FR3083414A1 - AUDIOVISUAL MOUNTING PROCESS - Google Patents

Abstract

The invention relates to an audiovisual editing method comprising; - a step (100) of entering a plurality of data flows, - a step (105) of handling a single work flow formed from at least one said data flow from the previous step , in an expression space targeted in three spatial dimensions plus time, during this step, a virtual camera is used which has: either ▪ a first ability to perform rotations along the three axes, a step (130) of management of a 360 ° environment and a step (135) for calculating intermediate positions; ▪ a second ability to move on the three axes, a step (130) for managing a three-dimensional environment using at least one planar object, and a step (135) for calculating intermediate positions.

Description

TECHNICAL FIELD OF THE INVENTION

The present invention relates to an audiovisual editing method.

It relates more precisely to the field of video editing. In known manner, video editing consists of producing the visual component of a film from a plurality of audiovisual content.

STATE OF THE ART

Numerous audiovisual editing programs are known which are suitable for editing two-dimensional images.

One of the means that can be used to obtain a multi-point auto-stereoscopic view media consists in using software known as “compositing”, such as for example “After Effect”, from Adobe (registered trademarks).

These tools have many drawbacks:

- the manual editing, by the editor, of the geometric consistency of the two-dimensional image streams and the management of colors to obtain a good spectator rendering. It is the editor who must ensure that any modification of space-time placement, color of a media on one of the data streams, colorimetric and sound retouching. Mastering the software is therefore essential, which is not possible for someone who does not have real notions of editing and compositing,

- the operations required are very redundant and tedious and delivery times to customers are therefore very high.

- with the editing tools usually designated by two dimensions plus time, such as "Premiere Pro" software from Adobe (registered trademarks), are generally obtained by artifices or by projection of two-dimensional objects with scenarios predefined in three dimensions plus time. The possibilities to modify colors or manage a 3D environment is impossible or complex,

- editing tools on two spatial dimensions over time, do not allow 3D mapping of a panorama in a sphere, it is therefore not possible to make video rushes (defined below) from panoramic content.

- Editing tools using 3 dimensions plus time (Like After Effects) are often very complex to use because it is often necessary to carry out the definition of the keys yourself without being constrained by rules. It is just as necessary to have knowledge of texture management. It therefore requires many successive actions by the editor and often many adjustments to obtain the desired result.

In general, editing in two dimensions plus time is therefore tedious and does not take into account the visual characteristics of three-dimensional images or change colors. The transformation of colors on software of this type is possible but again complex and tedious. The use of design made with non-primary colors would not allow the colors of our users to be obtained, but it is possible to get closer with some of their tools. However, only the human eye can then define whether the color is conforming or not.

There is the Pre-calculated rendering which allows to obtain the best possible results by calculating all the elements of the scene (3D elements, lights, textures). For each position of the camera, the calculation will be carried out in exactly the same way and therefore within the same time. This type of rendering is widely used in cinema because it allows you to get as close as possible to a so-called “Photo-Realistic” quality (as realistic as possible and close to what you could get by a photo in reality ). However it is very time consuming, indeed a single image often takes several hours of calculation, knowing that a second of video in a conventional system contains 30 images per second, an animation of about 4 seconds requires 120 images, rendering this animation will therefore take at least 2 days of calculation. It was therefore not conceivable that we would choose this type of solution, too costly in time.

It is noted, as of now, that the invention relates in particular to images which are physical entities and / or representative of physical phenomena or of supervision of industrial processes.

OBJECT OF THE INVENTION

The present invention aims to remedy these drawbacks.

To this end, the present invention relates to an audiovisual editing method remarkable in that it comprises:

- a) a step of entering a plurality of data streams, at least one said data stream representing an image, a panorama, a video or made from a 360 ° image,

- b) a step of manipulating a single workflow formed from at least one said data stream from step a), in an expression space targeted in three spatial dimensions over time, during from this step a virtual camera is used which has:

either a first ability to perform rotations along the three axes, said three axes being horizontal, vertical and depth, and is fixed on said three axes, • a step of managing a 360 ° environment using the first capacity of the camera virtual and a data flow in three spatial dimensions from a panorama or a 360 ° video, during this step of managing a 360 ° environment, a point of depat P1 is defined ( x1, y1 and z1), a point of arrival P2 (x2, y2 and z2) on said flow in three spatial dimensions, and a time, • a step of calculating intermediate positions, corresponding to different viewpoints of the space in three spatial dimensions plus the time between the starting point and the point of arrival on the flow in three spatial dimensions.

