Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for speech synthesis. The methods, systems, and apparatus include actions of obtaining an audio representation of speech of a target speaker, obtaining input text for which speech is to be synthesized in a voice of the target speaker, generating a speaker vector by providing the audio representation to a speaker encoder engine that is trained to distinguish speakers from one another, generating an audio representation of the input text spoken in the voice of the target speaker by providing the input text and the speaker vector to a spectrogram generation engine that is trained using voices of reference speakers to generate audio representations, and providing the audio representation of the input text spoken in the voice of the target speaker for output.

Abstract: Different implementations are described, particularly implementations for processing a medium dynamic range video signal are presented. In a method for processing such a video signal a medium dynamic range video signal and associated metadata are received, wherein the metadata include data representative of a peak luminance value of the medium dynamic range video signal. In addition, data representative of a peak luminance value of a presentation display are received. It is determined whether the peak luminance value of the medium dynamic range video signal is greater or lower than the peak luminance value of the presentation display. A processor is configured based on the determination, wherein the processor has a first mode to reconstruct a high dynamic range video signal based on a received standard dynamic range video signal and associated metadata, and a second mode to optimize a received high dynamic range video signal for the rendering device.

Abstract: A method for decoding a picture block is disclosed. The decoding method comprises:—decoding (10) at least one stream S_diff into decoded data and into one information for identifying a re-constructed reference picture in a decoder picture buffer;—reconstructing (12) a special reference picture from at least the identified reconstructed reference picture and from the decoded data;—reconstructing (16) the picture block from at least the special reference picture, wherein the at least the special reference picture is not displayed.

Abstract: A method for decoding a picture block is disclosed. The decoding method comprises: —decoding (10) at least one stream S_diff into decoded data and into one information for identifying a reconstructed reference picture in a decoder picture buffer; —reconstructing (12) a special reference picture from at least the identified reconstructed reference picture and from the decoded data; —reconstructing (16) the picture block from at least the special reference picture, wherein the at least the special reference picture is not displayed.

Abstract: A method for decoding a picture block is disclosed. The decoding method comprises: —decoding (10) at least one stream S_diff into decoded data and into one information for identifying a reconstructed reference picture in a decoder picture buffer; —reconstructing (12) a special reference picture from at least the identified reconstructed reference picture and from the decoded data; —reconstructing (16) the picture block from at least the special reference picture, wherein the at least the special reference picture is not displayed.

Abstract: A method for decoding a picture block is disclosed. The decoding method comprises: reconstructing reference picture from another reference picture of a decoder picture buffer of a first layer of a multi-layered stream and from data decoded from a second layer of said multi-layered stream and storing said reference picture in a decoder picture buffer of said second layer, wherein said reference picture is indicated as not to be displayed; decoding a flag indicating that a subsequently decoded picture of the second layer is not using any inter-layer prediction; and reconstructing a picture block of said subsequently decoded picture at least from said reference picture.

Abstract: Different implementations are described, particularly implementations for processing a medium dynamic range video signal are presented. In a method for processing such a video signal a medium dynamic range video signal and associated metadata are received, wherein the metadata include data representative of a peak luminance value of the medium dynamic range video signal. In addition, data representative of a peak luminance value of a presentation display are received. It is determined whether the peak luminance value of the medium dynamic range video signal is greater or lower than the peak luminance value of the presentation display. A processor is configured based on the determination, wherein the processor has a first mode to reconstruct a high dynamic range video signal based on a received standard dynamic range video signal and associated metadata, and a second mode to optimize a received high dynamic range video signal for the rendering device.

Abstract: A medium dynamic range video signal and associated metadata are received, wherein the metadata include data representative of a peak luminance value of the signal. The medium dynamic range video signal is processed in a first mode to reconstruct a high dynamic range video signal based on a received standard dynamic range video signal and associated metadata if the peak luminance value of the medium dynamic range video signal is smaller than the peak luminance value of a presentation display, and in a second mode to optimize a received high dynamic range video signal for a rendering device if the peak luminance value of the medium dynamic range video signal is greater than the peak luminance value of the presentation display.

Abstract: A medium dynamic range video signal and associated metadata are received, wherein the metadata include data representative of a peak luminance value of the signal. The medium dynamic range video signal is processed in a first mode to reconstruct a high dynamic range video signal based on a received standard dynamic range video signal and associated metadata if the peak luminance value of the medium dynamic range video signal is smaller than the peak luminance value of a presentation display, and in a second mode to optimize a received high dynamic range video signal for a rendering device if the peak luminance value of the medium dynamic range video signal is greater than the peak luminance value of the presentation display.

