Compressing High Dynamic Range Photos Using the JPEG Standard

In the consumer digital photography market we have nearly all images stored in the JPEG compression format. While many new compression methods were proposed for providing somewhat better compression, and also a variety of new features, most users currently do not see compelling reasons for adopting new formats. The high dynamic range feature seems to be an exception, standing out in its appeal to consumers, because it can effectively improve photo quality. This feature is not supported by the baseline mode of JPEG, since it requires more than 8 bits of precision for light and color intensities. What is not well known is that the JPEG standard has one extended mode allowing 12-bit precision, which can be implemented with very small changes to current software and hardware, because the standard uses a technique that allows complexity to grow linearly instead of exponentially with the number of bits. We consider how this extended JPEG mode can enable high dynamic range photography in the consumer market, with much smaller investments than required for adopting a completely new format. We explain the technical differences among modes, what are the required changes, and the consequences in terms of complexity and compression efficiency.