Or a second ability to move on the three axes, horizontal, vertical and depth and is fixed on the rotations of said three axes, • a step of managing a three-dimensional environment using at least one planar object using the second ability to the virtual camera and a two-dimensional spatial data flow from a video or an image, during this step of managing a three-dimensional environment using at least one planar object, a starting point P1 (x1, y1 and z1), an arrival point P2 (x2, y2 and z2) on said flow in two spatial dimensions, and a time, • a step of calculating intermediate positions, corresponding to different points view of space in three spatial dimensions plus the time between the starting point and the point of arrival on the flow in two spatial dimensions.

The advantage is that we can recreate the environment in 360 ° without distorting the image. So we can generate a quality video in which our users can choose the most interesting positions. Some of our customers choose these plans carefully so as not to show unwanted items.

The calculation of the intermediate positions allows the fluidity of the animation. Whether on a linear or damped animation, the management of intermediate positions guarantees the quality of the final video. The advantage of linear animation and that it retains the dynamic, animation with amortized it retains a certain aesthetic at the beginning or at the end of the video sequence. The intermediate sequences are therefore always linear.

Regarding the management of 360 panoramas, we are mapping on a sphere because it is the 3D element which represents our environment at 360 °. The basis of a sphere is that all the points of this sphere are equidistant from the central point, and in our case from the camera sensor. The creation of a spherical panorama is achieved by assembling multiple photos from different points of view but whose common point is the origin in physical space. Each shot is taken in exactly the same place except at a different angle.

Regarding the management of an image, it is a single point of view, and this point of view already contains in part deformations related to the lens of the camera. It is therefore not necessary to add a deformation to it. And so do not touch the focal length because it will alter the initial focal length, the one used when shooting. So we are using a camera with a standard focal length. Since it is a photo, it is therefore flat or the environment in three dimensions using at least one planar object.

The invention is advantageously implemented according to the embodiments and the variants set out below, which are to be considered individually or according to any technically operative combination.

In one embodiment, for the first capacity which comprises during the step of calculating the intermediate positions either an identical difference between the different positions of the flow in three spatial dimensions or an increase in the number of intermediate positions near the point of departure and arrival point.

In one embodiment, for the second capacity which comprises during the step of calculating the intermediate positions either an identical difference between the different positions of the flow in three spatial dimensions or an increase in the number of intermediate positions close to the point of departure and arrival point.

In one embodiment, the method comprises:

- a step of storing the data flow on a remote server,

a rendering step from the data stored remotely and information from the starting point P1, the finishing point P2 and the time, said rendering step processes the calculation of each image on the server creating a graphical environment according to the OpenGL specification (registered trademark).

So the solution is real-time rendering which allows rendering almost immediately, that is to say at least 30 images per second. This is a type of rendering that allows for smooth and very fast rendering. By using this type of rendering, we save a lot of time allowing us to offer a quality service.

In one embodiment, the method comprises:

a step for manipulating a first color defined by three numerical values between 0 and 255 in the RGB color space corresponds respectively to the color Red, Green and Blue, the first value being defined as being Rci, the second value being defined as Vci, the third value being defined as Bci,

- a step of calculating the proportion of red, green and blue from the numerical values Rci, Vci, and Bci for a first color resulting from step b) according to the following matrix M1:

Ren

2sW ^{C [0]}

M2 _{= Vc [0]}

M12 _{= file [0]}

VCI

2T5 = ^RCÎ1]

Ben

255 = ^{flC [2]} (1 - WD (1 ~ /? C [l]) = Vc [l] = Bc [l] (1 - M2]) (1 - 7? C [2]) = Vc [2 ] = Bc [2]

a step of applying the values calculated in step c) to a data stream to be colored, at least one said data stream to be colored representing an image or an expression, according to a colorimetric filter constituted as follows:

color filter = Rf: Vf: Bf with Rf being defined by the values of the matrix Rc [0]: Vc [0]: Bc [0], Vf being defined by the values of the matrix Rc [1]: Vc [1 ]: Bc [1], and Bf being defined by the values of the matrix Rc [2]: Vc [2]: Bc [2], the data flow to be colored becoming the colorized data flow,

- a step of generating the colorized data flow over at least one of the data flows of the manipulation step of a first color.

Thanks to these provisions, it is possible to edit any type of media from different sources (Video, images and virtual tour) without having any specific skills. There is a saving of time and many possibilities.

The advantage of using only one color is that the graphic consistency is respected, the main color of our user is kept without accompanying colors that potentially do not go together.

The color filter is written as follows color filter = Rc [0]: Vc [0]: Bc [0]: Rc [1]: Vc [1]: Bc [1]: Rc [2]: Vc [ 2]: Bc [2],

The colorized image is for example a banner on which an indication is mentioned. According to an example, it is possible to indicate on images representing a property to locate the interior of the exterior. The font of this expression (inside or outside) can also be colorized to go with the graphic consistency of the entire video.

