اساسيات دمج الأصوات

Crashing swords, destructive lasers, intimidating helicopters, whispered words of love, jubilant screams of joy, and wheels that go round and round are all part of any good story in today’s visual medium. But so many films seem to overlook this most important aspect of a production.

Why?

I believe it is due in great part, to a lack of understanding on the part of many independent producers and filmmakers with regard to audio.

Audio is 70% of what the audience sees, and what they see drives the storyline. Dialog that is lost in noise, ambiences, or just plain poor recording, causes the audience to easily become disinterested and find distractions from what may be an otherwise excellent picture and story.

If the audience cannot comprehend dialog, they cannot follow the storyline; it’s as simple as that.

Building upon this premise, dialog should be the most important aspect of any project. Always. If the line is worth speaking and occupying screen time, it’s worth taking the time to be sure the audience can comprehend the words.

Comprehension is different than merely hearing. Being able to hear the words and know how they relate to the storyline is different than the audience having to strain to hear the words as the scene flies by. Viewers become frustrated with a story if the dialog they interpret in their head doesn’t match the imagery on the screen.

Getting dialog “right” begins long before the post-process begins; it starts in pre-production and planning for the shoot, and finishes in production with dialog properly recorded. From there, it’s only a matter of cleaning up the dialog and “fitting it” into its place in the video mix.

Recording proper levels is the most important aspect of getting good dialog. In today’s digital world, old-schoolers are too caught up in the concept of average recording levels. Averages are important, peaks are more important. It’s all a game of numbers. If you record audio too quietly, you’re leaving numbers on the floor that can never be recovered. It’s similar to taking a low resolution photograph in a dark room and then expecting to pump it up in resolution and quality in Photoshop. It can be done, but not very effectively. You’re better off shooting with lighting and high resolution camera in the first place.

With audio, you want to record as close to the subject as possible, and at the greatest level possible without reaching 0dBFS (Zero decibels Full Scale). Recording low levels and then normalizing in post isn’t the same as recording properly in the first place.

Practice recording dialog cleanly and clearly using either lavalier mics or boom-mounted mics. Practice, so that when the time comes to record for a project or client, you’ll know exactly what to do.

~Don’t use on-camera microphones for any purpose outside of reference audio or wild audio in a large crowd.

~Don’t put higher quality microphones on the camera either. Microphones belong near the source, not near the photographer.

~Good sound stems from good proximity. The closer a microphone to the subject, the better the audio quality will be. A $5.00 microphone five inches from the subject will sound better than a $5,000.00 microphone placed five feet from the subject in almost any recording situation.

~Don’t run high impedance cables over long distances. Use low impedance, balanced lines in all possible situations. If your camcorder does not allow for low impedance, balanced inputs, look at tools such as those from Beachtek, Studio Devices, and others to provide low impedance input to your camcorder.

A typical mix should incorporate dialog, music, and effects. Some NLE systems allow for bussing of audio so that all three components may be processed as individual units, but whether your NLE allows for this or not doesn’t really matter, it merely makes the work easier to mix, package, and deliver.

Audio is made of frequencies, and the human ear can detect frequencies in the 20Hz to 20KiloHertz range. With a reasonably broad set of frequencies for the ear to listen to, mixing audio should really be easy, right? Not quite. Most sounds that are within specific frequencies ranges and combinations of frequencies, overlap other sounds that contain identical frequencies or harmonics of those frequencies, which can make it difficult to discern one sound from another.

For example, the “kiss, kiss, kiss” sound of a slightly open and then closed high-hat in a musical mix can easily render the higher frequencies of a human voice unintelligible. A kick drum can easily do the same to a male voice, and a piano’s numerous frequencies and multiple harmonics may often be very challenging to mix with the human voice. Getting a grip on all these various frequencies can easily present a significant challenge.

The typical response of the less-experienced producer is to use volume/levels as a means of mixing ambiences, music, dialog, and special FX into the project, but in real practice, this often doesn’t work. When dialog becomes quiet, the common response is to raise the volume of music by a significant amount, while also raising ambience and perhaps effects depending on how the audio is routed. A better means of going about this, is to find the sweet spot frequencies for each element in the mix as it relates to other elements of the mix. For example, although the human voice occupies a very broad frequency range, does the audience really need to hear all of the frequency content in order to comprehend the dialog completely? Of course not. The very high frequencies and the very low frequencies can be removed entirely from the mix, freeing space and energy for other elements of the mix.

There are of course, other techniques, but creating donuts or holes in which the dialog can float, is a simple and fast technique for making space in a mix.

 

In this screen shot, compression has been applied to the voiceover, set to a compression value of 2:1, starting at -18dB

Voice/dialog should rarely be compressed at more than 4:1 with around -3 to -6dB of reduction, although this value of course, can change depending on the recorded levels of dialog.

Compression may also be applied to the master output channel of the mix in most NLE or DAW systems. This is a great means of achieving two goals at once; levels may be controlled, while maximizing levels for optimal output. iZotope and WAVES both offer excellent mastering/maximizing plug in processing tools that support both Apple and PC platforms.

Still yet another technique is to pan the music slightly to one side or the other, and pan the dialog by a similar amount to the opposing side of the stereo spectrum. The sample file linked above pans the bed 15% to the left, and the dialog 12% to the right.

As always, use your own ears to determine whether your envelopes, compression, or EQ are providing benefit to your audio vs damaging the intelligibility of the audio. Learn to compress various sources to gain experience. Acquiring great audio chops is most easily learned in practice sessions and not during times of pressure to finish a mix.
 

Some additional tips:

~Be sure your monitors are as flat as possible. Low grade, plastic computer monitor speakers are not optimal for mixing/finishing audio. Do not attempt to mix with headphones. Headphones are more like having a magnifying glass or microscope, they're great for drilling in to see the details or general monitoring when you're not seriously considering the actual mix. They're acceptable for working with in noisy areas or areas where you cannot make noise, but you'll need to find a time when you can make all the noise you need to in order to properly set the mix.

~Proper level of monitors is a must. Trying to mix with the volume turned low is virtually impossible. At the same time, listening to a mix at very loud levels is a bad idea as well. Periodically check your mixes at very low levels, and periodically, pump up the volume. Both extremes can expose glaring errors not normally heard.

~Isolate your monitors from surrounding objects/mass. Tools such as the Auralex MoPads are inexpensive and exceptionally effective.

~Mix with fresh ears. Don’t let a deadline force you to mix late at night after a long day of listening to audio. Walk away from mixes for periods of time, listening to other audio sources to keep ears fresh.

~Use quality tools when possible. Although most NLE’s/DAW’s include compression, EQ, and other tools, third party tools such as those from iZotope, WAVES, dBWorks, WaveArts, Digitalfishphones, and other tools are terrific to add to your arsenal of great plug ins. Both VST and Direct X plugs are available for virtually all NLE systems. Find a list of a few favorite plugins here.

~Learn what to listen for in your mixes. Listen to top-level films in the same genre as your work, and strive to mix on par with those films.

Gaining ground with great audio will significantly improve the quality of your video project whether you’re delivering to Youtube or the big screen. Better audio means your audience will better relate to your message, and stay focused on your visuals.

Listen up, because your audience surely is!