Spectral band replication

Spectral band replication (SBR, English: Spectral Band Replication ) is a patented method of parametric audio coding the company Coding Technologies for lossy audio data compression.

It is used to allow coding method for lossy compression of audio at low bit rates (typically 24 kbps to 64 kbps) of acceptable quality. These frequencies are above a cut-off frequency (eg 4-8 kHz ) is no longer directly encoded, but regenerated by control signals from directly encoded middle and lower signal areas by similar sounding signals.

This is made possible by two properties:

• The human ear is toward higher frequencies always less demanding. For one phase positions are no longer recognized from about 2 kHz, on the other hand the sound difference between pure tones and narrow-band noise is always lower at higher frequencies.
• The spectrum of many sounds has characteristic envelope; between the envelope in the low frequency and high frequency range there is a clear correlation. So overtones are periodically continued, the temporal envelope of medium- and high-frequency tones usually look very similar.

To synthesize the complementary high-frequency signal of the signal is usually carried out with a frequency-transformed representation. Thus, the frequency bands can be replicated in a simple manner. For this, there are various methods. In the sub-band method of synthesizing complement to signal appropriate sections of the low-frequency part dekodiererten signal is selected for different sub-bands and joined together in an adapted volume. The encoder determines the control parameters for the reconstruction process.

Similar to many other coding methods (eg intensity stereophony ) this encoding method to be used for transparent encodings shown little or hindrance, as the process of its natural synthesized independently of the corresponding portion of the original signal after signal components. The main application is in the encoding middle data rates at medium quality.

As a non -biased patents alternative method, referred to as a frequency band convolution technique, for example, at CELT set up to afford something similar and it has much lower implications for the codec latency and complexity ( computational effort ).

Examples

• Mp3PRO Audio Player, 64 kbps, conventional coding of the spectral range 0-8 kHz, SBR 8-16 kHz.
• Audio format MPEG -4 High Efficiency Advanced Audio Coding (HE -AAC, AAC , AACplus v1, v2 AACplus ), 32 kbps, conventional coding of the spectral range 0-5 kHz, SBR 5-15 kHz.
• Audio Enhanced Audio Codec 2.0