|
Common description
Robust MELP vocoder with rate 4800 bit/s (RMELP 4800
bps) is intended for using in discrete channels with bit error rates (BER)
up to 7-9%.
Robust MELP vocoder is built on the base of low bit rate
MELP vocoder and original channel coding scheme using codes with high
bit error correction capability.
Low bit rate MELP (Mixed Excitation Linear Prediction)
speech coders use classical parametric model of description of speech
signals. This model is implemented on the base of linear prediction filter
with mixed excitation in several frequency bands.
Independent estimation of voiced/unvoiced decision in
different frequency bands is providing more natural synthetic speech sounding
then quality of classical LPC vocoders with binary voiced/unvoiced decision
for single band.
RMELP 4800 bps is implemented as fixed-point ANSI-C models
for PC. It also has real-time implementations on the base of TMS320c54xx
DSP families.
DSP implementations have following features:
- C call interface for top level procedures
- real time work
- multichannel support - independent processing of several duplex channels
on single DSP
- code compatibility within DSP family
- high optimization of program code - all internal signal processing
procedures are implemented on assembler language
Parameters of RMELP 4800 bps TMS320C54xx implementation:
| |
Coder |
Decoder |
DSP peak load per channel, MIPS |
22.3 |
18.3 |
Program and data memory, K words |
23 |
N – number of channels
Parameters of RMELP 4800 bps TMS320C55xx implementation will be sent
on demand. Please ask.
Here you can test speech quality of following speech samples of RMELP
4800 bps vocoder with different channel bit error rates (BER) with or
without error bit grouping. (In case of modeling of error bit grouping
we use 2 state Hilbert model with BER probability and PGB probability
that mean probability of transition channel from "good" state to "bad"
state and vice versa.)
Original english speech |
|
RMELP vocoder 4800 bps, BER = 0% |
|
RMELP vocoder 4800 bps, BER = 3% |
|
RMELP vocoder 4800 bps, BER = 5% |
|
RMELP vocoder 4800 bps, BER = 7% |
|
RMELP vocoder 4800 bps, BER = 9% |
|
RMELP vocoder 4800 bps, BER = 5%, Pgb = 1% |
|
RMELP vocoder 4800 bps, BER = 5%, Pgb = 3% |
|
RMELP vocoder 4800 bps, BER = 7%, Pgb = 3% |
|
RMELP vocoder 4800 bps, BER = 7%, Pgb = 5% |
|
RMELP vocoder 4800 bps, BER = 9%, Pgb = 5% |
|
RMELP vocoder 4800 bps, BER = 9%, Pgb = 7% |
|
For example here you can listen the standard USA FS-1016 CELP 4800
bps vocoder in the same channel conditions:
Original english speech |
|
CELP vocoder 4800 bps, BER = 0% |
|
CELP vocoder 4800 bps, BER = 3% |
|
CELP vocoder 4800 bps, BER = 5% |
|
CELP vocoder 4800 bps, BER = 7% |
|
CELP vocoder 4800 bps, BER = 9% |
|
CELP vocoder 4800 bps, BER = 5%, Pgb = 1% |
|
CELP vocoder 4800 bps, BER = 5%, Pgb = 3% |
|
CELP vocoder 4800 bps, BER = 7%, Pgb = 3% |
|
CELP vocoder 4800 bps, BER = 7%, Pgb = 5% |
|
CELP vocoder 4800 bps, BER = 9%, Pgb = 5% |
|
CELP vocoder 4800 bps, BER = 9%, Pgb = 7% |
|
|
 Common description
Implementation
Speech samples
All technical and commercial questions about products, prices,
terms of delivery, software libraries, source code etc. please send us by e-mail:
|