In one embodiment, the method comprises in the step of handling a first color, if a numerical value of a color is given in hexadecimal composed of six digits stored on three bytes, said value is transformed into three numerical values in the RGB color space.

The most common color notation on web browsers is hexadecimal. Companies creating graphic charters also always provide this data, whether hexadecimal or RGB. It is therefore relatively easy for our customers to find the information. There are also many plugins on browsers allowing pipetting in a photo of the color of a logo for example. The results are also given in general in hexadecimal and RGB.

In one embodiment, the method comprises:

the manipulation step comprises a second color defined by three numerical values between 0 and 255 in the RGB color space corresponding respectively to the color Red, Green and Blue, the first value being defined as being Rû2, the second value being defined as Vc ₂ , the third value being defined as Bc ₂ ,

the step of calculating the proportion is replaced by a step of calculating the proportion of red, green and blue from the numerical values King, Vci, Bci, Rc ₂ , Vc ₂ , and Bc ₂ for a first color and a second color from step b) according to the following matrix M2:

Bq 7q Bqί _ ₌ M0] - = M11 ^ 5 = M2] ^Rc 2 r π ^VC 2 r π ^SC 2r Π - = M0] ^ = ^ [1] - = M2] .1 - Bc [0] - 7c [ 0] = Bc [0] l-Bc [l] -7c [l] = Bc [l] 1 - Bc [2] - Vc [2] = Bc [2] _ The use of two colors allows greater variety in templates. Indeed, a graphic charter is often composed of two or three main colors which are coherent. The user can then improve the quality of the skins using these two colors and adapt the video rendering to his own colors.

In an embodiment linked to the previous one, the method comprises:

- the manipulation step includes a fourth alpha numerical value between 0 and 1, the fourth value being defined as Ac,

- the matrix M1 of the proportion calculation step is replaced by the following matrix:

rc,

25W ^{C [O]} (1 -Æc [0])

--- = MO] = Ac [0]

VCI

2 ^ = ^{RC [1]} (1 - 7? C [l])

--- ^ - ^ = ^ [1] = Ac [l]

Bci

2T5 = ^RCÎ2]

2 ^ 2 _{= Bcl2]} = Ac [2] = Æc [3] = Vc [3] = Bc [3] = Ac [3] _ and the matrix M2 of the proportion calculation step is replaced by the matrix next :

Bq _r, Vq - _BC = [O] - = Rc [i] ^^ ^^ -1 2 r 2 r = Κ Ί _ _{£ [°]} _ _C = K _[1]

Bq r 1 ^ s ^{= Be [21}

Bc _{2 r Ί}

255 = ^{VC [2]} = AEC [3] = Vc [3]

1-7? C [0] -Ec [0] = Bc [0] 1-Æc [l] - Ec [l] = Bc [l] 1 - 7? C [2] - Vc [2] = Bc [ 2] 0 = Bc [3] 0 = 4c [0] 0 = 4c [l] 0 = Ac [2] 1 = 4c [3].

- the colorimetric filter of step d) is completed with the values of Ac, ie:

color filter = Rf: Vf: Bf: Af with Rf being defined by the values of the matrix Rc [0]: Vc [0]: Bc [0]: Ac [0], Vf being defined by the values of the matrix Rc [1]: Vc [1]: Bc [1]: Ac [1], Bf being defined by the values of the matrix Rc [2]: Vc [2]: Bc [2]: Ac [2], Af being defined by the values of the matrix Rc [3]: Vc [3]: Bc [3]: Ac [3].

This fourth value defines transparency in an image. Most of the time, the template elements contain transparency. If a template is 200x200 pixels, this entire area would become opaque, and therefore all the elements behind would be invisible. Besides, it is important to keep maximum visibility on the media.

BRIEF DESCRIPTION OF THE FIGURES

Other advantages, aims and characteristics of the present invention emerge from the description which follows, given for explanatory purposes and in no way limiting, with reference to the appended drawings, in which:

FIG. 1 represents, in the form of a flow diagram, steps implemented in a particular embodiment of the method which is the subject of the present invention,

FIG. 2 represents, in the form of a flow diagram, steps implemented in a second particular embodiment of the method which is the subject of the present invention,

FIG. 3 represents a diagram with elements of a particular embodiment of the audiovisual editing method which is the subject of the present invention,

FIG. 4 represents another diagram with elements of a particular embodiment of the audiovisual editing method which is the subject of the present invention,

- Figure 5 shows another diagram with elements of a particular embodiment of the audiovisual editing method object of the present invention.

DEFINITIONS

Throughout the description, terms are used, the definition of which is given below:

"3D": having, or perceived as having, three spatial dimensions (horizontal, vertical, depth).

