File structure analysis continued

Refer to the SWF file format document officially released by Adobe, but it is not quite clear. In Versions later than Flash 6, the generated SWF files support compression, so it is of little use for parsing uncompressed SWF files.

Let's take a look at how to analyze it! I analyzed it with a compressed file. (In fact, I cannot find the uncompressed SWF file of the lower version)

SWF file header information

The SWF file header contains a large amount of information. It tells people the basic information about how SWF files work.

The Macromedia Flash file format is the file format where the Macromedia Flash Player publishes vector images and animations on the Internet.

The SWF file format is designed not to share images between image editors but to efficiently interpret the format. It is designed to meet the following conditions:

A: real-time display on the screen: The main purpose of this format is to display and support instances on the screen, quick playback of color images, animation and interactive buttons.

B: Scalability: This format is a label format. Therefore, its player can be upgraded by adding new features to be compatible with earlier versions of its player.

C: network publishing: This format can be transmitted with limited network resources and unpredictable network conditions. Files are compressed to a very small size and stream playback is supported. A swf file is a binary file that cannot be directly read and understood like an HTML file. SWF files use methods such as bit compression and optional field structure to reduce the file size.

D: Simplicity: This simple structure makes flash players very small and portable. In addition, flash players only rely on a limited set of operating system features.

E: file independence: file display does not depend on any extended resources such as fonts.

F: variability: it is important that files work well in a limited hardware environment and use better hardware as much as possible, because the user's computer has different display settings and color depth.

G: Speed: files can be played quickly and with high quality.

H: script supported: This format contains tags in the agreed format, which specifies the sequence in which the stack machine interprets bytecode. Bytecode supports an Action Script Language. Flash Player specifies the runtime Action Script Mode, which allows the interaction between the original rendering, service, and Flash Player features.

The extension name of the SWF file is .swf. It is a mime (multi-purpose Internet Mail Extension protocol) type application (X-Shockwave-flash ).

Swf format has gone through several versions. In the 5th versions, SwF tag settings have been enriched and improved on a large scale. After the first version, the file format changes little. For example, more and more new flash features or full implementations are available only in ActionScript. Therefore, if you want to add more new features to the SWF file, you should be familiar with the ActionScript object model provided by Flash Player. The best reference for these is O 'Reilly's ActionScript: the definitive guide, published by Colin Mok.

SWF file header: All SWF files start with the following header:

The file header starts with a three-byte identifier, Which is 0x46, 0x57, 0x53 ("FWS ") or 0x43, 0x57, 0x53 ("CWS. The "FWS" identifier indicates that the file is an uncompressed SWF file, and the "CWS" identifier indicates that the first eight bytes of the file are later than the file length field) all the data is compressed by the open-source standard zlib.

The data format used by the zlib library is described in the request annotation from 1950 to 1952, and is detailed in the internet standard documentation (rfcs. CWF File compression only applies to version 6 or later.

The identifier is a byte version number, which is not an ascii code but an 8-digit number. For example, version 4 is identified by 0x04, instead of the ASCII code "4" (0x35 ).

The file length field is the length of the entire file including the file header. If it is an uncompressed SWF file (the identifier FWS), the length field should exactly match the file size, if it is a compressed SWF file (identifier CWS), the file length field refers to the total length of the file after being decompressed. Therefore, it generally does not match the file size, the uncompressed file size can improve the decompression efficiency.

The frame size field defines the width and height of a video. It uses a rect structure for storage, which means that its size can vary according to the number of digits required for coordinate encoding. The minimum value of the X and Y axes used by the rect structure of the frame size is 0. The maximum values of X and Y axes indicate the width and height (see bit values)

The frame rate is the frame playback rate per second. This rate is not guaranteed when the SWF file contains streaming audio data or when the flash player runs on a low-speed CPU.

The number of frames is the total number of SwF video frames.

