Not relevant. There are a handful of items in a file header that don't contribute to the bitrate, but they are negligible when dealing with files of more than a few MB. Also, I do not feel that packet information contained within the stream exempts itself from being classed as part of the bitrate. Even though it is negligible in the same way the file header is, at least it repeats itself depending on the duration of the film. You said it yourself:Devil wrote:Not necessarily true. There are many things embedded into files that have nothing to do with the bitrate.
Devil wrote:This is a tiny part of the data transmitted with a MPEG-2 video stream, large amounts of which are optional.
Also - you're confusing the issue by trying to split out the audio bitrate. I didn't, it would be kinder to the OP if you had the same courtesy.
I sometimes wonder if you get your jollies from trying to confuse novices, or whether you really know what you're talking about.
Why does any of that have any relevance to what we're saying here? For example, why does the algorithm matter? Surely that just affects the picture quality at a given bitrate? What you're suggesting is that a two-minute 2mbps file encoded using ulead's encoder will be a significantly different size to a two-minute 2mbps file encoded using CinemaCraft encoder.Devil wrote:Let me put it this way, you have a 30 Mb file containing 2 minutes of MPEG-2. What is the video bitrate? 2 Mbit/s? WRONG:
1. Is it low level, main level, high 1440 level or high level?
2. Which profile/I/P/B frame macroblocks are used?
3. What is the GOP?
4. What quantisation is used
5. What audio layers are used?
6. Which audio encoding is used?
7. What is the audio bitrate?
8. Is motion compensation used?
9. Who makes the encoder (algorithms change with the maker)?
That's just garbage. The quality will be different (I can testify to that - it's why I use CCE) but the filesize and bitrate will be more or less identical.
What you're suggesting is that a two-minute 2mbps file encoded using I frames only will be a significantly different size to a two-minute 2mbps file encoded using a standard IBP structure.
That's just garbage. The quality will be different (IBP frames will yield better quality) but the filesize and bitrate will be the same.
In a video stream, the effect of metadata is pretty much negligible. It's a bit like planning a 100 mile journey without accomodating the extra distance you'll travel because of lane changes or driving over studs in the road.
I fully agree. Finally.Devil wrote:...Put it another way: if you encode exactly the same clip with three different encoders at the same constant bitrate, you will obtain three different file lengths and they can vary quite considerably.
And then when you run those three files through a program such as BitRateViewer, you'll see three different bitrates reported. If the encoder you use is incapable of sticking to the bitrate you request, you need to accommodate that shortcoming when starting the process.
For example, I was getting pretty consistent results with all the encoders I tried, until I put DivX on its highest quality setting. The resulting file was 20% bigger than it should have been. Does that mean the file was full of extra cr*p, or does it mean that DivX did not keep to the requested bitrate, concentrating instead on quality? (For those of you trying to keep up with this rant, it was the latter.)
Yes - I know CBR is basically a narrow-deviation VBR. On most encoders there's a "padding" option that allows you to fill out the data packets if there's not enough picture data to do it.Devil wrote:BTW, did you know that even CBR varies in both bitrate and quantisation? Not much, but the bitrate is typically ¡Ó5-10%.
Yep - it's a balancing act between all the factors I mentioned - I've had the option of dropping the bitrate way below what I've been comfortable with, or going to half D1 resolution, to produce a wedding DVD. Neither were acceptable, so I ended up spreading the content over two discs rather than edit stuff out.troppo wrote:...I guess what I meant, is that by reducing the frame rate and frame size, you can also then reduce the bitrate without affecting quality too much. These 3 things should all be reduced together to find a happy medium between quality and file size.
I didn't know that - it's very bad that it behaves that way, because you're left "chasing the gauges" (a flying term).troppo wrote:To illustrate my point, in MSP8, if you increase frame size in the MPEG encoder options dialog, then it automatically increases the bitrate, to keep quality high.
