Thursday, April 25, 2013

Toshiba Announces 1/3-inch HDR Video Sensor

Business Wire: One more time Toshiba announces a full HD (1080p60) 1/3-inch image sensor for the security/surveillance and automotive markets featuring 2.7um pixel, HDR and color noise reduction (CNR). According to the company, the new TCM5117PL has an industry-leading DR of 100dB. The output data format is 10b in normal mode or 12b in HDR mode.

Toshiba's single-frame, alternate-row, dual-exposure HDR implementation in the TCM5117PL is said to be well suited for capturing fast-moving objects compared to conventional multi-frame systems. With frame rates at 30 fps for conventional multi-frame systems, the single-frame approach, at 60 fps, promotes faster, less blurry and higher-quality images especially in high-contrast light conditions.


  1. how do you cover 100dB DR with 10~12 bits? I thought that you need at least 17~18 bits.

  2. You need 17-18 bit to cover 100dB SNR, non DR. Consider that usually the SNR of pixels is below 40dB. You can cover 100dB DR for instance with a floating point coding.

    My questions are:

    Who is the patent holder of the alternate row, dual exposure technique? Many big players are using this.

    If the effective nr of pixels is 2008 x 1168 and 2 rows are used as a single pixel, how can they claim a full HD single exposure HDR?

    1. 12 bits can be used to cover 100 dB the encoding uses a log scale rather than a traditional linear scale. For example if 1 digital number was 0.025 dB, then 12 bits would cover just about 100 dB


All comments are moderated to avoid spam and personal attacks.