Noah Yuan-Vogel
Well-known member
- Joined
- Apr 25, 2007
- Messages
- 1,054
- Reaction score
- 0
- Points
- 0
- Location
- New York, NY
- Website
- www.noahyv.com
Anyone know why R3D file size numbers dont calculate out that well? R3Ds always seem to end up smaller than expected. REDCODE is VBR, right? So it must usually be undershooting I suppose. It seems to undershoot a lot if it is using 16bit since my calculations that assume CBR and accurate compression ratios seem to show that the actual file sizes are even smaller than 12bit CBR files would be.
I was trying to see if I could use file size to determine the 12bit vs 16bit behavior of REDCODE which was never revealed. It seems we are better off doing our calculations assuming 12bit pixel depth since those are much closer to the actual file sizes. If you assume 16bit, you'll usually be off by about over 30%.
Here is a little worksheet I was using to investigate using various R3Ds I had lying around, most of which were shared on this forum.
https://docs.google.com/spreadsheet/ccc?key=0AgEPvD3dxujLdC1hNk9ncFJrcFBWNF9oQlhnWWptMnc
I was trying to see if I could use file size to determine the 12bit vs 16bit behavior of REDCODE which was never revealed. It seems we are better off doing our calculations assuming 12bit pixel depth since those are much closer to the actual file sizes. If you assume 16bit, you'll usually be off by about over 30%.
Here is a little worksheet I was using to investigate using various R3Ds I had lying around, most of which were shared on this forum.
https://docs.google.com/spreadsheet/ccc?key=0AgEPvD3dxujLdC1hNk9ncFJrcFBWNF9oQlhnWWptMnc