Upmix & Downmix

Upmixing or downmixing within File-Based workflows

As we continue to try and offer an improved user experience, technology keeps changing. Sometimes this causes problems for broadcasters, and a good example is the move to surround audio. Ideally all content is produced in both stereo and surround sound versions so that the customer can select the appropriate source for their viewing/listening environment. But older content is probably only available in stereo, yet the surround must still be provided. If you have access to the original master audio, and can afford edit suite time, you can create the surround audio in high quality. Usually though, you don’t have access to the master, and only have the released stereo soundtrack. In any case, apart from particularly high value material, time in edit is likely to be prohibitively expensive, meaning another solution is required. Enter Engine with our Up/Down mix modules.

Engine supports stereo to 5.1 upmixing, and 5.1 to stereo downmixing, plus stereo to mono downmix. Like other options, upmix and downmix can be added flexibly at any point in a workflow.

Upmix in particular is a complex process, and everyone will have heard bad examples of automated upmix, which illustrates the need to choose your technology carefully.

Engine incorporates world leading Upmix and Downmix technology developed by Soundfield. Soundfield Upmix technology is reviewed time and time again as a leading example of high quality audio processing, repeatedly giving consistent results with a variety of program material. The Soundfield algorithm was also carefully developed to provide excellent downmix compatibility, should this be required in the consumers environment.

The Upmix technology uses a unique algorithm which extracts the ambience from the direct sound and differentiates itself from other approaches to upmixing through its adaptive nature. The detailed analysis of the stereo input material will determine how the upmix algorithm behaves at that instance in time.

This technology produces a very stable and natural sounding 5.1 without destroying the original stereo image. Keys to achieving this are:

Good extraction of mono sources to feed the centre channel – such as dialogue, commentary, etc.
Maintaining the frontal stereo image by keeping the direct sound sources of the stereo mix at the front.
Only the extracted natural ambience is fed to the rear surround sound channels

When creating a workflow including Upmix, you choose the channels containing the source stereo, then choose which channels will hold the created upmix. By default the channels use the standard order shown in this diagram, although the ‘Advanced’ control lets you configure any order. At last – no more difficulties with SMPTE and Film channel ordering!

The Upmix profile is editable, with a list of standard presets being supplied, plus the ability to create new settings.

Note that after Upmix, the overall level of the audio may have changed. This can conveniently be fixed by using a Loudness step directly after Upmix.

The Downmix Algorithm

The SoundField downmix algorithm offers the standard Lo/Ro (Left only and Right only) and Lt/Rt (Left total and Right total) downmix options but also offers some unique features such as adaptive EQ and independent direct and ambient downmix levels for the surround channels. The Downmix algorithm can be used as a reference to verify the quality of your stereo downmix in both Lo/Ro and Lt/Rt or as a high quality adaptive downmixer to create a separate stereo transmission feed.

There are two features unique to the SoundField downmix algorithm, the Adaptive EQ and the Ambience Mode.

The Adaptive EQ is a gentle equalizer that corrects for frequency colouration that may occur due to phase related downmix artefacts. Although you would only perceive tonal difference in the most extreme of cases the effect is far more pronounced when listening to the resulting Stereo image. These frequency colourations tend to make the Stereo image narrower – tending towards mono in the most extreme of cases – the adaptive EQ will combat this narrowing effect and returns the Stereo back to its intended Stereo image width.

The Ambience Mode is a unique and very effective way of combatting ambience build up in a downmix, the biggest challenge faced in downmixing. In a standard downmix the only control to combat ambience build up is to alter the Ls/Rs downmix level – typically -3, -6 or -9dB – although this can help it will also reduce by the same level any relevant direct sounds from surround channels. The ambience mode uses the technology developed for the upmix algorithm to split the surround channels in its Direct sound and Ambient sound components, making it possible to set individual downmix levels for the Ls/Rs Direct sound and Ls/Rs Ambient sound components.

The workflow using the downmix option is straight forwards. Choose the six channels containing the 5.1, and select the destination for the newly created stereo. This can replace existing channels within the original source file structure, or can be routed to a new stereo WAV file.