What is video encoding? Part 3

MP 3

MP3 (MPEG-1 Audio Layer-3) is a standard technology and format for compressing an audio stream into a very small file (reduced by approximately 20 times the size of the original file) while maintaining the same level of quality as the file original audio. . MP 3 technology uses psychoacoustic modeling to reduce the size of the audio file and maintain a high level of sound quality.

MPEG
MPEG stands for Motion Picture Expert Group and MPEG stands for the generic name for video formats. A group of experts defines standards for digital video such as MPEG -1 (used in Video CD), MPEG -2 (used in DVD and SVCD), MPEG -4 (used in DivX video technology), as well as some audio standards, among others that MP 3 and AAC.

MPEG -4
MPEG-4 is a standard developed by the 11th MPEG Working Group of ISO (International Organization for Standardization) in October 1998 (the date of the first draft of the standard). MPEG-4 is the standard for the mature digital age. With its additional features, MPEG-4 offers better compression, interactivity, and versatile wireless data / Internet access.

Multipass encoding – multipass encoding
True multi-pass encoding is currently only available for WM8 and MPEG-2 (SVCD and miniDVD). The multi-pass encoder analyzes the video stream on the first pass and writes everything it finds to log files. Let’s say we have a shortcut that starts with a dialogue scene where we cut something out and the camera remains stationary. Then we move on to karate shots, with a host of rapidly changing scenes and actively developing action (people flying in the air, kicking, punching, etc.).

Usually with a constant bit rate the encoder receives a higher or lower bit rate every second (this can only be called a constant 100% bit rate, but these are details). While in multi-pass variable bit rate mode, the encoder will use the bit rate according to the knowledge of the video stream, that is, the dialogue scene will receive a dedicated bit rate and the fight scene it will be much greater.

The more passes there are, the more perfect the bitrate distribution will be. In a single variable bit rate pass, the encoder must base its decision on where and what bit rate to use solely on knowledge of the composition of its previous encoding.

PCM
Pulse code modulation is the simplest binary representation of digital audio. The audio signal is converted into samples (samples) corresponding to the frequency of the signal. Then each sample (sample) is written in sequence, without using heavy compression techniques.

What is video encoding? Part 2

Bit rate – bit rate

The bit rate is the bit rate of the data transmission, that is, the number of bits transmitted per unit of time, generally measured in bits per second. When encoding video, a distinction is made between video and audio bit rate; During encoding, each parameter is set to its own value and does not depend on the other.

Shine – shine
(1) Intensity of color, measured from black (zero brightness) to white (maximum brightness).

CBR / VBR constant / variable bit rate
Constant Bit Rate / Variable Bit Rate: Constant / Variable Bit Rate. With a constant bit rate, the same number of bits is allocated for each frame of the movie. With a variable bit rate, frames that require better quality get more bits and vice versa. Note that dynamic movie scenes require more bits than smooth scenes.

Codec -codec
COder / DECoder – COder-DECoder (short codec): A codec is a piece of software that allows you to encode data (usually audio or video) in a specific format, and it can also decode data encoded in this format. Popular codecs: MPEG -1, MPEG -2, MPEG -4, Indeo, etc.

AVI, ASF and others are not codecs, but data formats that can be obtained using codecs.

Composite video – composite video
This is a signal in which luminance, chrominance, and timing are combined into a single signal used in the television broadcast standard.

Decoding – decoding
The term Decoding describes the process of converting a compressed (encoded) file into a viewable image.

Deinterlace – deinterlace (remove interlacing)
It is the process of removing artifacts caused by the nature of interlaced video (two fields per frame). Otherwise, it is called the “comb effect.”

Encoding – encoding
It is the process of compressing a “raw” uncompressed file by encoding it in a given format while maintaining a certain quality.

FilmFX
This post-processing algorithm adds “warmth” to video for those users who like warm tones rather than the clarity of digital video. The FilmFx filter is best suited not only for adding warmth to a movie, but also for reducing perceived blockage in digital video and decreasing processor load during decoding.

Frame – frame
This is the basis of the entire film, one frame represents one image. Film generally runs at 24, 25, or 30 frames per second, which is equivalent to displaying 24 (25 or 30) images per second. Imagine 24 images of one bird each. In the first image, the bird is on the left, in each subsequent image, it gradually shifts to the right, in the last image, the bird is on the right edge. When these 24 images are displayed in sequence very quickly, the human eye will see the bird flying from left to right.

Global motion compensation: general motion compensation
Global Motion Compensation (GMC) – Global Motion Compensation helps enhance complex scenes including panning and zooming. The ability to reduce the amount of data from one frame to the next may be diminished as there are some similarities between panning and zooming an image. To more effectively compensate for movement, this similarity can be exploited in the groups of blocks found most frequently in such scenes.

Keyframe – Keyframe
Full frame, but highly compressed (compressed).

What is video encoding?

I suppose I will not be wrong to say that this is a rather delicate and complex process, fraught with many difficulties and problems.

Well, first of all, you need to understand clearly enough for yourself that the video encoding process, in any scenario of the main and accompanying factors, always leads to an overall decrease in the quality of the video stream. But the whole question and all the excitement, so to speak, is how much will we lose in quality and whether our loss will be proportional to the size of the resulting video file. In general, we try to repackage the video in such a way that with a slight decrease in quality, we get a visible gain in size.

Now let’s move on to more “mundane” things … “It is known for sure” 🙂 that any video sequence is a sequence of images that alternate at a certain speed, the so-called FrameRate (frame rate). Each image has a set of certain parameters such as: frame size (FrameSize), color depth, brightness, contrast, etc. All these parameters are very important and leave a serious mark on the final video. Unfortunately, video quality is a subjective factor in our perception of this or that video stream, for each person the line between acceptable quality and disgusting “screen” is different. In this regard, I would like to point out that although the quality of video material is an important benchmark in the world of encoding, it unfortunately does not have strictly regulated characteristics.

Now in a nutshell, directly about the process itself …

As an example, let’s take the well-known and fairly easy-to-use DrDivX program. Let’s start our debriefing with the terminology …

AC3
Also known as Dolby Digital from Dolby LabsTM. It can support up to 5.1 channels of audio.

AVI
Audio Video Interlacing – Audio Video Interlacing (joining them together) – This is a video and video format commonly used on Windows operating system computers. This format is defined as a kind of combination of audio and video data, without specifying any specific codec.

B-frames / bidirectional encoding
There are three types of frames that are possible in a DivX video stream. These frames are called I-frames (intra), P-frames (predicted), and B-frames (bidirectional). Before the release of the DivX 5.0 codec, only I and P-frames were used. I-frames are encoded using information only from the encoded frame itself and do not use information from other frames (time compression). I-frames are based on conventional single-frame compression to the JPEG format. P (predicted) frames predict next frames and can also refer to I or P frames, that is, E. P frames are encoded using information from previous frames. In any video sequence, there will always be a group of frames, many of which will be the same and will contain the same image. For example, if you are watching news, and you pay attention to any moving character, you may notice that for several frames the background behind him almost always remains unchanged. (Remember that normally the frames are rotated at 25 or 30 frames per second). So instead of encoding each frame in JPEG format independently, you can take advantage of the redundancy of previous frames by applying P-frames. Basically, P-frames are future frames, defining how a block from a previous frame has been moved. at the current P-frame. So instead of spatially encoding the frame, the P-frame just says “Hey block, in the frame above, move to point (X, Y)”. This time encoding algorithm requires much less data than the spatial encoding of each frame. Basically we are passing the difference between adjacent frames, which is more efficient,