ok, so i've seen everyone talk about making the horizontal rez divisible by 16.....why is this? does it have to do with the decoding or something?
Printable View
ok, so i've seen everyone talk about making the horizontal rez divisible by 16.....why is this? does it have to do with the decoding or something?
Actually, horizontal and vertical pixel counts need to be even numbers and the horizontal counts need to be 16 pixel multiples because of the core frameworks that Quicktime uses.
The simple answer and beyond the scope of Catalyst
Is that the line count or pixel width counts are stored as 32 bit integers where 16 of those bits are the line numbers in the group and 8 of those bits are an address, the remaining 8 bits are header and checksum.
This isn't just true of Quicktime, it applies to all graphics files that are derived from OpenGL frameworks. Some codecs autocorrect for this by populating the vacant lines in a user's output with black empty lines. Other codecs simply apply null data.
Unfortunately when catalyst runs, it uses an aspect of QT which misinterprets the null data and considers it render attributes.