MultimediaKB
This page is an attempt to create a multimedia knowledge base, i.e. a collection of all sorts of information about digital and analog video, audio and image formats.
References
- The Wikipedia article about Aspect Ratio in digital images has a wealth of information about movie/video and still image formats.
Glossary
- Anamorphic
- A format and technique to record a widescreen picture on standard 35mm film.
- DAR
- Display Aspect Ratio. This term is used both by MPEG2 and MPEG4-AVC.
- FAR
- Frame Aspect Ratio. This term is used by MPEG4-AVC. It has the same meaning as "Storage Aspect Ratio" in MPEG2 terminology.
- PAR
- Pixel Aspect Ratio. This term is used by MPEG2. It has the same meaning as "Sample Aspect Ratio" in MPEG4-AVC terminology.
- SAR
- Can have two different meanings. 1) Storage Aspect Ratio in MPEG2 terminology. This is the same as "Frame Aspect Ratio" in MPEG4-AVC terminology. 2) Sample Aspect Ratio in MPEG4-AVC terminology. This is the same as "Pixel Aspect Ratio" in MPEG2 terminology.
Aspect ratios
The information in this section pertains to video and was taken from these sources:
- This superuser.com answer
- This videohelp.com forum post
- This Wikipedia article
A video file has three different aspect ratios. The MPEG2 and MPEG4-AVC standards use different terms for these aspect ratios.
- A video file consists of frames. Every frame in the video file has the same dimensions, i.e. every frame is the same number of pixels wide and high. A frame's aspect ratio is calculated by dividing the frame's width by its height. For instance, if a frame is 1920 pixels wide and 1080 pixels high, the frame's aspect ratio is 1920 / 1080 = 1.777777777777778, commonly known as 16:9. MPEG2 calls this the "Storage Aspect Ratio" (SAR), MPEG4-AVC calls this the "Frame Aspect Ratio" (FAR).
- The pixels in a video file also have an aspect ratio. This describes the shape of the pixel. A square pixel with the same width and height has an aspect ratio of 1.0. Why is this needed? Today when we think of modern computer displays we don't consider pixel aspect ratio because we naturally assume that the computer display has square pixels. Historically, however, other display devices such as PAL or NTSC TV displays had rectangular pixels, i.e. pixels that did not have the same width and height. Now that we know that non-square pixels can exist in the physical world at some level, it is conceivable that non-square pixels also enter the digital world due to some digitizing or conversion process, and that therefore the pixels in a video file might need to have a custom aspect ratio. MPEG2 calls this the "Pixel Aspect Ratio" (PAR), MPEG4-AVC calls this the "Sample Aspect Ratio" (SAR).
- A video file is supposed to be displayed in a certain aspect ratio. Both MPEG2 and MPEG4-AVC call this the "Display Aspect Ratio" (DAR).
The mathematic equation to relate the three aspect ratios to each other looks like this:
DAR = SAR * PAR (MPEG2) DAR = FAR * SAR (MPEG4-AVC)
Anamorphic
For details see the Wikipedia article about the anamorphic format.
Anamorphic refers to the process of recording a so-called "widescreen" picture onto a recording media (typically film) that has a non-widescreen format. Basically this means that ...
- ... the lens captures the picture in the desired playback format
- ... the lens distorts the picture so that it can be projected onto the recording media
- ... at playback time, the process needs to be reversed
You begin by comparing the desired playback format and the format of the recording media. There are two possible outcomes:
Playback format Recording format
+-----------------------+
| |
+--------------------------------------+ | |
| | | |
1) Playback format | | | |
is wider than | | | |
recording format | | | |
| | | |
+--------------------------------------+ | |
| |
+-----------------------+
+-----------------------+
| |
| | +--------------------------------------+
| | | |
2) Playback format | | | |
is higher than | | | |
recording format | | | |
| | | |
| | +--------------------------------------+
| |
+-----------------------+
A picture is anamorphosed only in one dimension:
- The dimension of the playback format that in comparsion with the recording format is larger is projected onto the recording media without distortion
- The dimension of the playback format that in comparsion with the recording format is smaller is anamorphosed, i.e. is projected onto the recording media with distortion
- In example 1 above, anamorphosis occurs vertically, i.e. the picture is stretched vertically
- In example 2 above, anamorphosis occurs horizontally, i.e. the picture is stretched horizontally
This is the only way in which anamorphosis makes sense, because the goal is to maximise overall image detail. If anamorphosis were performed in the opposite direction, the process would lose detail.