Question about FPS/bitrate<->filesize relationship in containers.

[Gew]

Hiya!

Question #1)
Just curious. I sat playing with Xvid bitrate calculator. I took 2h/700MB (no audio) for example parameters. Funny thing then, I noticed, when selecting AVI-Legacy, calculated bitrate differed 8kbps in the range from 15fps to 60fps, even though it's the very same duration. I then selected AVI-OpenDML. 6kbps difference. I then tried selecting Matroska. Now the difference was only 4kbps. Finally, I tried selecting OGM in the bitrate calculator. Now, the bitrate was stuck (806kbps) regardless of if I picked 15,25,30, or 60fps. Strange, huh? Of course, there is some logical explanation to this. I have a feeling that it could have something to do with different container holding different amounts of 'overhead' for each second(?), determined by the amount of frames its holding. Something like that, not entirely sure, that's why I posted here. I'm 100% sure that someone could offer me clarification on this one.

Question #2)
Also, when I was messing around with this calculator, something else struck me. Is the relationship between duration<->bit rate absolute, sort of like Einstein's well known E=mc2 formula. I'm thinking the following. I have this movie. It's exactly 2h long. I wish to encode it into 700MB. Bit rate calculator tells me 816kbps is right. I then try something daring/brave to achieve higher bit rate, without exceeding the desired 700 MB. I simply AssumeFPS(25) and squeeze the audio to match. The duration for my film is now only 1:55:05. Calc now allows me to go slightly higher on the bit rate (700MB). This is where I wonder if there is some sort of equivalence to the Einstein formula, saying that the resulting video will have the exact same quality (compression-by-quantization loss), simply because 2h/23.976fps@816kbps equals(!) to 1h55m5s/25fps@851kbps. Is this correct?


Ty in adv~
Regards~

[jagabo]

Frames per second makes no significant difference in bitrate calculations. The differences you see have to do with container overhead (AVI vs MKV vs ODM).

file size = bitrate * running time

Plus a little container overhead. Container overhead can vary a little depending on the frame rate and audio interleaving because each "chunk" of data (a frame, a bit of audio) has a few bytes that identify the chunk and its size.

[Cornucopia]

The ONLY time frame rate has a DIRECT influence on bitrate calculations would be be when you're talking about Uncompressed video.
Then, you determine bitrate in this way:

bitrate = (horizontal resolution * vertical resolution * pixel colordepth) * framerate
**The part in parentheses is the spatial component, the framerate is the temporal component.

Framerate DOES have a big INDIRECT influence on bitrate, in that the higher the framerate, the higher the bitrate will need to be to maintain the level of quality (all other things being equal). Or to put it the other way, if you keep the bitrate the same (on the same material) and raise the framerate, your per-frame quality will lower.


SO, in answer to Q#1: No, that's not strange at all. It's what I'd expect. You are correct about the OVERHEAD thing.
Q#2: Yes, you are correct. And what jagabo said is correct. Look at that formula and think of the units...
Filesize is in MB or Mb. Bitrate is in MBperSec or Mbps, kbps or similar. Duration is seconds or minutes, etc.
That's how you know you got the right formula, when you can go from one unit type to another, cancelling out the conversion factors.

Hope that helps,

Scott

[Gew]

Ty alot, jagabo and Cornucopia.

In fact, I read jagabo's post before going to bed the other night. I then thought a little closer on the whole deal, and just before I shut my eyes I came to the conclusion that his answer to Q1 is in fact also an indirect answer to Q2.

Case closed!