"Color channel mixer >>: for a color channel mixer in French, which is a filter allowing the modification of colors using a" Focal distance "matrix: The focal distance defines the distance between the focal point and the camera lens . The focal point is the point of convergence of the light rays directed from the lens. Changing this distance allows you to change the appearance of the image and zoom in or out.

“FFMPEG >>: collection of free software intended for the processing of video and audio streams (recording, playback or conversion from one format to another).

"Image": visual or mental representation of an entity; from the Latin "imago", which means "portrait", "representation", "effigy", and which designated the death masks; by abuse of language, the term image is commonly used to speak of a 2D image.

"2D image", "planar image", or "flat image": image defined in a two-dimensional space.

"Panorama": represents a width view of a physical space. The 360 ° panorama is distinguished by the representation of the entire physical space from the point of view.

"Rush" (English term for original media in French): Original document or data stream output directly from the camera.

"Rendering": production of an image based on a data flow.

"Sequence of images": temporal succession (with a more or less fixed temporal spacing) of images stored in independent files.

"Templates": in French, models mean models comprising several design elements linked by the same graphic universe.

"360 ° video": a 360 ° video allows you to see from all points of view the physical environment surrounding the cameraman.

The verb form "may", as well as its derived forms define an option.

DESCRIPTION OF EXAMPLES OF EMBODIMENT OF THE INVENTION

As illustrated in FIG. 1, in a first particular embodiment, the audiovisual editing method which is the subject of the present invention comprises:

a) a step (100) of entering a plurality of data streams, at least one said data stream representing an image, a panorama, a video or produced from a 360 ° image,

- b) a step (105) of manipulation of a first color defined by three numerical values between 0 and 255 in the RGB color space corresponds respectively to the color Red, Green and Blue, the first value being defined as being Rci , the second value being defined as Vci, the third value being defined as Bci,

- c) a step (110) of calculating the proportion of red, green and blue from the numerical values Rci, Vci, and Bci for a first color resulting from step b) according to the following matrix M1:

Rex

25W ^{C [0]}

VCI

2T5 = ^RCÎ1]

Bex

2sM ^[2] (1-MW

-— = Mi] (i-Mi])

-— = Mi] (1 - “C [2D (1 - 7? C [2]) = Vc [2] = Bc [2]

- d) a step (115) of applying the values calculated in step c) to a data stream to be colored, at least one said data stream to be colored representing an image or an expression, according to a colorimetric filter consisting of as follows:

color filter = Rf: Vf: Bf with Rf being defined by the values of the matrix Rc [0]: Vc [0]: Bc [0], Vf being defined by the values of the matrix Rc [1]: Vc [1 ]: Bc [1], and Bf being defined by the values of the matrix Rc [2]: Vc [2]: Bc [2], the data stream to be colored becoming the colorized data stream,

- e) a step (120) of generating the colorized data flow over at least one of the data flows of step a).

In a variant, the method comprises a step of storing the data stream on a remote server.

In another variant, the method includes a step of rendering remotely from user data. The remote rendering server processes the calculation of each image via a graphical environment created according to the OpenGL specification. A colored data stream can then be applied over this newly created stream.

Indeed, various known libraries conforming to OpenGL can be used.

In another variant, the method includes a step of adding a soundtrack selected from a list.

For step b), if a numeric value of a color is given in hexadecimal composed of six digits stored on three bytes, said value is transformed into three numeric values in the RGB color space.

In another example, the manipulation step b) (105) comprises a second color defined by three numerical values between 0 and 255 in the RGB color space corresponding respectively to the color Red, Green and Blue, the first value being defined as Rû2, the second value being defined as Vc ₂ , the third value being defined as Bc ₂ , step c) is then replaced by a step of calculating the proportion of red, green and blue from numerical values Rci, Vci, Bci, Rc ₂ , Vc ₂ , and Bc ₂ for a first color and a second color from step b) according to the following matrix M2:

ffCiM _ ₌ ^R ₄ o] - =

Rc _{2 r} , Vc - _{= M} o] ^ = r ^BC 1 r 1

255 = ^RCP1

Bc ₂

255 = ^ 2] ll-Æc [0] -7c [0] = 5c [0] 1-Lc [l] = Bc [l] 1 - ffc [2] - Vc [2] = Bc [2] J

In another example, the manipulation step b) (105) includes a fourth alpha numerical value between 0 and 1, the fourth value being defined as Ac, the matrix M1 of step c) is replaced by the matrix next :

Γ R ^c i

255 = * 40] MM = ^ 40] (1 - 7? C [0])

---- 2 “^ = ^{Sc [0]} = Ac [0] and the matrix M2 of step c) is replaced by the following matrix:

Γ ^Rc ir 1 Mr-. ^Bc i Γ 1 - = = 2SS ^{= RC} [ ² 1

Rc ₇ Vc ₇ Bc ₇

255 = ^ 01 = O = v _e [ ₃ l

1-Æc [0] - Vc [0] = Bc [0] 1-7? C [l] - ùc [l] = Bc [l] 1 - 7? C [2] - ùc [2] = Bc [ 2] 0 = Bc [3] 0 = Ac [0] 0 = Ac [l] 0 = Ac [2] 1 = Ac [3] _

- the colorimetric filter of step d) is completed with the values of Ac, ie:

Vc.

