The addition of non-destructive region (clip) gain would be a powerful and valuable addition to the editing interface of Pro Tools.
BackgroundSound editors are often required to adjust the level (volume) of a particular region. For musicians this is fairly easy, because most songs are produced with a one-track-for-one-instrument work flow - simply push the fader up or down, or adjust the input gain on an equaliser.
However, in the world of film, television and radio, many regions from different sources may need to be combined on a smaller number of tracks, according to the sound's function in a mix (dialogue, atmospheres, effects, foley, etc).
Working with correct Gain Structure.
Gain structure is an important part of any good recording, editing, and mixing process. There is an enormous amount of headroom in the Pro Tools mixing system, which can be effectively realised by the use of master faders in a mixer. [1]
Just as gain structure was important in analogue studios to find the right balance between distortion and noise, appropriate gain structure is important to ensure that digital audio signals are processed (eq, compression, reverb, etc) without excessive distortion or noise.
The K-System, designed by Bob Katz [2] [3], is a very effective way to measure to align digital signals with analogue signals, and with "loudness". Editing in a K-20 production environment is very appropriate for work in film and television. However, editors will find that many of the 16-bit sounds that are sourced from mastered CDs (music and sound effects) will sound too loud and will require attenuation. Region specific level adjustments will greatly aid the combination of these existing recordings with the current sessions configuration (K-20, for example).
Current Work FlowVolume automation is a commonly used tool to adjust the level of adjacent sounds in a session. This method is very simple, and quick to use. Press the "-" key to switch to volume automation, and then use the trim tool to adjust the volume up or down.
This method has a number of distinct disadvantages, however:
- Distortion Can Occur Pre-Fader.
Volume automation is applied at the output of a track's channel strip. There are many processes that are applied to an audio signal before the channel strip's output fader, including: edits, fades and cross-fades, and plug-in inserts. The use of volume automation to adjust volume of a specific region is too late in the signal path to provide level adjustments during any of those earlier processes. This is discussed in more detail below.
- Volume automation is not preserved in region groups.
Complex edits of many regions that have volume adjustments to them are effectively useless when exported as region groups because the volume balance between the individual regions (components of the greater sound effect, for example) are not exported.
- Sudden volume automation changes are not mixer friendly.
When using a control surface, these square volume automation changes cause faders to suddenly jump up and down whilst playing a session. Making slight changes to volume automation is very difficult when the fader keeps jumping suddenly up and down for each region on a timeline. (Sure there is trim mode on HD systems, but trim mode is cumulative, so each successive pass needs to compensate for accumulated volume changes made by all previous trim passes.)
Pre-Fader Distortion
Cross-fades, AudioSuite processes, and plug-in inserts all occur pre-fader. Using volume automation as a method of adjusting the level of a specific region cannot prevent distortion, or improve gain structure during any of these processes.
Editing Automation Problem With Cross-FadesConsider this example: An editor has placed two regions on a single track, one after the other. The first region is attenuated by 5dB, using the volume automation method discussed above, and the second region is attenuate by 10dB. As the regions are moved from left to right and their handles trimmed, the volume automation is automatically adjusted to follow those regions.
Now suppose the editor wants to fade in the first region cross-fade to the second region and then fade out the second region. The fades are applied to the regions. There is an immediate problem with the volume automation jumping suddenly from -5dB to -10dB half way during the cross-fade. The volume automation can be edited to be a diagonal line from -5 to -10dB from the beginning to the end of the cross-fade, but this is a manual and tedious process. If the fades need to adjusted at a later point in time, so must the the volume automation.
In addition, the level adjustments are not applied to each region individually, which means that asymmetrical or customised fades do not preserve the level adjustments for each region during the cross-fade. The level change is applied to the output of the cross-fade, which may not create the desired result.
This problem also presents itself in the same way when using the plug-in automation method mention below.
Levels and Distortion During Cross-FadesFor the majority of cross-fades, an equal power fade is a very good starting point. This fade has a property of maintaining equal loudness (power) from different recordings. The mid-point of this fade has 3dB attenuation of both outgoing and incoming regions.
Mathematically, a cross-fade is the sum of a fade out and a fade in. The sum of two identical samples causes a 6dB increase in level at that point in time. Therefore, it is theoretically possible for the center of a cross-fade to peak at +3dBFs. Since this value cannot be recorded to disk in a 16 or 24 bit file, clipping occurs.
The only current way to prevent this distortion is to apply gain reduction using an AudioSuite plug-in to the region on the track before the cross-fade is applied. This introduces a more problems:
- No handles in AudioSuite rendered files - limits further editing.
- Wasted disk space - the gain reduced file is written to disk.
Plug-in Insert HeadroomPlug-ins generally have no headroom above 0dBFs. Some plug-ins have internal headroom that can be utilised with input and output gain controls, but the signal path between plug-ins, and from the final insert to the volume fader remains at 24-bit in a TDM system. (I'm not suggesting this needs to be increased - 144dB dynamic range is more than adequate.)
There are many plug-ins that can cause a significant increase in peak levels. (For example, a simple high-pass-filter). To avoid distortion in these processes, it is a common feature of many plug-ins to have an input gain control. However, these gain controls are typically global (constant through the duration of a session), and do not fit the requirement of being able to adjust the level of a specific region.
Using Plug-in Automation as an Alternative to Volume AutomationAn alternative to using volume automation, is to automate an RTAS "Trim" or "Time Adjuster" plug-in as the first insert on a track. This alleviates the problem of jumping faders, and allows for gain adjustments to be made before other inserts.
However, the user has to sacrifice one insert per track to achieve this, and a similarly configured plugin (with automation enabled) must be instantiated on every track that this region group may need to be used.
Whilst this may seem to solve a number of the problems highlighted above, this method does not provide any support for level adjustments during cross-fades, and even further complicates the problems of maintaining region level in exported region groups.
Automation Is Not Exported In Region Groups
Since no form of volume automation is exported in a Region Group, neither the volume automation nor the plug-in automation method can maintain the balance between individual regions in the design of a complex sequence of regions (for a complex sound effect, for example).
This significantly reduces the usefulness of Region Groups, because every time one is imported from disk, or even from the region list, no automation is present, and the balance between the components of the group is not as it was in its original form. Re-using these region groups without support for volume automation is difficult and laborious.
Using AudioSuite Gain Instead Of Automation
Another alternative is to use the AudioSuite Gain process to adjust the level of specific regions.
Using this method provides a solution to all of the problems above, except for these limitations:
- AudioSuite files have no handles.
- AudioSuite files are rendered to disk - this is an inefficient use of disk space for a simple level adjustment.
- AudioSuite gain adjustments would not be adjustable except by further AudioSuite processing. A volume automation curve can be altered, but once AudioSuite gain is applied, that amount of gain cannot be easily adjusted.
Keeping Up With The IndustryLet's face it, Nuendo has had non-destructive clip-based gain for some time now, and this is a real pleasure to work with.
ConclusionThe addition of non-destructive region-specific gain would be a valuable feature in a wide number of situations. Its incorporation into Pro Tools as an editing tool would directly benefit signal quality by allowing better control of gain structure right from the edit window; and in crease productivity for editors and mixers alike by alleviating the need to manually use automation to overcome differences in levels between regions.
References1.
"Mixing In The Box", by Stan Coney
2.
"Level Practises Part 1", by Bob Katz3.
"Level Practises Part 2 includes the K system", by Bob Katz
