DVD Basics

DVD contains VTS's (Video Title Sets) which containe PGC's (ProGram Chains) which contain Cells.

Typical picture sizes and their associated applications

* some DVD recorders use a HHR (half horizontal resolution) video frames (352x480) to record (fit) twice the time duration of video onto the tape

Standard MPEG Resolutions
 

Video Resolution  

used for . . .

NTSC
Resolution
(W x H) 

PAL
Resolution
(W x H) 

Comments

D1  

DVD

720x480  

720x576

DVD player takes the "height" of the video frames, and adds signaling for vertical retrace.  For example, the full vertical NTSC interlaced resolution is 525 lines, so an analog TV:

  1. scans 240 "odd" lines of visible video from top to bottom

  2. scans 22.5 "invisible" lines (beam shuts off) as the scanning position moves back to the top

  3. scans 240 "even" lines of visible video from top to bottom

  4. scans 22.5 "invisible" lines (beam shuts off) as the scanning position moves back to the top

cropped D1

 

704xNN 704xNN

the 8 pixels on either edge of the video frame aren't supposed to contain useful information. Therefore, some programs will prefer the cropped D1 resolution to save bandwidth

2/3 D1  

SVCD

480x480  

480x576

 

1/2 D1
or
HHR (Half
Horizontal
Resolution) 

VCD
or
DVD

352x480  

352x576

Many ask, "why 352 instead of 360, which is exactly 1/2 x 720 ??"  This is because the MPEG structure uses 16x16 "macroblocks", and therefore the horizontal dimension must be divisible by 16.  352 is divisible by 16, and 360 is not.

SIF
aka "Low Level"

VCD

352x240  

352x288

sometimes inaccurately called "1/4 D1"

1/4 SIF PC video clips 180x120 N/A

used to be common - now rare due to high resolution displays and video cards

NOTE:  some people use the term "D1" with an assumtion that it is NTSC
it can be either so make sure to always indicate NTSC or PAL !!

 

A VOB is simply an MPEG-2, Right?

WRONG.  This is a myth.  They're very similar, but a VOB has extra data and different headers.  The truth is that a VOB is an MPEG-2 wrapped in VOB headers, which is why it is possible to convert a VOB to an MPEG-2 quickly, if you use a converter that DOES NOT RE-ENCODE.

Some even say you can just change the .vob extension to .mpg and it will then be an mpeg-2.  But changing the extenstion doesn't turn it into an mpeg. If you can play a file after you've renamed it from .VOB to .mpg, it's only because your player is recognizing that file as a .VOB (even with the name change).  If you didn't have a software player capable of playing vob's, for example, you wouldn't be able to play your renamed vob to mpg file.

VOB stands for "Video Object" as defined in DVD specifications. Although a vob file is also an MPEG file, it has additional data that a standard MPEG decoder will not understan, e.g., the private data, which only a DVD decoder will understand. Those data include information about the file offsets and time duration of all chapters included in the VOB file.

Whenever a vob file is edited, some of those information will no longer be valid and may be discarded by the editor; and some other data ("user data" in the MPEG specs) which include the caption texts, will be preserved by the editor. This is why the Womble VOB editor, for example, does not have a VOB format for saving. 

There is no mpeg-2 software that has the capability of saving to a VOB file, since the VOB file format depends on the structure of the rest of the DVD.  Only DVD authoring software can create VOB's

MPEG-2 Standards

to make DVD's that play on any player, you MUST adhere to these standards !!!

*** also see MPEG-2 FAQ

The two types of Aspect Ratios - Display and Pixel

Display Aspect Ratio
(Width:Height)
Typical Pixel Aspect Ratio
(Width:Height)
Typical Use
4:3 525 line (NTSC) 10:11

NOTE:  10:11 is more commonly called .9 (in many Video applications)
10:11 is a ratio, which is the same as 10 11 = 0.909
and 0.909 rounded to one decimal place = .9

NTSC TV
4:3 625 line (PAL) 12:11 PAL TV
16:9 525 line (NTSC) 40:33 Widescreen NTSC TV
16:9 625 line (PAL) 16:11 Widescreen PAL TV

Television Standards
Your DVD player looks at IFO files, which are index & information files that
tell the DVD player how to play VOB files (which are MPEG-2 files wrapped in a header)
The DVD player then converts the video streams to these TV formats

