor
[– Telos Hx1 Digital Telephone Hybrid
[– Tascam SS-R200 Solid State Recorder

Optional:

[– QCart for Mac OSX
[– KRK Rokit 5 Powered Studio Monitors

-paul.

spotPoint Lighting – Simultaneously …

I’m working on the next version of spotPoint Lighting. The next version will include simultaneous use and control of Spot and Point Lights. The example below is actually a Title as opposed to the Effect that was initially released. I’m trying to decide which format would be more useful. Having two independent text layers right within the package is definitely a plus. OTOH Effects are much cleaner implementations, and least for me – all due to their ability to be applied to individual clips. Titles are fine for timeline compositing. They do add a bit of clutter to the mix …

Below I used the Spot Light to warm up the sky independent of the Point Light.

-paul.

spot-matrix

Film Zone

Cinemascope Toolkit ver.1.2 has been released. The Crop Guides popup now displays one of three options:Letterbox, Film Zone, Letterbox and Film Zone. The Film Zone is essentially a set of colored cropping guides less the letterbox matte(s). Viewing the underlying clip with the Film Zone displayed on it’s own makes it easy to view what is being cropped. Also, the Film Zone display works well when the underlying clip is very dark at the top and/or bottom of the frame. You can set the Film Zone color to orange (default), black, or white.

Also new in this release is the capability to reposition the clip manually by clicking and dragging the center point object (Drag Target). When doing so the clip positioning sliders in the EFX UI will update accordingly.

Here is a look at the new controls:

In the image matrix below you can see the top clip was repositioned (and scaled). The visible Film Zone clearly displays the 2.35:1 frame. In the middle image the 2.35:1 Safe Zones are displayed along with the Film Zone. Note the clips reduced opacity. The bottom image is the cropped output.

Please note you must set the FCP X Player Background to Black when using Cinemascope Toolkit. Do this in the application Preferences/Playback. When you switch on the Safe Zones display the clip opacity is reduced. This provides a clear view of the zones. If the player background is set to Checkerboard, there’s nothing behind the clip – it’s transparent. The clip’s opacity reduction will be prevelant and this feature will be useless.

Also – I designed the matting system to be independent of the clip’s image layer. Any agressive grading or exposure adjustments will have no effect on the visual state of the letterbox matte(s).

-paul.

Download Cinemascope Toolkit

colorFloat

Last eve I was sifting through the Apple App Store looking for a simple utility to quickly convert RGB color values to corresponding float values (RGB integer / 255 = float). I decided to build my own Cocoa application with a few added enhancements.

High-res Image: colorFloat

Run the standard OSX Color Picker and press the second toolBar option (Color Sliders). Select the RGB Sliders option in the popup menu. Notice each RGB value changes as you move through the color spectrum. We can divide each one of the displayed values by 255 to return float values that can be used in source code authoring. In colorFloat the user adds an input RGB value (x3), converts, and appends each conversion result to the desired color channel. The final action displays the corresponding color for confirmation.

I also built in support for what I refer to as Dynamic Floats. Notice the Dynamic Floats HUD located in the high-res image. The Float value strings change dynamically as you move around the color wheel or change the values of the RGB sliders. This feature allows the user to easily sift through the color spectrum to view corresponding floats in real time.

Lastly, I added a simple Palette that consists of five Color Wells. The user can store colors for future access.

The app. turned out pretty well. I found it interesting to take a break from QTKit and explore a few unfamiliar Cocoa Classes.

Notes:

When I find the time I’ll be writing about a bunch of new stuff, mainly Adobe Audition for the Mac, Final Cut Pro X, and a new media playback application that I am finishing up with interesting support for images captured with one of my favorite iPhone apps. – Panascout. Lastly, FiRe 2 … an awesome iPhone audio recorder that supports waveform editing and audio processing.

-paul.

Final Cut Pro Offline Edit with ProRes Proxy…

It has been documented that the newly released feature film “Eat Pray Love” staring Julia Roberts was edited entirely on a Final Cut Pro workstation.

I found this most interesting:

“The editors found an efficient solution in ProRes 422 (Proxy), a new version of the Apple ProRes codec introduced with Final Cut Pro 7. As soon as dailies arrived from EFILM, Assistant Editor Doc Crotzer would transcode the files from ProRes 422 (HQ) to ProRes 422 (Proxy), organize the footage into bins, and prepare the material for editing.”

Review this chart, and notice the variations in data rates of the ProRes family of codecs:

Obviously lower data rates = smaller file sizes. The bottom line is working with ProRes Proxy files (Offline copies of original ProRes 422/HQ files) creates a much more efficient workflow that is less taxing on any system.

I’ve adopted a slick method using my iMac for rough cutting ingested AVCHD footage that has been transcoded to ProRes Proxy via Final Cut Pro’s Log and Transfer. Depending on the complexity of the finished project sequence, I can finish and output on the iMac, or – move the project and it’s assets over to my MacPro for finishing. The key is prior to outputting, the edited Proxy clips can be re-captured and replaced with higher quality ProRes versions.

The Workflow:

  • Set up the FCP Project and Sequence. My sequences are typically 1080p/24.
  • Add a new Bin in the FCP Browser and designate it as the Logging Bin.
  • Mount media and run Log and Transfer. From the Import Preferences Action Popup menu, set ProRes Proxy as the ingest transcode format.
  • Log and Transfer clips into the project. Before closing Log and Transfer, reset the ingest transcode format to ProRes 422.
  • Close Log and Transfer and edit clips into sequence.

</br

Reconnect Media

  • Select the project sequence in the Browser. From the FCP menu, run Media Manager: File/Media Manager. Create an Offline Sequence and set to ProRes 422.

  • Name and save the new Sequence. A new tab will appear in the FCP Browser that includes the duplicate sequence and it’s offline clips.
  • Select the new sequence. From the FCP Menu: File/Batch Capture.
  • Choose All Offline Items in Selection from the displayed Re-Import popup menu. Log and Transfer will run and recapture the sequence clips as ProRes 422 versions.
  • Finish and output.

If you edit on an iMac, a Mac Portable with an external FW800 hard drive, or if you are looking for a more efficient large-scale project workflow – try this method. It works well for me …

-ptfigg.