Talk:SANM: Difference between revisions
Jump to navigation
Jump to search
No edit summary |
m (Talk:SNM moved to Talk:SANM: Changed codec name to its FOURCC instead of the files' extension (SNM).) |
||
(One intermediate revision by the same user not shown) | |||
Line 2: | Line 2: | ||
* In the preamble section, does the movie size count the preamble itself or not? | * In the preamble section, does the movie size count the preamble itself or not? | ||
* Same for the video header: Does the size count the entire header? | * Same for the video header: Does the size count the entire header? | ||
* If the video header is always known to be a constant size, it maybe useful to note that size. ''The video header size is not known for sure yet. I'll update the size when I find it out for sure.'' | * More on sizes: All of the size fields should stipulate what the size includes; alternatively, if they follow the same rule throughout, make a blanket notice earlier in the page. | ||
''Sizes indicate the contained chunk data size without the FOURCC and size field. I added this to the document.'' | |||
* If the video header is always known to be a constant size, it maybe useful to note that size. | |||
''The video header size is not known for sure yet. I'll update the size when I find it out for sure. It would appear that the size is constant because video data always starts at offset 0x238 of the Bl16 chunk, although a definitive quantity for the codebook size is yet to be figured out.'' | |||
* Are the padding bytes always 0? It might be better to list them all as reserved/set to 0. | * Are the padding bytes always 0? It might be better to list them all as reserved/set to 0. | ||
''As far as I can see, this is irrelevant because all that matters is that the annotation is terminated with at least one zero character. A robust implementation would seek to the end of the annotation anyway and ignore the padding.'' | |||
'' | * Do you know if the rle_decode() function is defined by the 'SRLE' tag in any way? I am working on that format for [[Humongous CUP]]. | ||
''Nope, the decompression method is specified only by the codec opcode. There are no "SRLE" tags in SNM.'' |
Latest revision as of 09:58, 30 April 2006
Questions/Notes:
- In the preamble section, does the movie size count the preamble itself or not?
- Same for the video header: Does the size count the entire header?
- More on sizes: All of the size fields should stipulate what the size includes; alternatively, if they follow the same rule throughout, make a blanket notice earlier in the page.
Sizes indicate the contained chunk data size without the FOURCC and size field. I added this to the document.
- If the video header is always known to be a constant size, it maybe useful to note that size.
The video header size is not known for sure yet. I'll update the size when I find it out for sure. It would appear that the size is constant because video data always starts at offset 0x238 of the Bl16 chunk, although a definitive quantity for the codebook size is yet to be figured out.
- Are the padding bytes always 0? It might be better to list them all as reserved/set to 0.
As far as I can see, this is irrelevant because all that matters is that the annotation is terminated with at least one zero character. A robust implementation would seek to the end of the annotation anyway and ignore the padding.
- Do you know if the rle_decode() function is defined by the 'SRLE' tag in any way? I am working on that format for Humongous CUP.
Nope, the decompression method is specified only by the codec opcode. There are no "SRLE" tags in SNM.