Assuming that Canon 5D Mark II and Canon 60D both have the same processor, DIGIC 4 (single), why does the former shoots a better quality video? Better quality I mean the technical aspect, the video/compression itself.

Let's ignore the visual aspect like more shallow depth of field, for example.

I know 5D is a full-frame, but the video is 1920px wide. The 60D shoots 18MPx, which scaled down to 1920px wide should still look almost the same as if shot on full-frame initially, right?

Therefore, and this is my point:


If we exclude the depth of field and a better ISO handling from our equation, it's possible that the uncompressed input is the same on both cameras, right? Then, it goes to the digic 4 processor to be compressed, which is the same on both cameras. Yet, we end up with a better quality video on one of those cammeras. Why?


1 Answer 1


The 5D Mk2 has a far larger sensor (because it's a full frame camera) and more pixels. My assumption would be the camera scales down the input to 1920x1080 so having more uncompressed data allows it to give a higher quality output.

If Andres's comment about it skipping lines whilst taking video is right, then the improvement in quality is more likely down solely to its lower pixel density. A lower pixel density means a bigger area per pixel, which means more light per pixel, more light per pixel means less amplification is required (amplification increases noise also)

  • 3
    \$\begingroup\$ Quality is related to surface area, not linear size. So while the diagonal is 1/3 larger the surface area is 2.25X bigger. Otherwise, the full-frame advantage would be much less. \$\endgroup\$
    – Itai
    Jun 4, 2011 at 15:35
  • 1
    \$\begingroup\$ Right you are @Itai, I'll update my answer \$\endgroup\$ Jun 5, 2011 at 15:35
  • 2
    \$\begingroup\$ Are you sure that the camera scales down the video? I thought the HD resolution was achieved by skipping lines. \$\endgroup\$
    – Andres
    Jun 6, 2011 at 1:41
  • 1
    \$\begingroup\$ I believe @Itai has it nailed. The 5DII has considerably more surface area, which provides a lot more information to interpolate when scaling down to final output size. I would be doubtful that sensor pixel rows are skipped...however I wouldn't be surprised if sRAW-style YCC encoding was used rather than normal bayer interpolation (which uses more bayer pixel data per pixel, and does not overlap pixel interpolation.) \$\endgroup\$
    – jrista
    Jun 9, 2011 at 3:01
  • 1
    \$\begingroup\$ This article: blog.planet5d.com/2011/10/… Seems to suggest that line skipping did occur in older DSLRs \$\endgroup\$ Nov 1, 2011 at 13:00

Not the answer you're looking for? Browse other questions tagged or ask your own question.