Results 1 to 2 of 2

Thread: DTMF Detection - How to adjust detection parameters?

  1. #1
    brent Guest

    Default DTMF Detection - How to adjust detection parameters?

    Is it possible to adjust the internal DTMF tone detection parameters in the same way that you can define level threshold, signal to noise ratio, twist, etc. with user defined tones and tone groups using inctonedet.cfg?

    Specifically, when talk-off is a potential problem, which of the default DTMF detection parameters that can be modified have been found to be effective in reducing talk-off?

    (Environment: HMP Analog using GP)

  2. #2
    cxu Guest

    Default DTMF Talkoff

    There is a parameter that can be adjusted at HMP low level API. This parameter has not yet been exposed to GrandPrix configuration files.

    To change the default value for GrandPrix applications, the following steps need to follow. It is recommended to get the parameter changed before the media processing activities start on each channel.

    Please note: reduce the possibility of talkoff will increase risk of DTMFs not being captured in some cases.

    1. Retrieve the channel handle through bypass function PKX_CHANNEL_GetBypassInfo (...).
    2. Call the HMP low level function call to retrieve the default settings for the DTMF detector PKH_DD_GetConfig(...).
    3. Set new value to "talkoffThreshold".
    typedef struct {
    PK_FLOAT talkoffThreshold; /* Threshold value between 0.0 and 0.9. */
    /* This threshold can be used to decrease the chance of falsely */
    /* detecting speech as a DTMF digit at the expense of the ability */
    /* to detect good DTMF digits in the presence of voice echos. */
    /* The default value is 0.5. A higher value decreases talk-off. */
    } PKH_TDDConfig;
    4. Set the parameters to DTMF detector PKH_DD_SetConfig(...).

    Quote Originally Posted by brent View Post
    Is it possible to adjust the internal DTMF tone detection parameters in the same way that you can define level threshold, signal to noise ratio, twist, etc. with user defined tones and tone groups using inctonedet.cfg?

    Specifically, when talk-off is a potential problem, which of the default DTMF detection parameters that can be modified have been found to be effective in reducing talk-off?

    (Environment: HMP Analog using GP)

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •