Sunday, February 14, 2021

ISSCC 2021: Sony SPAD-Based HDR Sensor

Sony presents an HDR counter & time stamp-based approach for SPAD sensors:

"A 250fps 124dB Dynamic-Range SPAD Image Sensor Stacked with Pixel-Parallel Photon Counter Employing Sub-Frame Extrapolating Architecture for Motion Artifact Suppression"
by Jun Ogi, Takafumi Takatsuka, Kazuki Hizu, Yutaka Inaoka, Hongbo Zhu, Yasuhisa Tochigi, Yoshiaki Tashiro, Fumiaki Sano, Yusuke Murakawa, Makoto Nakamura, Yusuke Oike.


  1. Can we all just take a moment and acknowledge the horribly badly written and presented Sony paper on automotive dToF? Why is this even accepted? The author takes aeons to summarize the obvious "frame time", "slot time", "trig" his CV as well as device specifics already presented at IEDM - but completely fails to explains his actual readout circuit! How does he go from 189x600 to 63x168 - why does every block of 3 rows provide 81 outputs? Why does he use 4 coincidence blocks with 48 inputs? How does he use the adder/SRAM/FIR? This was hands-down one of the worst ISSCC presentations so far... and the paper doesn't do a better job at explaining anything. I have the feeling these papers just get accepted because they come from Sony. At this point they could just submit datasheets without any explanation...

    1. This happens when you put SPADs people in charge of the "image sensor" session. I heard many great image sensor papers were rejected because the SPAD pro tech committee only like SPADs. In the meantime, SPADs are still mostly prototype after burning billions of investment money and decades of R&D.

    2. Don't get me wrong - SPAD papers are very exciting and certainly contain more circuit centric novelty than most recent CIS publications. Problem is, the paper was badly written so this could've been anticipated. The paper could've been rejected. It was not rejected most likely as we haven't really seen Sony-SPAD related publications yet other than the device centric (actually well written and well presented) IEDM paper! We know Sony makes great technology - would be great to see that effort reflected in their papers. Some Sony papers are amazing - this one, unfortunately, wasn't.

    3. Which paper is pointed out? A paper in this post is not automotive dtof paper. I think this post presented the circuit novelty well as ISSCC paper.

    4. "A 189×600 Back-Illuminated Stacked SPAD Direct Time-of-Flight Depth Sensor for Automotive LiDAR Systems"


All comments are moderated to avoid spam and personal attacks.