How to Tame Wild Storytellers

Multimedia, TV advertising and radio commercials have a common denominator: storytelling. However, sound bursts, breathing noises, and wheezing, among others, are difficult problems to correct when recording the human voice.

Fortunately, digital hard disk systems allow you to use Digital Signal Processing (DSP) to modify and solve many storytelling irregularities. Here are the techniques:

Approaching the microphone and enhancing the bass tones produces an atmosphere of intimacy, similar to that of a professional FM announcer, but also creates the conditions for p-popping: the pe that stands out in a dominant way.

Let's look at three ways to control the problem:

• Profit changes. The sound of the pe is consistent and facilitates its identification among other sounds. Select the audio region that contains a pe and reduce the gain by 50% (or more if necessary). This reduces the burst but leaves the quality of the doorbell intact.

• Changes in the EQ. Bursts are basically low-frequency energy. By defining the zone of the burst and cutting off all frequencies below 40 Hz, the resonance of the burst is eliminated. Sometimes you can remove all the bursts from the track, automatically, without removing the overall "warmth" of the narrative.

• Splice. To eliminate the burst outright, for example if it spoils a sector considered "sensual", define or mark the section of the pe (similar to the procedure already mentioned in the gain changes) and cut it so that the rest of the file fits on the splice.

If you are using Digidesign's Sound Designer program (for Macintosh), make sure that the Smoothing feature is enabled while performing any of these operations.

Compressing or limiting can strengthen the voice but, at the same time, produces a deflated and flat quality that obscures the inflections of the speaker and tends to be boring after a while. A better option is to select individual phrases (they are easy to observe, since you see spaces between them) and normalize each one to the maximum gain. This procedure will maintain a high level of average, but will leave the dynamics intact. (Warning: If the narrative is too long, normalize the words before closing the sentence spaces to decrease the overall time.) If one word stands out too much over the others in the sentence, reduce your profit to get it online before normalizing the entire sentence.

If you expected the voice of James Earl Jones, but instead got that of Bart Simpson, try to lower the overall tone of the track by 10 to 20 hundredths. (Turn on any time correction feature so that words don't lengthen but barely decrease in tone.) This procedure can give you a deeper, more authoritative timbre.

This defect is related to the bursting of pe. It can be defined as the excessive presence of high-pitched sounds that makes the that dominate the track. This defect is easy to observe: the wave on the screen has a dense and pointed appearance. Mark the region that includes the ese and decrease the profit by 50%.

Some inhalations of air before words sound natural. It is rumored, for example, that one of Barbra Streisand's engineers once removed all inhalations from the runway, only to have to put them back in since the result sounded unnatural.

Sometimes, however, inhalations can be distracting enough to overshadow the words that follow. Of course, you can use the gain change trick mentioned above for pe and ese sounds. My experience shows me that the result is even better, in most cases, if the region of inhalation is defined and a fade in is applied.

It is very tempting to remove the sounds from the background between words or phrases, simply by cutting them. Usually, the result is an uneven sound. It is better to apply a fade to silence and one from silence.

Another possibility is to remove the affected region (leaving a digital silence) and then add a little reverb. This will cover the transitions between signals, noise and silence.

Suppose you have the best storytelling imaginable, but then your customer decides to emphasize the name of the product by lengthening its duration a bit. You can troubleshoot the problem by using the time scale function.

The timescale algorithm of the Alchemy program is particularly good. Sound Designer can only process the timescale of an entire file. The best option, if you are using that program, is to copy the word into a file, process it independently, and place the new version in the original file. On a few occasions I have also changed the tone of a particular word to add more emphasis.

In conclusion, working on storytelling is another way to make the narrator look good, to create more professional commercials, to develop a better product and to get customers who come back for more after hearing how good their commercials sound.

And even though there's nothing better than getting quality storytelling out of the box, if it's about turning good storytelling into excellent storytelling, there's nothing better than hard drive editing.