![](/shared/image/trans.gif) |
![](/shared/image/trans.gif)
|
![](img/clear.gif) |
![Technologies in ATRAC3](img/technologiesinatrac3_top.gif) |
![](img/clear.gif) |
![](img/clear.gif) |
![](img/4.gif) |
![](img/clear.gif) |
An ingenious signal analysis method
|
• |
Simultaneous masking and hearing threshold in the lower audible ranges are
used for quantization noise control in the frequency domain |
• |
Forward masking and backward masking are used for quantization noise control
in the time sequence domain |
![](img/clear.gif) |
![](img/clear.gif) |
Removal of data redundancy |
• |
Quantized data are efficiently coded applying Entropy Coding |
|
![](img/clear.gif) |
![](img/clear.gif) |
![Band division filter](img/banddivisionfilter.gif) |
![](img/clear.gif) |
Optimal processing is performed on signals that change over time in each sub-band by splitting input signals into 4 sub-bands. |
![](img/clear.gif) |
![](img/clear.gif) |
![Quadrature transform](img/quadraturetransform.gif) |
![](img/clear.gif) |
Signal compression is performed with virtually no loss in sound quality by
setting the frequency resolution at 21.5Hz. In addition, a smaller area for the
transformation and frequency resolution processing is made possible by hybrid
processing using the band division filter. |
![](img/clear.gif) |
![](img/clear.gif) |
![Spectral signal coding](img/spectralsignalcoding.gif) |
![](img/clear.gif) |
1. |
Tonal signal isolation |
• |
Efficient bit allocation is performed by isolating tonal signals such as that
of trumpet or flute sounds, which have signal components that are unevenly distributed
throughout certain frequencies. |
2. |
Coding table |
• |
Simple coding is achieved with less memory by using a smaller
coding table. |
|
![](img/clear.gif) |
|
|
|
|
![](/shared/image/trans.gif) |