|ML2216||ADPCM2 Algorithm-Based Speech Synthesis LSI|
Oki Electric Industry
|ML2216 Datasheet PDF : 19 Pages |
Memory Allocation and Creating Voice Data
The ROM is partitioned into 4 data areas: voice (i.e., phrase) control area, test area, voice area, and phrase control
The voice control area manages the ROM’s voice data. It controls the start/end addresses of voice data, usage/not
usage of the phrase control table function and so on. The voice control area stores voice control data for 256 phrases.
The test area stores the data for testing.
The voice area stores the actual waveform data.
The phrase control table area stores data for effective use of voice data. As for the details, please refer to the Phrase
Control Table Function.
There is no phrase control table area if the phrase control table is not used.
The ROM data is created using a development tool.
ROM Addresses (ML2216)
Voice control area
(16 Kbit Fixed)
Phrase Control Table area
Depends on creation of ROM
Playback Time and Memory Capacity
The playback time depends upon the memory capacity, sampling frequency, and playback method.
The equation showing the relationship is given below.
1,024 × (Memory capacity – 16) (bit)
Playback time [sec] = Sampling frequency (kHz) × Bit length
(Bit length is ADPCM, ADPCM 2 = 4 bits; PCM = 8 bits.)
Example: Let the sampling frequency be 16 kHz and 4-bit ADPCM algorithm. If one 8 Mbits ROM is used, then the
playback time is obtained as follows:
1,024 × (8 × 1,024 – 16) (bit)
Playback time =
16 (kHz) × 4 (bit)
≅ 131 (sec)
The above equation gives the playback time when the phrase control table function is not used.
|Direct download click here|
|Share Link :|