Learn about DSP audio processing

Aguilera

Learn about DSP audio processing [Copy link]

Most audio processing is still done in the analog domain today, because early digital processing solutions based on general-purpose DSPs and external analog-to-digital converters (ADCs) and digital-to-analog converters (DACs) add significant additional hardware and software programming costs. As a result, implementing such solutions is difficult, time-consuming, and costly. Solutions are now available that integrate an audio-specific DSP and high-performance audio data converters on a single integrated circuit. These provide professional-quality digital sound processing with 112dB signal-to-noise ratio (SNR), full graphical user interface development and programming tools, and a good price-performance ratio, allowing traditional analog systems to adopt digital technology with superior sound quality. The AD1954SigmaDSP is an example of a complete 26-bit, single-chip, 3-channel digital audio playback system with built-in DSP functions. Its main features include: 3 digital audio channels; a 7-band 48-bit stereo equalizer; delay for loudspeaker position adjustment; Phat Stereo spatial enhancement module; and a dual-band professional-quality dynamics processor. 1. Audio Dedicated DSP Core A DSP core optimized for audio processing requirements. This user-configured DSP core has significant advantages over general-purpose DSP cores because it provides many features, such as hardware accelerators for double-precision filter calculations and dynamic processing. These features can significantly reduce the number of instruction cycles required for a given audio algorithm. This DSP core is based on a 26x22 multiply-accumulate engine with dual 48-bit accumulators. When the input word length is 24 bits, the internal resolution of the core is 26 bits according to the 3.23 format (3-bit exponent and 23-bit mantissa). Many audio algorithms require +12dB gain, and the additional 2 bits provide up to +12dB of gain, ensuring that no additional gain is required in most applications. All filters are calculated with 48-bit double-precision resolution using dedicated hardware accelerators. Double-precision operation ensures that low-frequency infinite impulse response (IIR) filters can work correctly and avoid the finite cycle problem, which would otherwise produce artifacts. Graphical User Interface Graphical user interfaces (GUIs) make it easy for experienced digital engineers and analog engineers who are familiar with their own audio systems but do not want to get bogged down in low-level DSP programming with "bits and bytes" to add DSP to their systems. This tool not only allows intuitive operation, but also allows real-time control of the entire signal flow. It shows the signal flow graphically, so it is really intuitive to use. Designers can directly access and modify every parameter in the signal chain, including filter coefficients, volume settings, and dynamic processing functions. The GUI can be connected to the evaluation board through the PC's printer port. In this way, any parameter changes can be sent through the SPI serial port and take effect immediately. Design user configuration programs according to needs Even if using internal programs is the simplest and saves the most design time, designers may still want to configure the signal flow specifically for their system needs. The Graphic Compiler is a fully graphical program development tool. The graphical input tool allows system design engineers to draw the signal flow configured by the user and compile it into a DSP program at the push of a button. There is no need for line-by-line code development, making this workflow particularly friendly to analog circuit engineers. The DSP is used as a processing engine for 3-channel speakers. The processing process includes: total equalization module, 3-channel crossover module, single driver equalization module, each single driver dynamic range control module, delay module for driver position adjustment (phase correction), and DAC with 112dBSNR. Fourth, the importance of professional quality dynamic processing Small and medium-sized audio systems are often limited by the power of their amplifiers and speakers. In addition, due to the small size of the speakers, the frequency response of the low-frequency speakers often has a premature natural roll-off at lower frequencies. Therefore, it is very popular to use a fairly strong equalization, especially in the bass area (bass boost circuit) to compensate for this imperfect sound setting. Finally, it is often desirable (if not required) for a system to have a high maximum volume. The combination of a system's limited amplifier power, heavy bass equalization, and high overall system loudness quickly saturates the amplifier and begins to produce heavy distortion, which can be frustrating or even annoying. Several previous attempts to address this problem have used simple clipping detectors that avoid clipping distortion, but the resulting artifacts are just as undesirable as clipping distortion. However, using the AD1954 SigmaDSP professional-quality, dual-band dynamics processor, the limitations of the system can be controlled without producing artifacts. V. Improving System Clarity and LoudnessOne transfer function without any dynamics processing and one with a compressor and limiter function with adjustable breakpoints. Because of the use of dynamics processing, the natural clipping level can be handled without distortion in the high volume range. This effectively allows the user to increase the system volume by about 10dB. A 10dB increase in volume means that the sound pressure level has doubled, so the user can increase the loudness of the system to twice the original level. This requires adjustments to the real-world factors that affect sound. The transfer function of the user-configured DSP dynamic processor can be adjusted arbitrarily. It can combine several dynamic processing functions into a functional curve. It has four typical functions, including: compression, limiting, expansion and noise threshold. Since this transfer function is fully programmable, these functions are very easy to implement and can be used individually or in combination. VI. Conclusion The introduction of audio-specific DSP technology has brought the embedded design of audio systems into a new era. The processing performance, transformation technology and complex algorithms in the digital field all appear in graphical form, making their application more convenient and economical. This DSP technology enables design engineers to quickly develop or transplant their systems to the digital field, so they can give full play to the high quality of digital media. World Electronic Components