Μ ^ΦΙ

MM =

--- γ ^ = Βο [1] = Ac [l]

Bci

2sM ^[2]

MM

MM = * 42] = Ac [2] = Æc [3] = M [3] = Bc [3] = Ac [3] _ = 7? C [3] colorimetric filter = Rf: Vf: Bf: Af with Rf being defined of the matrix of the matrix of the matrix by by by by the values the values the values the values

Rc [0]: Vc [0]: Bc [0]: Ac [0], Vf being defined

Rc [1]: Vc [1]: Bc [1]: Ac [1], Bf being defined

Rc [2]: Vc [2]: Bc [2]: Ac [2], Af being defined from the matrix Rc [3]: Vc [3]: Bc [3]: Ac [3].

Colorization allows a very important qualitative contribution. It is possible to make almost all the possible animations by coloring them. We therefore combine a customizable skin for optimal rendering.

Conventionally, colors are defined via the color space (Color representation in a color synthesis system) RGBA (Acronym for Red Green Blue Alpha, for RVBA red green blue blue alpha in French is a computer coding system for colors , the closest to the material and including the opacity rate), comprising four channels, each of which can include a value between 0 and 255. Computer screens reconstruct a color by additive synthesis from three primary colors, one red, a green and a blue.

We also manage the use of the hexadecimal triplet (6-digit hexadecimal number stored in three bytes). Each byte of the triplet corresponds to an RGB hue, the alpha is not managed. For example, if we take the color # B8202F, we can break it down like this:

1. B8 -> 184

2. 20 -> 32

3.2F -> 47

The translation of # B8202F is therefore RVBA (184, 32, 47, 1), the alpha being in our case always 100% opaque. So regardless of the input color space, we always convert to RGBA notation.

We will via a matrix and quantities of colors in each channel transpose the colorimetric values of our users using an image or an expression, of template type including the primary colors. For example banners which are positioned on a part of the video.

For this, we have adapted the Color channel mixer present in the library of the FFMPEG software. This library allows the transformation of colors, so we developed a color transposition calculation system which was then automated. The user's colors reflect a visual identity. It is therefore necessary that the colors are perfectly identical on request. The calculation is very precise.

For the transformation to take place, the template video files must be made with so-called primary colors so that these colors can then be modified.

At first, we want to perform a transformation to a color, we will use the Red Channel, or even Red layer.

In addition to the colors to be transformed, our templates can contain shades of gray (from white to black), the values in each channel must be equal. The objective is therefore to be able to transform the color red to the color of our user without transforming the shades of gray. We can therefore integrate the graphic elements with a background including for example a shading effect, this one not being impacted. If the notions of gray are slightly tinged, then they are also subject to modification.

Example for a color:

A user selects the RGBA color (80, 136, 207, 1) which we place in the variable C1. This variable is seen as an array composed of four values.

R V B AT C1 80 138 207 1

Transposition requires a calculation using each value in the table. Each of them is divided by 255, the maximum value it can reach in this color space.

Matrix showing primary colors:

Red Green Blue Alpha Red primary 255 0 0 0 Green primary 0 255 0 0 Blue primary 0 0 255

Primary red is a color that could be described as maximum red, that is, its color channels are composed of the maximum value in the Red channel and the minimum in the other channels. The other two primary colors are formed in the same way. The Alpha channel is special because it is a proportion ranging from 0 to 1 (0 indicating the minimum opacity and 1 maximum). Our templates are always created from the three primary colors, necessary for transposition.

C1 is a non-primary color, so it is a shade of color in the RGBA color space. The red of the initial templates therefore becomes a shade of color when the calculated proportions are applied.

The step of calculating the proportion of red, blue and green in each channel to use in the FFMPEG filter. It is necessary to start by calculating the transposition of the first color on all the channels. The total value per column must be 1, so it is necessary to compensate the other channels so that the shades of gray are not affected. In the following case, the distribution is balanced between the two remaining channels (V & B), but it is possible to compensate on a single channel by declaring the last null.

