ISP Technologies was presented with a design challenge from one of our integrators: to improve the voice clarity and intelligibility of announcers for broadcast sporting events as heard in the sports bar environment. The fundamental challenge is that most sports bar audio suffers from a number of detrimental sonic problems that collectively mask the announcer’s dialogue, destroying the intelligibility. The typical problems include poor midrange performance of commonly installed speakers, less than desirable acoustics from highly reflective surfaces in the environment, plus the masking effect of a room full of people eating and talking with a typical SPL noise floor of 85 dB to 90 dB. This combination makes it impossible to clearly understand sporting-event audio when using common sound reinforcement systems. The spectral build up of the masking sound manifests in the lower midrange portion of the audio spectrum; simple EQ, while helpful, simply cannot overcome the problem and may actually cause clipping of the sound system, creating further problems with intelligibility. Adding to the problem, the typical installed speakers are not flat; many have a peak in their response in the upper midrange portion of the spectrum, which over-emphasizes the sibilance of the announcer and the squeak of tennis shoes on the basketball court, further denigrating sound quality. The broadcast audio also suffers from equalization enhancement that can cause overly sibilant sports announcer vocals.
To overcome these problems, ISP Technologies has developed the VIP (Vocal Intelligibility Processor). The VIP is designed to dynamically analyze the spectral energy distribution of the voice band of broadcast sporting events and dynamically enhance the formant energy and the harmonics of the broadcast vocal so as to greatly improve vocal intelligibility, clarity and definition. A second dynamic enhancement of high-frequency spectral compression, similar to de-essing, is applied to remove the sibilance of broadcast vocals and the squeak of tennis shoes on the court for basketball broadcast. The VIP requires a fixed reference level of the broadcast audio in order to properly track and will dynamically enhance the critical voice band and apply high frequency de-essing as needed. The resulting intelligibility can improve vocal clarity by up to 10 db, which can make a huge difference in the enjoyment level for sports bar customers.
ISP Technologies has been a leader in the area of real-time noise reduction and has sold hundreds of thousands of channels of Decimator noise reduction to remove unwanted noise for professional and semi-professional guitar players. ISP holds multiple signal-processing patents for the Decimator noise-reduction processes, and relied on this knowledge to develop the VIP system. Decimator noise reduction operates by dynamically removing the background noise encountered by musicians and operates on the masking principals of noise reduction, employing a combination of low-level expansion and dynamic high-frequency low-pass filtering. Life would be great if we could apply a Decimator to the background noise in the sports bar, but it would be impossible to remove the masking noise in the sports bar environment without removing the people. However, it became apparent that the same basic dynamic functions of the Decimator control could be applied in the design of the VIP to dynamically enhance the voice band. The core patents in the Decimator are based on what we term “Time Vector Processing,” allowing the Decimator noise reduction to dynamically change its time constants and track the envelope of the audio signal for controlling the response of downward expansion and dynamic filtering. In essence, we can track the input signal and change the release time constant of a downward expander and dynamic filter. The processing adapts to the audio signal (amazingly accurately) so as to maintain masking of the noise by the audio signal and avoid noise intrusion as the signal decays. By applying the patented Time Vector Processing circuit to control the response of the vocal formant enhancement and the upper harmonics, we can provide a dynamic spectral boost as required to enhance or elevate the voice band formants and harmonics in the broadcast signal. The Time Vector Processing circuit is used to control the dynamic spectral boost specifically in the critical voice band using a combination of compression and expansion. Most dynamics-based signal processors incorporate a fixed or pre-defined attack and release response. Using a variable response for attack and release of the dynamics processing circuit, we can apply much more specific enhancement of the voice, rather than a simple EQ boost in the voice band. Another critical aspect of the VIP operation is that the dynamic enhancement is above the spectral portion of the masking background noise in the sports bar. It would require improved headroom to process the 100 Hz to 400 Hz portion of the audio spectrum where the majority of the masking noise is present (the masking noise will extend far above this portion of the spectrum, but this frequency range has the largest spectral build up). By analyzing the low frequency formants in this portion of the spectrum, and providing enhancement of the harmonics above the masking spectrum, we can greatly improve the clarity and intelligibility of the voice band of the broadcast audio.
The VIP hardware system is designed with two independent channels that are inserted between the selected audio source and the power amplifiers of the system. The VIP becomes operational when broadcast sports are selected. For those interested in more detailed reading of the operation of ISP Technologies’ “Time Vector Processing,” the main patent is US patent number 7,532,730.
Buck Waller is the founder and CEO of ISP Technologies.