Link to inside page

Software IP

Speech CODEC

Contributing to system cost reduced.

Demo Board

Speech CODEC Experience the Demo Board

* You need Adobe Flash Player to play the demo.
Adobe Flash Player

Sound Quality Sample Playback

8 kHz (1/2 compression)

Background

 Recent years have seen advances in the recognition of accessibility issues, and more and more products in our daily life are being equipped with speech output capability for giving operating instructions, alarm messages, etc. by speech sounds. However, such speech output capability is normally realized by a dedicated IC, which has prevented its widespread use in portable equipment and other products that must meet exacting space-saving and low-cost requirements.
 The Speech CODEC software IP offers a simple solution for implementing the speech output feature by eliminating the need for a separate IC. By combining the Speech CODEC with an application program in one of our supported MCUs, the added value of speech output can easily be realized with a single IC.
 Toshiba’s proprietary CODEC technology enables compression of speech data into less than half the original size by combining 1/2 compression and anacoustic zone compression. It is also possible to break up speech data into reusable parts and to play back speech sounds synthesized from these parts. This allows effective use of ROM area.
 The Speech CODEC is offered in two package types: for playback only and for recording/playback. The supported MCUs include the TLCS-870/C Series, TLCS-870/C1 Series and TLCS-900 Family, offering customers a wide variety of choices.
 The Speech CODEC is adopted in various application fields, such as alarm devices, sensors, medical appliances, home appliances, automotive security, personal portable equipment and educational toys.

The Speech CODEC is adopted in various application fields, such as alarm devices, sensors, medical appliances, home appliances, automotive security, personal portable equipment and educational toys.

Features

  1. Supports the TLCS-870/C Series, TLCS-870/C1 Series and TLCS-900 Family microcontrollers.
  2. Compact software size and Toshiba's proprietary CODEC technology capable of compressing 8-bit speech data into less than half the original size enable the implementation of speech output capability in unused ROM area.
  3. Eliminates the need for a dedicated IC, offering customers a low-cost, space-saving speech output solution.
  4. Supports various sampling frequencies, allowing speech quality to be set freely. It is also possible to play back speech data synthesized from speech parts of differing sound quality.
  5. Allows on-site modification and tuning of speech data when used with a Flash microcontroller.
Example of Synthesized Speech Output
For example, time announcements can be synthesized from just a few speech parts, such as "it's", "nine", "a.m." and "p.m."

Example of Synthesized Speech Output

Speech CODEC Implementation Image

Speech CODEC Implementation Image

Estimated Maximum Playback Time when Using Speech CODEC (*1)(*2)
Sampling Frequency (*3)
6 kHz
8 kHz
16 kHz
Memory Size
60 KB
20.5 s
15.4 s
7.7 s
96 KB
32.8 s
24.6 s
12.3 s
128 KB
43.7 s
32.8 s
16.4 s
256 KB
87.4 s
65.5 s
32.8 s
512 KB
174.8 s
131.0 s
65.5 s
  • *1： Playback time is calculated assuming that the entire memory area is used for speech data.
  • *2：  In actuality, longer playback time is possible due to the effects of anacoustic zone compression.
  • *3：  Sampling frequencies other than those shown in this table can also be set.
Key Specifications
CODEC Toshiba proprietary CODEC
Quantization 8 bits/sample
Supported Output Format Monophonic
Sampling Frequency 4 to 16 kHz (*4)
Data Compression Rate Less than half the original size by using Toshiba proprietary compression technology
Target MCU TLCS-870/C Series, TLCS-870/C1 Series, TLCS-900 Family
CPU Load (*5) Recording: 50%
Playback:  30%
Resource Consumption (*5) Recording & playback: ROM 390 bytes, RAM 13 bytes
Playback only:              ROM 250 bytes, RAM 10 bytes
  • *4： Depends on the operating environment.
  • *5： MCU = 870/C Series (@16 MHz), Sampling frequency = 8 kHz
Speech CODEC Summary Sheet
Software IP MCU/Series Features Summary Status
Speech CODEC TLCS-870/C
TLCS-870/C1
TLCS-900/L1
TLCS-900/H1		 Toshiba original compression algorithm. The CPU load and memory usage are optimized to enable also the control of peripheral devices with embedded state in single-chip microcontroller. Speech codec application development environment and evaluation board are available. [PDF:53KB] Available

Top of this page