By Rudy Helm, Quality Assurance Tech, Visual Purple, LLC
With respect to design, the following scenarios could be weighed.
-Is it narrative VO? Or
-Is it dramatic dialog?
If narrative in design, consider trying a condenser microphone, and be consistent with lip distance and use a pop-filter. Ensure that VO talent’s lips are about 2 inches from the pop filter screen, and attach the screen about 6 inches from the and measure your distances with a tape measure, write the measurements down, and enforce those measurements throughout the session. Try to do the session in one sitting (talent shouldn’t eat food while recording). If not possible to do one voice actor in one session, replicate the environment exactly and measure those distances! Realize that if doing the session with the same voice talent over a span of days, barometric pressure and any other environmental variances from one session to the next can work against you.
If the audio design is multi-voice talent drama/comedy, etc. Try using a high-quality lapel type mic in addition to a suspended room mic (the lapel microphones need to be placed equally in distance on each actor, from clothing point to lips). And with each voice actor, to ascertain a proper peak level, test-record the loudest passages of the dialog first. If the character in your drama needs to shout or raise her voice, start with testing those readings.
Let me relay to you an experience that a colleague shared with me. He recorded a character (let’s call her Sue) initially with his team’s mixer set to “where it was left at”. Then with their second character (call him Bob…he was brought in as a replacement for their previous ‘Bob’), the new Bob was clipping because the old Bob was so much louder. “So I had to find some knob to bring it down”, my colleague explained. “Then I had to have him repeat his exclamations at a lower volume to avoid clipping.”
So, the lesson is, always test-record the loudest sequences first. If you can achieve a good recording that doesn’t clip (distort) on the more intense passages, you are assured that all other passages from the same voices talent will not clip. And do not use the same mixing board levels for just any other voice talent. Test-record the loud passages from each separate talent. Sometimes dynamic compression can be deployed. Compression is better used during the session itself (to help stave off clipping issues and to mitigate the ‘proximity effect’, etc) but ultimately better avoided if there are not dramatic differences between normal passages and loud passages within the dialog.
Do not mix the two microphones (lapel plus room) to mono until all talent has been
recorded. Digital recording media is cheap. Make archives. Never overwrite originals! In
this scenario, your mono mix-down is not the original (the multi-track lapel plus room
formats are considered original takes). When mixing the finals to mono, mix such that the overhead microphone is barely noticeable. A good rule of thumb is to set such a level that you almost swear that track is not even ‘on’, yet if you were to mute the overhead track, you realize that you can tell the difference when only the lapel track is playing.
Even a desk (including for instance, if a table microphone stand is being used) can bounce the sound around. It’s a good idea to use a soft table pad and place something like a cloth napkin around the microphone pedestal. If the microphone location is suspended, rather than on a table, reflections can be problematic if it’s close enough to a wall (a wall corner, even worse). Another culprit can be simply a small-ish room. Reflections bounce off plaster, glass from mirrors and hanging photos/artwork as well. Oddly, a well insulated isolation booth can still be problematic, simply because it’s a small enclosure, and sound reflects back and forth even before it gets a chance to diffuse in the absorptive or dispersive materials of a small isolation booth.
A Colleague reported that he moved the microphone from in front of his VO talent’s mouth to above it. This technique is referred to as ‘off-axis’ and is a method to avoid direct sound, which can result in a mix of environment with the intended source, depending on the microphone and its configuration.
This may not have turned out to be a method that worked well for him because his team was recording with a big monitor outside the isolation booth. This was a 2-foot tall speaker, reportedly with high volume, sitting about five feet outside of the booth on the mixer desk next to the engineer.
My colleague pointed out that two qualified engineers didn’t have a problem with it, but he was still able to detect issues with the resulting recordings. Myself, I can only respond that if the sound booth provides near 100% isolation, I wouldn’t have a problem with it either. But with only 5 feet of separation, it would have to be one hell of an iso booth…though I guess it’s possible. To which my colleague adds that one can hear a raised voice outside the booth (and you don’t have to yell). Well, I replied to him that it sounds like the monitor speaker would easily approach the dB level of a ‘raised voice’.
Hmmm…To that my colleague agreed, positing that he was pretty sure that if the voice actor wasn’t wearing headphones, which said talent would be able to hear a playback on the speaker.
I think it stands to reason that if the voice talent is in fact able to hear a playback from the monitors, then that means the microphone can also hear the speaker playback. That means it gets mixed back into the microphone (especially if the microphone is off-axis as my colleague reports). The off-axis effect (usually) isn’t as worrisome on a condenser microphone as it is on a dynamic microphone, which typically is configured for unidirectional. On the other hand, condensers are (usually) more sensitive, and therefore keen to pick up external sources.
Let’s conclude with a comment on the end-user’s sound system…
Given the popularity of subwoofers and better-than-multimedia quality speaker systems attached to modern PCs, it probably is a reasonable consideration to roll off frequencies below 100 Hz on certain types of audio, especially VO and ambience loops, lest you are prepared for unnatural representation of those kinds of sounds. Obviously this approach might not be considered for high impact noises, such as car crashes, weapons, explosions, etc, where the subwoofers are best at their job.
By the way, I was happy to report that most of my friend’s sound issues have been resolved with his team’s proactive efforts to avert future problems at the recording level. And they have gone far at making their projects exude that professional sonic ‘sheen’.