Abstract: A system and methods is provided for performing, by a processor, an initial tone mapping function on a first image to produce a second image. The first and second images are of different dynamic ranges. The method also includes applying, by the processor, a second tone mapping function to the second image when it contains any saturated portions. The second tone mapping function lowers luminance by a first fraction. The overall luminance of the second image is also determined and a fine tuning filter is applied when at least portions of said second image contain areas darker than an amount.

Abstract: A method and device of encoding both a HDR picture (IHDR) and a first SDR picture (ISDR1) obtained from said HDR picture, in at least one bitstream (F1, F2, F3, F4). The method comprises: —obtaining (210) a second SDR picture (ISDR2) by tone-mapping the HDR picture (IHDR); —obtaining (230) a color mapping function (CMF) that allows the mapping of the colors of the second SDR picture (ISDR2) onto the colors of a third SDR picture (ISDR3) obtained (220) from the first SDR picture (ISDR1); —encoding (240), in a bitstream, an information (INF) representative of the color mapping function; and —encoding (260), in a bitstream, a fourth SDR picture (ISDR4) obtained (250) from the first SDR picture (ISDR1). The present disclosure further relates to a method and device of decoding.

Abstract: The present disclosure generally relates to a method and device for obtaining three color difference components from three components C1, C2 and C3 representing colors picture data in an input color space, a first of said color difference components being defined from the component C1, a second of said color difference components being defined from the first color difference component and the component C2, and a third of said color difference components being defined from the first color difference component and the component C3. The method is characterized in that it further comprises: —obtaining a first component by applying a process to the component C1; —obtaining said first color difference component by applying the inverse process to the first component.

Abstract: The invention concerns scaling color difference components prior to encoding in order to extend the dynamic range. In one embodiment the color difference components are encoded prior to quantization. In another embodiment the color difference components are encoded after quantization. A decoder is configured to process data representative of the gain factor for a decoding process.

Abstract: A method for decoding a picture block is disclosed. The decoding method comprises: decoding decoded data, and information for identifying a reconstructed reference picture in a decoder picture buffer; reconstructing another reference picture from the identified reconstructed reference picture identified by the information and from the decoded data; replacing in the decoder picture buffer the reconstructed reference picture identified by the information with the another reference picture; and reconstructing the picture block from the another reference picture.

Abstract: A method for encoding a plenoptic image divided into blocks is disclosed. The method comprises: —encoding (S120) a block in a central region of the plenoptic image, called central block, without spatial prediction; and —predictively encoding (S130) blocks immediately surrounding said central block by spatially interleaving spatial unidirectional prediction and spatial bi-directional prediction.

Abstract: Encoding or decoding a stack of images of a same scene focused at different focalization distances from one image to another can involve encoding or decoding information representing an image of the stack of images, where the image meets an image sharpness criterion, reconstructing the image into a reconstructed image, and encoding or decoding at least one other image of the stack of images by prediction from at least the reconstructed image.

Abstract: A method for encoding a current focal stack is disclosed that comprises a set of images focused at a different focalization distance from one image to another. According to present disclosure, the method comprises: —encoding (31) information representing an image of the current focal stack, the image being selected in said current focal stack according to an image sharpness criterion, and reconstructing the image into a reconstructed image; —encoding (32) at least another image of the current focal stack by prediction from at least the reconstructed image.

Abstract: A method for coding an HDR picture is disclosed that includes mapping the HDR picture to obtain a SDR picture, determining color correcting parameters, color correcting said SDR picture responsive to said color correcting parameters, and encoding the color corrected SDR picture in a stream. In the method, determining color correcting parameters includes determining an intermediate SDR picture so that at least one color appearance value is preserved between said HDR picture and said intermediate SDR picture, and determining the color correcting parameters from said intermediate SDR picture and from said SDR picture.

Abstract: A system and methods is provided for performing, by a processor, an initial tone mapping function on a first image to produce a second image. The first and second images are of different dynamic ranges. The method also includes applying, by the processor, a second tone mapping function to the second image when it contains any saturated portions. The second tone mapping function lowers luminance by a first fraction. The overall luminance of the second image is also determined and a fine tuning filter is applied when at least portions of said second image contain areas darker than an amount.

Abstract: A method for encoding and decoding, a scalar integer into at least one parameter representative of a pivot point comprised in a set of pivot points representative of a piece-wise linear function, said scalar integer being used when modifying pixel values of a picture, and corresponding apparatus, are disclosed. Said encoding method comprises:—a step E1 of selecting a subset of pivot points from the set of pivot points, according to a criterion, said subset comprising a number M of pivot points less than the number N of pivot points of said set of pivot points, —at least one step of coding E2 one bit of a binary representation of said scalar integer in a least significant bit of a component of a pivot point comprised into said subset.