Find palette entry with minimum error from difference of actual pixel value and predicted pixel value. Propagate error of pixel to its top and left pi

Returns two ranges (<left, width> pairs) at row 'canvas_y', that belong to 'src' but not 'dst'. A point range is empty if the corresponding width is 0

Finalize and transmit a complete row. Return false in case of user-abort.

Finalize and transmit a complete row. Return false in case of user-abort.

Replace similar blocks of pixels by a 'see-through' transparent block with uniform average color. Assumes lossy compression is being used. Returns tru

Create an optimal Huffman tree. (data,length): population counts. tree_limit: maximum bit depth (inclusive) of the codes. bit_depths[]: how many bits

Create an optimal Huffman tree. (data,length): population counts. tree_limit: maximum bit depth (inclusive) of the codes. bit_depths[]: how many bits

Returns the tightest dimension for the canvas considering the image list.

Returns best predictor and updates the accumulated histogram. If max_quantization > 1, assumes that near lossless processing will be applied, quantizi

Get a reference to appropriate chunk list within an image given chunk tag.

Given the input 'filter' option, return an OR'd bit-set of filters to try.

Given the input 'filter' option, return an OR'd bit-set of filters to try.

Outputs image data given a bitstream. The bitstream can either be a single-image WebP file or raw VP8/VP8L data. Also outputs 'is_lossless' to be true

See section 13-2: http://tools.ietf.org/html/rfc6386#section-13.2

Returns the maximum number of hash chain lookups to do for a given compression quality. Return value in range [8, 86].

Returns the maximum number of hash chain lookups to do for a given compression quality. Return value in range [8, 86].

horizontal accumulation. We use mirror replication of missing pixels, as it's a little easier to implement (surprisingly).

horizontal accumulation. We use mirror replication of missing pixels, as it's a little easier to implement (surprisingly).

Performs out = a + b, computing the cost C(a+b) - C(a) - C(b) while comparing to the threshold value 'cost_threshold'. The score returned is Score = C

Same as HistogramAddEval(), except that the resulting histogram is not stored. Only the cost C(a+b) - C(a) is evaluated. We omit the term C(b) which i

Same as HistogramAddEval(), except that the resulting histogram is not stored. Only the cost C(a+b) - C(a) is evaluated. We omit the term C(b) which i

Partition histograms to different entropy bins for three dominant (literal, red and blue) symbol costs and compute the histogram aggregate bit_cost.

Partition histograms to different entropy bins for three dominant (literal, red and blue) symbol costs and compute the histogram aggregate bit_cost.

Construct the histograms from backward references.

Construct the histograms from backward references.

Merges some histograms with same bin_id together if it's advantageous. Sets the remaining histograms to NULL.

Combines histograms by continuously choosing the one with the highest cost reduction.

Combines histograms by continuously choosing the one with the highest cost reduction.

Compact the histogram set by removing unused entries.