Comments

john_dennis wrote on 9/18/2014, 11:42 AM
The original comment was here. The thread didn't take off, but I've found the concept useful and thought about it ever since Nick posted it.
musicvid10 wrote on 9/18/2014, 12:35 PM
Thank you, John Dennis!
The Ben Waggoner "^.75 Rule" is exactly what I was looking for!
PeterDuke wrote on 9/18/2014, 8:28 PM
I deleted my question because I read the article title AFTER I posted.
NickHope wrote on 9/19/2014, 10:13 AM
Thank you John :)

(...up to my neck in new PC configuration, hence slow reply)
Hulk wrote on 9/19/2014, 10:54 AM
When looking at the total number of pixels in the same video, the one with the higher resolution will generally have more redundancy, which makes compression more efficient. I write "generally" because if you have a lot of noise, blooming, or other artifacts then the encoder could actually use the increased pixel count if the HD video to encode these artifacts!

It has been my experience that with a good quality, clean source and a high quality AVCHD encoder SD video can be compressed down to 2mbps, while 1080p needs more like 4 or 5mpbs.

I'd say the "0.75" guideline is a good one.
musicvid10 wrote on 9/19/2014, 2:37 PM
". . .while 1080p needs more like 4 or 5mpbs. "
Might be fine for an interview, but football won't hold up at that bitrate.
musicvid10 wrote on 9/19/2014, 2:46 PM
Many things to be learned just by searching "bits per pixel" on the internet.
Ben Waggoner's recent thoughts on the "^.75 Rule" (note that's "to the power of," not "times") can be found here:

http://forum.doom9.org/showthread.php?t=167816

Here's a brief example: The area of 1080p is 2.25 times greater than 720p. If our 720p video looked great at 8 Mbps, we would need about 16.7 Mbps for our 1080p version to play at optimal quality, not 18 Mbps as the plain math would suggest.

(2.25^.75)*8Mbps = 16.7 Mbps

Thanks Nick, for teaching me that!