Earlier today we explored how thumbnail grids can be used to rapidly visually triage unenriched television news that lacks captioning, sufficient onscreen text or other recoverable metadata. One approach we found especially powerful was a simple fixed 1/4 fps framerate thumbnail grid. How does this grid look at 160, 250 and 500 pixel widths?
The 160 pixel width appears to offer a unique sweet spot in that images are highly understandable, yet the complete thumbnail grid is small enough to be rapidly scanned by a person. The 250 pixel images lend additional clarity for some detail, but do not offer a substantially improved experience. The 500 pixel width images offer an enormous advanced in which even onscreen text is OCR'able, but the resulting image is too large for human previewing. However, it may be uniquely useful for nonconsumptive sampled visual OCR processing.