SWF file structure

The file header is followed by a series of consecutive tag data blocks. All tags share a common format. Therefore, any program that parses the SWF file can skip unidentifiable data blocks. Data in a block can point to the offset in the current block, but cannot point to the offset in other data blocks. This allows the tag to be removed, inserted, or modified by the tool that processes the SWF file.

SWF File Format

========================================================== ========

Tag format

Each label starts with a type and a length value. There are two label header formats: short format and long format. The short label header is used for 62 bytes or smaller Label data. The long label header can be used for any label data of no more than 4 GB. In the long run, this will be very practical.

Note: The tag encoding and length fields are not a 10-bit field followed by a 6-bit field, but a double byte field, the uninterrupted byte sorting of SWF files makes the two different la S.

The length of this field does not contain the record header at the beginning of the tag (that is, the length occupied by this field ).

If the length of a tag is greater than or equal to 63 bytes, it is stored in the long label header. The long label header consists of a 63-byte (0x3f) Short label header and a 32-bit length.

Description label and control label

The label in SWF is divided into two types: Description label and control label:

The description label describes the content-shape, text, image, and sound of a SwF video. Each defined tag specifies a unique ID called "role ID" for its description. Flash Player stores these roles in a library called dictionary. The description label itself cannot drive the generation of an event.

The control label can create and drive role instances in the dictionary to control video playback.

SWF file tag Classification

Generally, the label in the SWF file can appear at will. However, a few rules must be followed:

1. A tag can only depend on the previously defined tag. No tag can depend on the tag defined later.

2. The description label must be defined before any control label that uses it.

3. stream sound labels must be stored in sequence. Irregular stream sound tags may cause abnormal sound playback.

4. The end tag is generally the last tag of the SWF file.

Dictionary

A dictionary is a database that stores defined roles. It can be applied by control labels. The establishment and use of dictionaries should follow the following rules:

1. The description label defines the shape, Font, bitmap, or sound.

2. Each description label is specified with a unique role ID.

3. The contents in the dictionary are stored after the role ID.

4. The control label can locate the desired content in the dictionary using the role ID and perform some operations on it, such as displaying the shape or playing sound.

A unique ID must be specified for each role ID. Duplicate IDs are not allowed. For example, the first role ID is 1, and the second role ID is 2. Role 0 is specified as the dedicated ID of the empty role.

Not only control tags can reference dictionaries. Description labels can also use data in the dictionary for more complex role definitions. For example, buttons and editing tags both use roles that define their content. Text labels also contain font roles to select different fonts for text.

* For details, see the display list.

Process SWF files

Flash players process all the tags in the SWF file before encountering the "display frame" label. In this way, the display list will be copied to the scenario, and the flash player will remain idle until the next frame is processed. The content of the first frame is the effect accumulated by all the control label operations before the first frame. The content of the second frame is the effect accumulated by all the control label operations before the second frame, and so on.

File compression Policy

As SWF files are frequently transmitted over the network, it is important to compress files as much as possible. There are several methods to achieve this purpose. The following methods are described:

The structure of the reusable role Dictionary makes it easy to reuse the elements of the SWF file. For example, a shape, button, Font, or bitmap can be stored and referenced multiple times.

Compression when the starting coordinate of a line is assumed to be the ending coordinate of the previous line, the shape can be compressed using a very effective triangular encoding method. The distance is usually expressed by the relative relationship with the last position.

Default matrix and color conversion types usually have more common fields than other structures. For example, for a matrix, the conversion field is the most commonly used field, while the scaling ratio and rotation are used less. Therefore, if the scaling field does not appear, the default value is 100%. If the rotation field does not appear, it is assumed that there is no rotation. Use the default value to reduce the file size.

According to the regulations, SWF files can only store changes between States. However, the shape data structure and position, movement, and moving model used in the display list are opposite.

Shape Data structure shape data structure uses a unique structure to reduce the file size and effectively play real-name shapes in the scene.

Summary

A swf file consists of a header and many labels. There are two types: Description label and control label. A description tag defines an object that is considered a role and is defined in the dictionary. Controls tag operations and roles, and controls the video process.