Red Green Blue

Rci Vci B Ci ^RC 255 ^{= RC [O]} 255 = ^{Rc [11} 255 = ^RcP1 (1 - Rc [0]) (1 —Rc [l]) _UM1 (1 —Rc [2])

Vc --------- = Vc [0] --------- = Vc [l] --------- = Vc [2] (1 - Rc [0] ) (1 - Rc [l]) (1-Rc [2])

Bc --------- = Bc [0] --------- = Bc [l] --------- = Bc [2]

Ac 0 = Ac [0] 0 = Ac [l] 0 = Ac [2]

Alpha = Rc [3] = Vc [3] = Bc [3] = Ac [3] _

Application to the present case:

- R V B AT- 0314 0541 0.812 0 0343 0.2295 0094 0 0343 0.2295 0094 0 - 0 0 0 1

It remains to apply the calculated values in the FFMPEG filter which is composed as follows:

Rf = Rc [0]: Vc [0]: Bc [0]: Ac [0]

Vf = Rc [1]: Vc [1]: Bc [1]: Ac [1]

Bf = Rc [2]: Vc [2]: Bc [2]: Ac [2]

Af = Rc [3]: Vc [3]: Bc [3]: Ac [3]

Colorchannelmixer filter = Rc: VC: Bc: Ac

Colorchannelmixer filter = 0.314: 0.343: 0.343: 0: 0.541: 0.2295: 0.2295: 0: 0.812: 0.094: 0.094: 0

The colorization operation is complete, the next step being mounting over a data stream.

In a variant not shown, the method includes a step of storing the data stream on a remote server.

Example for two colors:

We will use in this new demonstration two colors, it is necessary to compensate only the Blue channel. Our user chose the colors RVBA (80, 136, 207, 1) and RVBA (207, 80, 172, 1), placed respectively in the variables Colorl and Color2.

R V B AT C1 80 138 207 1 C2 207 80 172 0

Let's do the same calculation as for a single color, with the blue channel in addition, the following matrix:

Rc, ^ = ^{B2 [0]}

Rc _{2 r Ί} ^1.2101 - 7? C [0] - Vc [0] = Bc [0] = 4c [0] ^Vc i ⁸²¹¹¹

Vc ₂ - 7? C [l] - Vc [l] = Bc [l] = 4c [l]

Bc, ⁸²¹²¹

Bc ₂ ^1.2121 - 7? C [2] - Vc [2] = Bc [2] 0 = Ac [2] = 7? C [3] = Vc [3] = Bc [3] = 4c [3] .

Application to the present case:

R V BA0.314 0.541 0.8120

0.812 0.314 0.6750

-0.126 0.145 -0.4870

0 01 Colorchannelmixer filter = 0.314: 0.812: -0.126: 0: 0.541: 0.314:

0.145: 0: 0.812: 0.675: -0.487: 0

In a variant not shown, the method includes a step of storing the data stream on a remote server.

As illustrated in FIG. 2, in a second particular embodiment, the audiovisual editing method which is the subject of the present invention comprises:

a) a step (100) of entering a plurality of data streams, at least one said data stream representing an image, a panorama, a video or produced from a 360 ° image,

a step (125) of handling a single workflow formed from at least one said data stream from step a), in an expression space targeted in three spatial dimensions over time, at during this stage a virtual camera is used which has either:

a first ability to perform rotations along the three axes, said three axes being horizontal, vertical and depth, and is fixed on said three axes, or a second ability to move on three axes, horizontal, vertical and depth and fixed on the rotations of said three axes.

a step (130) for managing a 360 ° environment using the first capacity of the virtual camera and a data flow in three spatial dimensions from a 360 ° panorama or video, during this step (130) of managing a 360 ° environment, a starting point P1 (x1, y1 and z1) and a finishing point P2 (x2, y2 and z2) are defined on said flow in three spatial dimensions, and a time,

- a step (135) of calculating intermediate positions, corresponding to different viewpoints of space in three spatial dimensions plus the time between the start point and the finish point on the flow in three spatial dimensions.

During step (135) of the calculation of the intermediate positions either an identical difference between the different positions of the flow in three spatial dimensions or an increase in the number of intermediate positions near the starting point and the ending point.

According to another example, the method comprises the following steps:

a) a step (100) of entering a plurality of data streams, at least one said data stream representing an image, a panorama, a video or produced from a 360 ° image,

a step (125) of handling a single workflow formed from at least one said data stream from step a), in an expression space targeted in three spatial dimensions over time, at during this stage a virtual camera is used which has either:

a first ability to perform rotations along the three axes, said three axes being horizontal, vertical and depth, and is fixed on said three axes, or a second ability to move on three axes, horizontal, vertical and depth and fixed on the rotations of said three axes.

a step (130) of managing a three-dimensional environment using at least one planar object using the second capacity of the virtual camera and of a two-dimensional spatial data flow from a video or a image, during this step (130) of managing a three-dimensional environment using at least one planar object, a starting point P1 (x1, y1 and z1) and a finishing point P2 (x2, y2 and z2) on said flow in two spatial dimensions, and a time,

- a step (135) of calculating intermediate positions, corresponding to different points of view of space in three spatial dimensions plus the time between the starting point and the point of arrival on the flow in two spatial dimensions.

