Headroom on the Medium

Another idea emerged along with that of using more speakers for low frequencies. The headroom of the magnetic oxide stripe on the 70mm film used in 1977 was the same as it was before 1960 because the last striping plant that remained was still putting on the same oxide! Since the headroom on the print was limited to 1960s performance by the old oxide, it was easy to turn down the level recorded on the film by WdB, and turn the gain back up in the cinema processor by 10dB, thus improving low-frequency headroom on the film in the boom channel, although also suffering from a greater susceptibility to hum, but that could be looked after. At first, a 250 Hz low-pass filter stripped off the higher frequencies, so hiss was no problem. Within a few films, the frequency of the low-pass filter on the boom channel was lowered to 125 Hz. This was done so that if speech was applied to the boom channel, it would not get thick-sounding, because the usual boom channel content was just the sum of the other channels low-pass filtered. Even for Star Wars, special

content for the boom channel was being recorded that did not appear in the main channels, so a whole new expressive range was exercised for high-level low-frequency sounds. In fact at Lucasfilm about 1982 Ben Burtt asked me for a "boom sting" button, a button that when pressed just upped the gain temporarily by 6dB—faster than using a fader.

The idea of having more low-frequency than mid-range head room proved to be a propitious one. Modern day psychoacoustics tells us that human listeners need more low-frequency sound pressure level to sound equally as loud as a given mid-range level.This is known from the equal loud-ness contours of hearing, often called the Fletcher-Munson curves, but modernized and published as ISO 226:2003. (The 1930s Fletcher-Munson curves show the equal loudness curve corresponding to 100dB sound pressure level at 1 kHz to be nearly flat in the bass, but all later experimenters find you need more bass to keep up with the loudness of the mid-range.) Observing these curves demonstrates the need for more low-frequency level capability than mid-range capability, so that a system will sound as though it overloads at the same perceived level versus frequency (Fig. 2-9).

Fig. 2-9 Equal loudness contours of hearing, sound pressure level in dB re 20[ilM/m² (OdB SPL is about the threshold of hearing) versus frequency in Hz, from ISO 226. These curves show that at no level is the sensation of loudness flat with frequency;

more energy is required in the bass to sound equally as loud as the mid-range.

By the way, I found out about this in a completely practical way. Jabba the Hutt's voice in Jed'i started out with a talker with a very deep voice (a then-marketing guy from Dolby, Scott Schumann) recorded on a directional microphone used up close, thus boosting the bass through the proximity effect.

Then the voice was processed by having its frequency lowered by a pitch shifter, and a sub-harmonic synthesizer was used to provide frequencies at one-half the speech fundamental, adding content between 25 and 50 Hz corresponding to the frequencies in the speech between 50 and WOHz. All of these techniques produce ever greater levels of low bass. When I looked at the finished optical sound track of a Jabba conversation, I found that he used up virtually 100% of the available area of the track. Yet Jabba's voice is not particularly loud. Thus, really low bass program content can "eat up" the available headroom of a system quickly. This lesson surely is known to engineers who mix for CDs.

Дата добавления: 2015-10-30