Raw video file format support

Background

Currently, the edgertronic always encodes and saves videos in H.264 format. We are often asked to add support for saving in RAW video format, either to avoid dropped frames or improve image quality.

The following information is intended to help you understand why these aren't issues and thus why we don't offer RAW video support. If after reading the following information you feel your use model requires RAW video support, please contact us so we can understand your requirements.

No Dropped Frames

The camera hardware and software is designed to never drop a single video frame, either during capture or H.264 encode. The camera contains a high-speed image sensor, custom hardware, and a fast DDR3 memory, The video frames go from the sensor, though the hardware, and are stored RAW in DDR3 memory. The hardware is designed to ensure that every frame from the image sensor is correctly stored in the DDR3 memory. After all the video frames are captured, the camera then encodes and saves the video to a file in h.264 format. Proprietary hardware and software algorithms ensure that all of the RAW frames in the DDR3 memory are encoded and saved. Even if the encode complexity and time vary from frame frame to frame, the camera system will record and save all the frames.

Since the edgertronic already prevents dropped frames during capture, encode and save, RAW mode is not necessary in this case.

Image Quality

The default H.264 encoder settings were selected so that the H.264 encoder is virtually lossless. In fact, any encode artifacts that exist are substantially less than the noise in the sensor.

There are several sources of noise that can effect image quality:

• Image sensor fixed pattern noise
• Image sensor random noise
• Other electronic noise sources
• Noise under user control (ISO and overclocking)
• H.264 encoder

Stated simply, you won't see any blockiness [1] or other encoder artifacts, and saving in RAW format won't make the video look better.

[1]blockiness in photography is the state of being blocky; indistinctness and unevenness of shading.

Image sensor limitations

The photons from the field-of-view that are being captured are accumulated by the camera's image sensor. SC1, SC2, and SC2X have different image sensors, with the SC2 and SC2X being very similar using the same technology but supporting different resolutions. The SC1 image sensor uses a different technology. That is why we have an SC1 series of cameras and SC2 series of camera.

SC1 is 10 bits / pixel from the sensor, but is not linear, so a lookup table turns it into 12 bit/ pixel data. Many (3-4?) of the LSBs are below the sensor noise floor.

The SC2 series also gets 10 bits / pixel from the sensor, but only stores 9 bits / pixel. The best guess is the SC2 series has 1-2 more usable bits than the SC1.

All told, for most customers 8 bits/pixel after we love up the data contains all the resolution we have confidence in.