Publisher review:Simulink Pro Audio Sample Rate Converter - Floating- and fixed-point models of 48 kHz to 44.1 kHz audio sample rate conversion. 1) Start R14SP2 MATLAB2) Run script run_this_firstThese demo models illustrate audio sample rate conversion of a 48 kHz (DAT sampling rate) input audio signal to a 44.1 kHz (CD sampling rate) output audio signal using a multistage multirate FIR rate conversion approach.Professional digital audio recordings are often recorded with a device such as a Digital Audio Tape (DAT) recorder at a 48 kHz sampling rate. These recordings are then edited and mixed by an audio engineer before they are ready to be transferred to a Compact Disc (CD) recording. CDs use a 44.1 kHz audio sampling rate, which means that the original mixed recording at 48 kHz needs to be resampled down to 44.1 kHz before CD mastering.There are many ways to convert a digital audio stream from 48 kHz to 44.1 kHz. The method illustrated in this demo uses three multirate polyphase FIR rate conversion filters in cascade to achieve the required overall fractional resampling ratio, while preserving the signal integrity between 0 and 20 kHz.The sequential ratios 5:7, 8:7, and 4:3 are used in the cascaded three-stage Sample Rate Conversion subsystem in the demo model. We obtain these ratios by first noting that the ratio 48:44.1 is equivalent to 160:147. Factoring the latter ratio we get (5*2*2*2*2*2):(7*7*3). Therefore we can use the three sequential ratios 5:7, 8:7, and 4:3 to achieve an overall ratio equivalent to 160:147, or 48:44.1.The original 48 kHz audio stream is first resampled at 7/5 the rate yielding a 67.2 kHz sampled audio stream. The audio stream is then resampled at 7/8 the rate yielding a 58.8 kHz sampled audio stream. Finally, the 58.8 kHz sampled audio stream is resampled at 3/4 the rate yielding the desired 44.1 kHz output.The FIR filter lengths are greatly reduced by splitting the rate conversion into three separate stages, as compared to a single rate converter. Thus, this three-stage implementation may reduce the overall computation and memory requirements. Fixed-point quantization noise may also be reduced due to the lower-order filter stages in the three-stage implementation.The all-platform versions of the demo use an audio frequency sweep source (0 - 24 kHz), and Signal Processing Blockset scope blocks to view the input (original) and output (resampled) audio signals in the time and frequency domains. The Windows PC versions of the model use a .WAV audio file source and a loudspeaker (.WAV device) sink, allowing you to listen to the resampled audio clip directly from Simulink.In the floating-point versions of the demo, single-precision floating-point data types are used. In the fixed-point versions of the demo, 16-bit word lengths are used in the filter block I/O signals, and 32-bit processing is used internally for the multiply-and-accumulate (MAC) operations in the polyphase FIR filters.During the linear phase FIR filter design, we took care to ensure an overall flat magnitude response between the 0 and 20 kHz audio frequency or passband range. We also designed the overall stopband to start at 22 kHz, or half of the 44.1 kHz output sampling rate, to minimize aliasing. Thus, the overall system frequency magnitude response resembles a unity-gain lowpass filter with transition band edges at greater than 20 kHz and less than 22 kHz. Requirements: · MATLAB Release: R14 · Signal Processing Blockset · Simulink
Simulink Pro Audio Sample Rate Converter is a Matlab script for Signal Processing scripts design by Steve Conahan.
It runs on following operating system: Windows / Linux / Mac OS / BSD / Solaris.
Simulink Pro Audio Sample Rate Converter - Floating- and fixed-point models of 48 kHz to 44.1 kHz audio sample rate conversion.
Operating system:Windows / Linux / Mac OS / BSD / Solaris
Simulink Pro Audio Sample Rate Converter script details 
