We're looking at using MPEG2 as an alternative to ProRes 422 for delivery & archiving.
There are many options in there which are beyond my knowledge; I would like to understand them and maybe get even better results. Is there a guide to these terms?
MPEG-2 works by compressing in two different ways. Within a frame, it looks at groups of pixels (blocks) and compresses them based on patterns it finds within them. This is kind of similar to how JPEG compression work. the Macroblock quantization level describes how loosely the encoded block can represent the original block. The lower the number, the higher the quality but the more data required to represent the block since less error is accepted when trying to compress.
VBV Buffer Size is the size of the buffer required in order to maintain a constant frame rate. Each frame of video in MPEG-2 requires different amounts of data from other frames. Assuming the stream arrives at a constant data rate, a certain amount of buffer is needed to ensure that data continues to arrive in time to present a frame since the rate of arrival may sometimes be lower than the rate of data being consumed. This value allows the player to judge how much data it needs to buffer in order to avoid running out of information to display to the viewer.
Noise Control is a means of trying to avoid encoding inefficiencies from noise present in the video. MPEG-2 and other predictive encoding schemes rely on the fact that video tends to not change from frame to frame. Noise, however, is random and not able to be predicted. This is a significant problem for high efficiency compression since it can't be predicted. Reducing the noise before encoding results in better compression rates as the video is more predictable, but also results in some loss of detail as it is not possible to distinguish between noise and fine detail in many cases.
Write SDE writes out information about the vertical and horizontal size of the frame.
Intra DC Precision is the number of bits used (thus the accuracy) of the DC coefficients. It also impacts how accurately blocks are quantized. Higher values mean more precision is used, but also that higher data rates are required.
Ignore Frame Interval I'm not 100% sure on for that encoder, but generally it allows the encoder to adjust the keyframe interval as it deems necessary. It may use more or less data as needed, low motion video will likely benefit as the keyframes can be further spread out, however high motion video is more likely to suffer if the interval becomes too long.
After some deep Googling, I did find explanations for a few of these options:
MPEG macroblock quantization level. Valid values are from 2 to 31. Low values of this variable mean higher quality and higher bitrate, so 2 is the highest quality and 31 is the lowest quality.
VBV buffer size in units of 16k bits. The meaning of this parameter is confusing because it is used to help predict if the MPEG decoder might suffer buffer overflows or underflows. VBV buffer over/under flows are more commonly a problem when in CBR mode. If significant artifacts are present in the generated movie, try increasing this parameter.
(http://vis.lbl.gov/NERSC/Software/express/help6.2/help/relnotes/mpeg.htm)
Intra DC Precision: the MPEG-1 DC value is mandatory quantized to a precision of 8 bits. MPEG-2 introduced 9, 10, and 11 bit precision set on a picture basis to increase the accuracy of the DC component, which by very nature, has the most significant contribution towards picture quality. Particularly useful at high bit rates to reduce posterization. Main and Simple Profiles are limited to 8, 9, or 10 bits of precision.
http://www.gadegast.de/frank/mpegfaq/mpe1736.html
In MPEG-2 video, Intra DC Precision indicates the number of bits for quantized DC coefficients of intra-coded blocks; it takes one of the values of 8, 9, 10, or 11 bits. The more bits are used, the more precise quantization is achieved.
