Table 12.
AVC/H.264 Video Encoder Advanced Settings – Page 1
Control
Description
Pulldown
Allows converting of video from 23.976 4 fps to 29.97 fps or 23.976 4 to
59.94 fps. The conversion is performed by adding special headers to the
stream. Therefore the number of frames is not changed in the stream.
•
No pulldown – Encoder ignores the presented pulldown headers and
forms new timestamps for every frame.
•
Auto mode – Encoder saves the information about pulldown headers
presence/absence in the input stream.
•
3:2i/2:3i – mode «i» composes the stream by fields «duplication». It is
designed for conversion of 24 into 30 fps.
•
3:2p/2:3p – mode «p» composes the stream by frames «duplication».
It is designed for conversion of 24 into 60 fps.
To convert 23.976 fps to 29.97 fps, select 3:2i pulldown and set fixed frame
rate to 29.97 fps. To convert 23.976 fps to 59.94 fps, select 3:2p pulldown
and set fixed frame rate to 59.94 fps.
B-frames mode
Sets the B-frames encoding mode:
•
All disposable — B-frames do not contain information needed for
another frames decoding.
•
All reference - B-frames contain information needed for another frames
decoding. It provides higher quality of compressed video.
•
Hierarchical - allows encoder to use hierarchical structure of B-frames
encoding. It increases quality of compressed video. The B-frames
number must be set to 3.
Recommended only for progressive video encoding.
Scene Change Detection
Enables the automatic insertion of I-frames (reference frames), when new
scenes are detected. If this option is enabled, the encoded stream quality and
positioning precision are improved. Some options will be inaccessible if the
scene change detection mode (SCD) is disabled. Recommended for
permanent use.
Adaptive GOP
Enables/disables adaptive B-frames placement. When video complexity is very
high, the encoder decreases B-frames count to increase encoding quality.
Recommended for permanent use.
Max rate
Specifies the hypothetic rate of the encoded data streaming. It is
recommended to set high bitrate value (2-3 times greater than average
stream bitrate). Is relevant only to the variable bitrate (VBR) mode.
Buffer size
Specifies the video buffer size.
Init Fullness
Specifies the initial CPB removal delay. Recommended value is 67500 (90000
clock units per 1 second).
Adaptive quantization
mode
Selects the adaptive quantization mode. The adaptive block quantization
improves visual video quality. Macroblocks will have different quantizers
according to the selected algorithm. The following modes are available:
•
Do not use – Adaptive quantization is not used
•
Complexity – Is based on macroblock complexity analysis
•
Complexity and Luminance – Is based on macroblock complexity amd
luminance analysis
Strength
Specifies the variation range of quantizers.
Transform size
Selects dimension of DCT transformation. Sometimes higher DCT dimension
provides better macroblock encoding and higher compression (for example
for HD video with massive details). The automatic selection is recommended.
Entropy coding
Selects entropy (lossless) compression methods:
•
CAVLC – basic compression method of video stream
•
CABAC – more effective (about 15-20%) but needs more CPU
resources when encoding and decoding
The CABAC method is recommended.
RD optimization
Selects the model of resource optimization for more effective compression:
•
SAD – quick (is based on sum of absolute differences).
•
SATD — slightly slow (is based on sum of absolute transformed
differences).
•
RDO - high-quality and high-effective model of resource optimization
(rate-distortion optimization).
Elecard Converter Studio
53
User Guide