Multi-pass encoding

Two - pass encoding, general multi- pass encoding is a compression method by which an audio or video file using a variable data rate (VBR ) can be encoded with optimum utilization of the given storage capacity.

Motivation

In compression process ( streaming, single-pass or one-pass encoding), in which the multimedia material is compressed in a single operation, there is the drawback that the complexity distribution throughout the time the material is not known, that is, the achievable compression rates are unknown. There are two methods can deal with it as the encoder, both at a significant disadvantage.

In the first variant, a quality of the encoder is determined satisfied, irrespective of the compression ratio achieved. This procedure is also known as variable bit rate encoding ( VBR). This has the disadvantage that the total amount of data recorded may be any that is a storage medium of fixed size, such as a data disc, is not fully utilized or not at all sufficient.

Therefore, a fixed amount of data per time period ( second or frame) is given in the second approach, the encoder, herewith, a desired total size, eg 680 MB will be achieved as a CD volume size. This procedure is also known as Constant Bit Rate encoding (CBR ). Disadvantage of this approach is that the quality achieved is not constant. This results in CBR, therefore, that for example with video material, simple video scenes (dark or with little movement ) too much and complex video scenes (rapidly changing ) too little space is given - other than for a consistent quality required.

Method

The two- pass encoding method now tries to combine the qualities of both single-pass method described above. The idea of ​​this is to perform a more efficient selection of the amount of data per time segment with additional information about the complexity of the material. For this purpose, a prior analysis compression run is performed. The encoder determines in the first run for each time period ( for videos about per image ), the data rate at optimum quality, and stores this information as a measure of complexity between temporarily. The compression result (e.g., the compressed video ) itself is not stored.

In the second pass, the encoder then uses the predetermined target data amount and the cached complexity measures from the first pass. He now compressed each time period with a data rate proportional to the determined complexity measure. The goal of a constant quality at a given target data set is typically good, with only slight variations, is reached.

This process can be described as iterative procedure with more than two passes expand (Multi- pass), with an even better homogenization of the quality can be achieved over time, or a more precise target dataset.

Advantages:

• With the optimum utilization of the available space ( such as a CD -ROM), the best average quality is achieved.

Cons:

• It is about twice the processing time as in the one- pass encoding is required.
• This method can not be used in live / sent in real-time audio-video data because of the important second passage can be ( ie at the end of the program ) is executed only if all the data.

History

In the video section Nandub was one of the first video compression programs, which dominated this technique, known there as Smart Bitrate Control ( SBC).

• Communications Engineering