"Standard" MPEG-2 for DVD's

It is definitely possible to make non-standard MPEG-2 files, and there are some DVD authoring packages that will even allow them as input and make a DVD.  You may even be able to play a non-standard DVD made from non-standard MPEG's on your DVD player.  However, only a 100% standard DVD will play on ALL DVD players.  So stick with that - or else you are just asking for trouble.  Here are the allowable standards for MPEG-2 video.  The framerate is also called "Hz" (Hertz):

Both 720x480 and 704x480 are from the NTSC's CCIR 601 standard
352x240 is for both MPEG-1 and MPEG-2, and is also called SIF (Source Input Format), also known as Low Level (LL)
352x480 is for non-interlaced, which is called "progressive" scanning - used at 24 fps

NOTE:  a common MPEG-1 Myth is that the max frame size is 352x240.  Actually, MPEG-1 permits sampling dimensions as high as 4095 x 4095 x 60 frames per second. The 352x240 size that most people think of as "MPEG-1" is really a kind of subset known as Constrained Parameters bitstream (CPB).

D1 and SIF

You may encounter some common terms such as "D1" (Digital 1 - which is broadcast video at 29.97 fps, MPEG-2 only) and SIF (Source Input Format, MPEG-1 and MPEG-2), SIF is also known as Low Level (LL) and, like D1, is at 29.97 fps.

Video Resolution  

NTSC  

PAL

D1  

720x480  

720x576

2/3 D1  

480x480  

480x576

1/2 D1  

352x480  

352x576

SIF  

352x240  

352x288

  • Full D1 is used for DVD's

  • 2/3 D1 is used for SVCD's

  • SIF is used for VCD's   

MPEG-2 Aspect Ratios (AR)

*** you will not always want exact aspect ratio's - because it is best to be able to divide evenly by a factor of 8 or 16.  The various video software editing and player packages work best that way

Display AR Terminology:

16:9  =  enhanced for widescreen TVs, anamorphic widescreen
4:3 = fullscreen, 1:1.33, pan&scan, standard edition, letterboxed widescreen.

When you encode MPEG-2 files for DVD, you use 720x480 for the source file aspect ratio.  But the actual displayed ratio is either 4:3, 16:9, or 2.21:1 (less common).  This is decided by the DVD player based on a parameter that is encoded into the mpeg-2 file. 

In addition, the Source File pixels can be square or oval shaped, which also affects the displayed aspect ratio.  Here are the various combinations of pixel aspect ratios and image aspect ratios - only one of these can be encoded into the mpeg-2 file

this is a screenshot from TMPGenc Express
so the term, "Image", which I have never seen to indicate
aspect ratios of videos . . . must simply be TMPGenc Express term for "Display"

If you have a square pixel (1:1) source file:

If your have a non-square, oval pixel source file:

the actual displayed aspect ratio of the encoded file can vary greatly !!  Therefore you are better off just selecting "square pixels" if this setting is even available in your encoder.  This will give you more constant results. 

Sample aspect ration settings using TMPGenc Express

TMPGenc Express allows you to select the pixel ratio in the "Clip Info" box.  Here are the effects of different settings selected for the source file.  Note the effects as far as the black bars are concerned:


NOTE:  10:11 is the same as the .9 non-square
pixel setting in many Video utilities

TMPGenc Express
720x480 Source Video clip Aspect Ratio Settings
Shown as displayed on a 4:3 Screen

The images show how the source file is interpreted and the resulting aspect ratio if it were to be displayed on a 4:3 screen.  However, the aspect ratio that you select for the final output will be either 4:3 or 16:9, which, as AVIcodec indicates, ends up being 16:9 or 21.1:1, which to add even more confusion, so long as square pixels were selected for the input, ends up being displayed by the DVD player as 4:3 or 16:9 !!!  The easiest way to understand all this, is to ignore the AVIcodec reported aspect ratio, and just realize the following:

 


the two DVD aspect ratios on a 4:3 screen  -  both using 720x480 video

NOTE1:  on a 4:3 screen, if using letterbox (widecreen) the 16:9 video is merely shrunk in size and retains the same aspect ratio, and therefore will have black bars on top and bottom.
If the DVD is played using "Pan and Scan" then it will fill the screen and the sides of the 16:9 are chopped off.