During step (135) of the calculation of the intermediate positions either an identical difference between the different positions of the flow in two spatial dimensions or an increase in the number of intermediate positions near the starting point and the ending point.

In one example, the method comprises

- a step of storing the data flow on a remote server,

a rendering step from the data stored remotely and information from the starting point P1, the finishing point P2 and the time, said rendering step processes the calculation of each image on the server creating a graphical environment according to the OpenGL specification (registered trademark).

The remote server on which we create an OpenGL graphical environment allows the rendering in real time of our audiovisual rushes. The client realizes according to the steps of the process an animation in his browser using a webgl tool (OpenGL in the browser) and sends the different information so that they are consistent in the OpenGL environment of the remote server.

FIG. 3 shows a diagram with elements of a particular embodiment of the audiovisual editing method which is the subject of the present invention.

To manage 360 ° panoramas, we use a virtual camera (in 3D space) which can rotate, but it cannot move on three axes, horizontal, vertical and depth. The management of the approximation in the texture is carried out using the focal distance, this allowing a faithful photographic aspect.

In this figure, three axes (x, y and z) allow the camera to perform rotation. The camera is blocked on translational movements, so it cannot leave the sphere and always remains at the same distance from each point composing it. The user will then play on the tilt of the camera on these axes to define a starting point P1 (x1, y1 & z1) and an ending point P2 (x2, y2 & z2). Once these two keys have been defined, an animation will be performed. The system automatically calculates an optimal duration allowing an aesthetic animation of the video rush. For each image of the final rush, an intermediate position between P1 and P2 is calculated. The camera will then perform between these keys a so-called Linear animation (the difference between the different positions is always the same) or ease-in-out to facilitate entry into French for a slow animation at the start and end, making a damping effect. Once all the images have been made, the editing of all these images is done and therefore composes the final video rush.

A treatment with the addition of a black strip is carried out if the format is not 2: 1 and the texture of the sphere is applied towards the interior of the object.

FIG. 4 represents another diagram with elements of a particular embodiment of the audiovisual editing method which is the subject of the present invention.

To manage the video rushes from images, we use a virtual camera which does not have the capacity to carry out rotations, but it can move on three axes, horizontal, vertical and depth. In this case, the focal length of the camera is not used because it would distort the 2D object.

The axes represent the possibilities of movement of the camera. We will then perform an animation over the duration defined by the user between the two camera positions. The user can create animations only from translation. Whether vertical (Y), horizontal (X) or deep (Z), the user can combine the different translations. It will be limited in depth so that it cannot exceed the limits of the image it wishes to render.

FIG. 5 represents another diagram with elements of a particular embodiment of the audiovisual editing method which is the subject of the present invention.

In this figure, the first position called P1 composed of the positions

X1, Y1, Z1 and the so-called P2 position composed of X2, Y2 and Z2 allow you to define an animation. The camera will then carry out between P1 and P2 an animation called Linear or ease-in-out. As for the panoramas, the intermediate 5 steps are calculated allowing the composition of the final video rush.

NOMENCLATURE

100 Flow input

105 Manipulation of color (s)

110 Calculation of proportion

115 Applying calculated values to a data stream to be colored

120 Generation of a colorized flow over data flows

125 Handling a single workflow

130 Management of a 360 ° or three-dimensional environment using at least one planar object

135 Calculation of intermediate positions

Claims (8)

