Entropy coding information delivery

2014-02 Delivery of Probabilities - PATENT FAMILY FOR SALE

This patented method enables efficient delivery of range coding frequencies, probabilities or range values. The method enables range coding to be utilized in cases where the amount of symbols to be coded is not very large compared to the amount of possible symbols. With known range coding techniques, it is not feasible to encode such data due to the amount of bits required to deliver the frequencies, probabilities or range values, without accumulating an impractical data overhead associated with such a delivery of frequency, probability or range information from an encoder to a corresponding decoder. This improved method uses an availability signal and a probability value to convey that a given symbol exists, and an indication of an associated probability of the given symbol, representing a considerable technical advance in data encoding technology and resulting in data size savings. This solution can produce a more accurate probability table than prior art solutions, thereby achieving entropy that is very nearly optimal.


The method is feasible to being used in data processing and data generating devices, for example in personal computers, portable devices such as smart telephones, digital cameras, video cameras, televisions, surveillance apparatus etc. The encoder and the decoder are implemented as a codec, to be used for example in portable signal capture devices, such as digital cameras, internet-capable multimedia devices, game consoles, environmental monitoring systems etc.

2014-02 Efficient Codetable Delivery - PATENT FAMILY FOR SALE

This patented method enables efficient delivery of code tables or frequency tables. It reduces a data communication and/or data storage overhead needed for table delivery and/or storage. The method enables a large data file to be split more efficiently to smaller portions (for example using our patented block coding technology), i.e. into data chunks, with a benefit that the delivery of code table or frequency table can be optimized for each individual data chunk. The method is capable of generating a variety of different tables based on the characteristics and properties of the data, and thereafter delivering them efficiently.

The split of the big data file enables substantial benefits in respect of modifying entropy of data involved, and so it is capable of highly reducing an amount of encoded data to be communicated. Thus, increased greater compression efficiency can be achieved, which means that the data storage capacity, the transmission bandwidth and the energy consumption can be reduced.

2015-06 Palette Compression - PATENT FAMILY FOR SALE

Method of compressing data of the palette in a lossless manner into the encoded data and delivering the compressed palette data from an encoder to a corresponding decoder, wherein palette entry values of the palette are provided consecutively within the encoded data.

2015-11 Adaptive Probabilities - PATENT FAMILY FOR SALE

In this patented method, the encoder is operable to adapt the probabilities corresponding to remaining elements of the at least one sequence, namely elements that are yet to be encoded. Thus, the encoder pursuant to embodiments of this method operates unlike conventional encoders, which do not have knowledge of the probabilities, but make an assumption that symbols that have been delivered previously will be delivered often in the future as well, and thus, adaptively increase the probability of a symbol as elements having that symbol are being encoded. The aforementioned method of encoding the data sequence, as described above, progressively improves a compression ratio of data achieved in the encoded data by using more accurate probabilities of those symbols that occur at the end of the data sequence.