NOTE2:   if these were displayed on a 16:9 widescreen, then the 4:3 would have black bars at the sides

The MPEG-2 standard defines several "layers" which contain both data and variables.  The Sequence layer represents either interlaced or progressive video sequences, (MPEG-1 is strictly progressive).  In addition, MPEG-2 sequence layer contains a variable called "aspect_ratio_information", sometimes called a "flag".  It refers to the overall display aspect ratio (e.g. 4:3, 16:9).  With MPEG-1, the same variable refers to the ratio of the pixels, not the display.

NOTE:  in addition, another important variable is "frame_rate_code", which with MPEG-2 refers to the display rate.  The same variable in MPEG-1 refers to the actual coded frame rate.

So, the mpeg-2 file structure has internal information that flags the DVD player as to whether it should be played as 4:3 or 16:9.  This can actually be edited without re-encoding, through the use of a utility called DVDpatcher.

In BOTH cases, the MPEG-2 file is the same pixel size of 720x480.  So, pixel-wise, that is an aspect ratio of 1.5:1, or simply 1.5  -  and to make it really confusing, here are the various aspect ratios normalized to 1 :      720x480 = 1.5   but 4:3 = 1.33  and  16:9 =   1.78:1

As you can see, none of them are equivalent !!  But today's TV's are either 4:3 or 16:9, and therefore, to fill those screens the DVD player, when given a video clip of 720x480, will NOT play it at the 720x480 aspect ratio of 1.5.  It will either go below that for standard TV, or above that for HDTV.  The DVD player inspects the header of the MPEG-2 file, reads the aspect ratio, and plays it accordingly.

Using the CucuSoft encoder as an example  .  .  .


AVIcodec Report of a CucuSoft Transcoded MPEG-2
CucuSoft Settings = 720x480, 4:3
Use for Standard TV viewing

 


AVIcodec Report of a CucuSoft Transcoded MPEG-2
CucuSoft Settings = 720x480, 16:9
Use for widescreen, HDTV viewing

CucuSoft, at the 16:9 setting, creates MPEG-2 clips at 2.21:1, which will give you a perfect DVD screen on a WideScreen TV.  However, if you have a standard 4:3 television, then select 4:3 in CucuSoft.  Surprisingly, the resulting video will be 16:9 !!!  But as it turns out, a 16:9 video will be changed to fit your 4:3 screen.  In both cases the video clip is indeed 720x480, but the clip contains elements that tell the player what aspect ratio to use.

This is a small but extremely capable utility.  It has tabs on top to select the type of output you want.  Do not select "DVD" because it will create VOB's ina DVD structured folder.  Instead, select MPEG-2.

Other Factors affecting the Displayed Aspect Ratio

To calculate and figure all this out - impossible - you need to consider things such as anamorphic squeezing of the picture, and non-round elliptical pixels.  Both of those features can take a 16:9 video and display it as 4:3

MPEG Transport Streams (*.ts files)

You will probably NEVER see a transport stream mpeg.  The ones we all deal with, *.mpg files, are called "program streams" or "system streams".

Transport stream (TS) is used in MPEG-1, MPEG-2, and MPEG-4 systems protocols.   Designed initially to allow multiplexing of digital video and audio and to synchronize the output.   Has features for error correction for transportation over imperfect media, and are used in broadcast applications.   The MPEG-2 transport stream is defined in the ISO/IEC standard (13818-1).

Multiple MPEG programs are combined then sent to a transmitting antenna. In the US broadcast digital TV system, an ATSC receiver then decodes the TS and displays it on your TV.

Multiple MPEG programs are combined then sent to a transmitting antenna. In the US broadcast digital TV system, an ATSC receiver then decodes the TS and displays it on your TV.


The transport stream consists of several layers of communications:

  1. Composition of the various programs.  
  2. Packetized elementary stream (PES)
  3. Elementary stream (ES) - audio or video (the below is for video only)
  4. Group of pictures (GOP)
  5. Slice - Lowest layer that encoder can change DCT coding values to manage bit rate
  6. Macroblock - Consisting of 4 DCT blocks
  7. Encoding block or just block - A DCT encoding block, 8x8 luminance pixels

The transport stream is broken up into fixed length 188 byte packets (for compatibility with ATM systems[1]). Each 188 byte packet contains a 4-byte header, with a Program Identifier (PID) to identify which program the packet belongs to.  The ATSC TS contains an Electronic program guide.