1. Audiovisual editing process characterized in that it includes: a) a step (100) of entering a plurality of data streams, at least one said data stream representing an image, a panorama, a video or produced from a 360 ° image, - b) a step (125) of manipulation of a single workflow formed from at least one said data flow from step a), in an expression space targeted in three spatial dimensions over time , during this step a virtual camera is used which has: either a first capacity to perform rotations along the three axes, said three axes being horizontal, vertical and depth, and is fixed on said three axes, • a step (130) of managing a 360 ° environment using the first capacity of the virtual camera and a three-dimensional spatial data flow from a 360 ° panorama or video, during this step (130) of managing a 360 ° environment, it is defined a depal point P1 (x1, y1 and z1), an arrival point P2 (x2, y2 and z2) on said flow in three spatial dimensions, and a time, • a step (135) for calculating intermediate positions, corresponding to different views of space in three spatial dimensions plus the time between the starting point and the point of arrival on the flow in three spatial dimensions. Or a second ability to move on the three axes, horizontal, vertical and depth and is fixed on the rotations of said three axes, • a step (130) of managing a three-dimensional environment using at least one planar object using the second capacity of the virtual camera and of a two-dimensional spatial data flow from a video or an image, during this step (130) of managing a three-dimensional environment using at least one planar object there is defined a starting point P1 (x1, y1 and z1), a point of arrival P2 (x2, y2 and z2) on said flow in two spatial dimensions, and a time, • a step (135) of calculation of intermediate positions, corresponding to different points of view of space in three spatial dimensions plus the time between the starting point and the point of arrival on the flow in two spatial dimensions. 2. Method according to claim 1, for the first capacity which comprises during step (135) of the calculation of the intermediate positions either an identical difference between the different positions of the flow in three spatial dimensions or an increase in the number of intermediate positions near the point of departure and the point of arrival. 3. Method according to claim 1, for the second capacity which comprises during the step (135) of the calculation of the intermediate positions either an identical difference between the different positions of the flow in three spatial dimensions or an increase in the number of intermediate positions near the point of departure and the point of arrival. 4. Method according to claim 1, which comprises - a step of storing the data flow on a remote server, a rendering step from the data stored remotely and information from the starting point P1, the finishing point P2 and the time, said rendering step processes the calculation of each image on the server creating a graphical environment according to the OpenGL specification (registered trademark). 5. Method according to claim 1, which comprises: a step (105) for handling a first color defined by three numerical values between 0 and 255 in the RGB color space corresponds respectively to the color Red, Green and Blue, the first value being defined as being Rci, the second value being defined as Vci, the third value being defined as Bci, - a step (110) of calculating the proportion of red, green and blue from the numerical values Rci, Vci, and Bci for a first color resulting from step b) according to the following matrix M1: rm 255 = ^{RC [O]} Vc. 255 = ^{RC [1]} Ben 255 = ^{flc [2]} (1 - fic [0]) (1 - MO]) = 7c [0] = Bc [0] (1 - fic [l]) (1 ~ /? C [l]) = 7c [l] = Bc [l] (1 - “C [2D (1 - 7? c [2]) = Vc [2] = Bc [2] a step (115) of applying the values calculated in step c) to a data stream to be colored, at least one said data stream to be colored representing an image or an expression, according to a colorimetric filter constituted in the manner next : color filter = Rf: Vf: Bf with Rf being defined by the values of the matrix Rc [0]: Vc [0]: Bc [0], Vf being defined by the values of the matrix Rc [1]: Vc [1 ]: Bc [1], and Bf being defined by the values of the matrix Rc [2]: Vc [2]: Bc [2], the data flow to be colored becoming the colorized data flow, a step (120) of generating the colorized data stream over at least one of the data streams of the step (105) of handling a first color. 6. The method of claim 1, wherein in step (105) of handling a first color, if a numerical value of a color is given in hexadecimal composed of six digits stored on three bytes, said value is transformed into three numeric values in the RGB color space. 7. Method according to claim 5, in which: the manipulation step (105) comprises a second color defined by three numerical values between 0 and 255 in the RGB color space corresponding respectively to the color Red, Green and Blue, the first value being defined as being Rû2, the second value being defined as Vû2, the third value being defined as Bc ₂ , the step (110) for calculating the proportion is replaced by a step for calculating the proportion of red, green and blue from the digital values King, Vci, Bci, Rû2, Vc2, and Bû2 for a first color and a second color from step b) according to the following matrix M2: FITT 255 = ^{RC [O]} Rc _{2 r Ί} 255 = ^ ¹ ° ¹ .1 - ffc [0] -Fc [0] = Bc [0] ^ = ^Rcl11 1 -ffc [t] -7C [t] = Bc [t] ^BC 1 ri 1 255 = ^RC Bc ₂ 255 = ^ ²¹ 1 -ffc [2] -Fc [2] = Bc [2]. 8. Method according to claim 7, in which: - the manipulation step (105) includes a fourth alpha numerical value between 0 and 1, the fourth value being defined as Ac, - the matrix M1 of step (110) of calculating the proportion is replaced by the following matrix: rc, 5S5 = ^{Rc [01} M2 _{= Kc [0]} (1 - AEC [0]) ---- 2 “^ = ^{Sc [0]} 0 = Ac [0] Fri 255 = ^{flc [1]} M » _{= M1]} (1 - 7? C [l]) --- γ ^ = Βο [1] Ben 2sh ^{Rc [21} Μ12 _{= νσ [2]} Μ ”= m2] 0 = Ac [2] 0 = Æc [3] 0 = 7c [3] 0 = Bc [3] 1 = Ac [3] _ and the matrix M2 of step (110) of calculating the proportion is replaced by the following matrix: Rc, ^ = ^{RC [01} Rc _{2 r Ί} ^ = ^ ⁰¹ l-Æc [0] -Lc [0] = Bc [0] 0 = 4c [0] ^Vc i ^ = ^{BC [1]} Vc ₂ 1 -7? C [l] - