We already have software to sort out tuning problems and now Synchro Arts are providing AudioSuite tools to cure timing problems. Paul White's musical skills prove ideal to test this software to the full!
The original self-contained version of Vocalign was developed from technology invented by Wordfit Ltd in order to enable film editors automatically to match the timing of replacement dialogue with original dialogue parts. It did this job extremely well, saving actors hours of tedious 'looping', but along the way musicians and engineers discovered it could also help them with problems that they were facing every day: backing vocals that were in tune but not quite in time, harmony vocals that weren't tight enough to a lead vocal, or real musicians injecting too much 'humanity' into their timing when replacing programmed guide parts. Because of these musical applications, Synchro Arts decided to launch an AudioSuite plug-in version of Vocalign for use with Pro Tools version 5 and higher.
Stretch To Fit
Vocalign works by first analysing the energy peaks and troughs in the section of audio to be used as the template that the processed track will be made to follow (the Guide audio). An analysis of the 'out-of-t
VOCALIGN AS £347
Straightforward user interface.
Good results in the majority of real-life situations.
Only works with Pro Tools 5 and above.
More sophisticated operations require a certain amount of user skill and experience.
Vocalign AS is highly effective in ironing out normal timing discrepancies, double-tracked parts or layers of backing vocals. It is also useful for creating some special timing effects.
Before using the program, it must be installed from CD, after which it has to be authorised using the included key floppy disk. Once correctly installed, Vocalign appears in the list of AudioSuite plug-ins within Pro Tools, and to ensure a smooth learning process, a step-by-step user guide and a set of example files are included.
In The Frame
The Vocalign window shows the envelope of the Guide audio part and that of the Dub audio part, one above the other. There are five basic settings that may be used depending on the nature of the audio being treated, the most neutral being Normal Flexibility. The Maximum Compression setting uses mainly time compression in order to correct the Dub audio track, whereas Maximum Expansion uses mainly time expansion to the same end. The remaining settings are Low Flexibility, which maintains better sound quality at the expense of absolutely accurate fit, and High Flexibility, which achieves the most accurate timing, though possibly at the expense of sound quality. It's within the Vocalign window that a destination track is selected for the processed audio, though if no track is selected, the audio is written back to the Pro Tools track that is currently selected by the cursor. This can speed up operation in a busy dubbing environment as the destination track selection can effectively be controlled from within Pro Tools just by selecting it.
Processing is done in sections selected by the user within the two windows -- in the case of vocals you'd usually work a phrase at a time. In any event, the selected region must be under two minutes and over a quarter of a second in length. Note, however, that if the Dub part includes extra material at the end not reflected in the Guide region then it won't be processed, but simply left on the end of the processed audio. A useful starting strategy is to select the Maximum Compression setting if the Dub audio is much longer than the Guide or Maximum Expansion if the opposite is the case. Where you're dealing with two similar performances with normal timing errors, then the Normal Flexibility mode is generally best. Clicking the Align button causes a processed file to be created and this may be auditioned within the Vocalign window to see if it has worked out OK. If the process doesn't deliver a good result, you can simply try one of the other settings to see if it produces better results.
At A Stretch
The manual suggests that a little silence be selected at the start of each region, so that the program can work out what is real signal and what is noise. It also apparently helps to get a little more silence before the audio in the Dub region than in the Guide. Apparently these 'suggestions' become more important if the audio being dealt with is in any way tricky for the program to handle. There are also techniques which no doubt become refined with experience, such as selecting all but the last second or so of the Guide if you don't want to mess up a reverb tail at the end of the Dub part, or selecting only the first few moments of the guide track if you just want the starts of the two sections to be lined up (without further fitting taking place).
In normal musical applications, the sections of audio being processed are likely to be roughly similar in both content and timing, but Vocalign can also be used when the two parts are somewhat different in length or when the sounds themselves are different. It's fair to point out that some experience is necessary before you tackle jobs such as foreign-language dialogue replacement, as much depends on the quality of the translation and on how much fitting is needed to achieve lip sync. Fortunately, few musicians will have to deal with this type of situation, though you may want to experiment getting two different instruments to track together. Similarly, you can process a string of words and a simple drum beat to give you a rapping effect where each word is sync'ed to a drum beat.
The procedure for using Vocalign AS is actually very straightforward and the results generally work out well, providing the timing errors are reasonably small. For example, a typical double-tracked BV is no problem to tighten up and the processing artifacts are generally insignificant. Where the timing errors are larger, the audio sometimes sounds a little processed and, in extreme cases, the software finds the wrong fit and comes up with a nonsense result. In such cases you'll probably need to select shorter sections of audio to work on -- pairs of words instead of phrases, or phrases instead of verses.
Things get a bit more 'suck it and see' when you're dealing with two dissimilar sounds, as Vocalign AS seems to work on the audio envelope. If the dub audio has a different number of peaks than the guide audio, the system can trip up. Again, working on shorter sections usually gets around the problem.
Vocalign AS does for timing problems what Auto-Tune does for pitching, and in both cases the best results are achieved when the original performance isn't too far out. The user interface is straightforward, and because a new audio file is created during processing, the original audio always remains intact. Though the software was originally designed for film work, it's actually very useful in a musical context and, for routine timing correction, very little user skill is required. The sound quality of the shifted audio is really very good in most normal situations, and even when more radical shifting is needed, the results are surprisingly good. In fact the process is so useful that it's rather a shame that only Pro Tools users have access to it. Perhaps a VST version is on the horizon?