24/96 is the de facto standard in the latest equipment for digital audio production. And to complicate matters, while any modern home camcorder offers 48Khz 16-bit sound, most movie theaters in our region offer a sound quality that could hardly be assimilated to that of 32Khz digital audio. In addition, some Home Teather systems offer D/A converters with sample rates that reach 192Khz at 24 bits... and paradoxically many consumers use them to listen to 16/44 mp3 files encoded at 128 kilobits per second. Where does this tangle of numbers take us?
A trained ear can feel the difference between digital audio with 48Khz sampling and that of 96Khz, especially if the tests are done in a sound studio with optimal acoustic design and high-quality monitoring. The obvious question for many of us is whether the average consumer finds any value in enhanced digital audio, which supposedly offers a more natural experience.
Numbers you hear
To try to answer this question let's briefly remember how digital audio is generated. The obligatory reference is the waveform of the original sound, which usually comes from analog sources - almost always from a battery of microphones. Once the sound is converted into electric current with very rapid voltage variations that approximately reproduce the sound waves, it is possible to implement a quantization process - a conversion of the signal into digital values.
The quantization of the audio signal is achieved by taking samples. The higher the sample rate, the more accurate the description of the original waveform. And the more accurate the digital description of the original waveform, the higher the quality of the recorded sound. The sample rate is a value that describes the temporal resolution of digital audio, then it is similar to the value of frames per second in a video signal. Put another way, the higher the sample rate, the better the frequency variations of the original audio signal will be transcribed. In the language of analog audio we speak of precision in the recording of the timbre, of the tonal variations.
Now, each sample uses a fixed number of bits to describe the state of the waveform. In digital audio slang this number of bits is known as "word". The size of the word that makes up each sample is equivalent – in some ways – to the number of pixels that are used to describe a video frame. More bits equals more detail recording. In video this represents a more defined image. In audio, the thing is a little more complex.
What advantages do we get by investing more bits in our digital audio streams? In principle higher resolution. This is a difficult thing to understand, because in many cases our ears, essentially fatigued by constant exposure to amplified sound, cannot discriminate enough to appreciate the effects of a variation in sound resolution. We have established that more resolution equals more detail. A 24-bit digital word carries more information about the waveform than a 16-bit word. The bits in each sample are used to describe changes in the amplitude of the waveform, which has led many sound engineers to confuse higher resolution with higher volume.
The Fallacy of Volume
Let us clarify this matter. And to begin to clarify, let's increase the confusion. From the purist's point of view, the music industry has accustomed consumer ears to enjoy recordings with excessively high volume levels. Taking advantage of the flexibility of 16/44 digital sound on compact discs, it has become customary to keep the level of digital audio signals as close to the ceiling as possible. To verify this, it is enough to check the output of a commercial CD in a PPM meter. Usually the "peaks" are coming just before the zero of the scale. Apparently the goal of many sound engineers is to achieve the highest possible level before saturating the signal – which in a digital environment represents clearly audible "pops" that should be avoided at all costs.
In the same way that we say that an increase in the sample rate allows us to obtain greater tonal accuracy in digital audio, we can affirm that the use of longer words for each sample gives us greater dynamic range. This means that the more bits are used, the greater the volume variations. The increase in resolution corresponds to more steps on the PPM scale, greater accuracy in the transcription of the amplitude changes of the waveform.
In the real world this is equivalent to acquiring a larger headroom. The headroom is one of the most desired features in audio equipment: It is the ability of equipment and supports to handle higher signal levels without reaching saturation levels. For the sound engineer it is good to have more space in the high levels because this way it is possible to accommodate peaks or transients with a very high level without generating audible distortions in the sound.
Having more headroom is exactly what popular music experts need to be able to operate at the highest levels possible. And here we find one of the biggest paradoxes in the world of digital audio: Taking the averages of signal levels to higher levels wastes the increase in resolution of digital audio. Recordings are usually made in which the signals are compressed to increase the volume, and this represents suppressing the transients that could be recorded thanks to the increase in the sample rate and the size of the word in each sample.
Compressing the signals at the top of the scale gives up "downward" level variations – which is equivalent to wasting the dynamic range that comes from using more bits. What the music industry is doing with popular music is exactly the same thing that was indispensable to do in audiocassette recordings in order to hide the high level of noise inherent in 1/8" tape recording: To raise the volume mercilessly.
This vice has moved to the world of video and television. Generally, the additional headroom provided by video recorders that offer 20/48 audio applies to increasing the volume, not to qualitatively improving the audio by taking advantage of the greater dynamic range. Each channel tries to "sound louder" than its competitors, and this is reflected in a higher volume... in less detail. And the problem is compounded because in general the signals are compressed twice: first when mastering and then in the transmission chain. And the issue is compounded because in some cases a digital production plant ends up delivering signals that are definitely too strong for an analog transmission network.
Considering that more and more viewers are using theater-at-home systems with reasonably decent amplification quality, perhaps this is the right time to start worrying about this issue. In fact, this may be one of the reasons why the ordinary consumer tends to perceive the compressed sound of DVD-Video as a better quality product than the sound "of the air". Attention: we are talking about digitally compressed sound to occupy less storage space, which means an appreciable loss of quality, not compressed signals to reach a constant volume. Why does mp2/mp3 audio from a DVD movie sound better than audio from free-to-air TV?
One of the reasons may be that DVD mixes are made specifically to sound good on a Home Teather, while TV has to make compromises to sound reasonably good on anything from an old 12" TV to the more sophisticated amplification equipment – to the cheap TVs that adorn most homes in Latin America. which in most cases are unable to reproduce sounds with a frequency greater than 8Khz.
Using each bit better
Quantized 24-bit audio can have greater detail or a higher volume, and audio sampled at 96Khz achieves a very accurate description of frequencies inaudible to the human ear. Is it worth increasing the sample rate to record something that can't be heard? According to equipment manufacturers, increasing the sample rate significantly improves the perception of space in surround sound recordings. According to them, by combining different signals in a controlled environment it is possible to faithfully reproduce the real listening experience and radically improve the perception of directionality of sound.
Now, what is the advantage of using 24/96 audio in video and TV? 24/96 digital audio offers a frequency response that can reach 24Khz and a dynamic range that can be estimated at more than 100DB. However, under normal conditions its apparent quality is not necessarily superior to that of the 48/16 audio of most DVD-Video titles available on the market. And definitely most installed TVs are not going to allow the viewer to appreciate these improvements in sound quality. Why worry then? In light of the bitter reality of the monophonic world of TV in Latin America, everything seems to indicate that it is best to continue using the headroom of digital audio technology to increase the volume in our old analog transmission systems...
However, the specter of DVD-Video, Super AudioCD and DVD-Audio threatens our industry in some way, as it will eventually contribute to raising consumer expectations for TV sound quality. It is true that the possibility of digitalization of transmission could give a respite to current operations. But it really is a problem of valuation of the product by the consumer, and the truth is that current engineering practices would lead us to offer a sound of inferior quality in the digital platforms that are getting closer. In a while we could be facing a competition between free-to-air TV and pay-TV platforms that begin to transport digital signals. Can our industry try to reach the level of new standards that the music industry is looking for? A good way to start can be to turn down the volume a little.
