Home | Site Map | Contact | |
| Products | Solutions | News | Technology | Company | Downloads





December 23, 2011

GritTec has updated the source code of sample rate converter (Version 1,30) under Intel x86/x64 platforms.

As known, at the moment all personal computers are delivered with the motherboards which includes a sound card. All such a sound cards support only a few basic sampling rate frequencies which help the sound card to quantize an analog signal into digital signal (by using ADC) and a digital signal into analog signal (by using DAC). As a rule, manufacturers of such sound cards choose values of basic sampling rate frequencies equal 22,05 kHz, 32 kHz, 44.1 kHz, 48 kHz and 96 kHz. And if developers of audio applications try to quantize an analog signal with frequency of sampling rate not coinciding with one of basic frequencies then as a result periodically fragments of "gluing" appear in the digital signal. So if the analog sinusoidal signal with frequency (example, 300 Hz) is input into sound card then it can be hear a few crackling. At At detailed stretching of osillogramm of such digital signal at time axis, it can be noticed fragments of "gluing", appearing with a periodic time interval and in which there is a sharp change of a phase of sinusoidal signal.

For decision of such problems it's recommended to use the technology of sample rate converter. Example, if it's necessary to quantize a audio signal with frequency of sampling rate 8 kHz then a first it's recommended to quantize an analog signal with one from existing of basic sampling rate frequencies in sound card (for example 32 kHz). And next to thin out the recorded digital signal by using sample rate converter technology up to 8 kHz. In result it will be received the digital signal with sampling rate 8 kHz without losses in quality of digital process.
The decimation procedure of initial digital signal up to low frequency will be useful for system in which the recorded digital signal is used for secondary analysis or his coding and transmition into network, for example into full-duplex applications (VOIP), when the recording and playback makes at once with one sample rate frequency. In this case it's necessary to use the technology of sample rate converter twice: at first time it's used for recording (from 32 kHz with decimation up to 8 kHz), second time - for playback (increasing from 8kHz up to 32 kHz).

GritTec has optimized a source code of sample rate converter in updated version for tasks under Intel x86/x64 platforms. The technology of sample rate converter is accessible to developers in the form of source code or SDK libraries with supporting under MS Windows x86/x64, Linux x86/x64 platforms and with examples of test projects under MS Visual Studio C ++ 2008 and gcc 4,x compilers. To test demonstration of a new version of GritTec's Sample rate converter (Version 1.30) engine, please download demo program via GritTec's website or contact us via Online Request form.

About GritTec
GritTec Laboratory (GritTec Ltd.) specializes on research and development of algorithms and technologies in the field of speech and audio processing. GritTec's research is focused on speech enhancement, speech concealment, voice biometric, speech recognition, speech synthesis and other speech and audio technologies.
Url: http://www.grittec.com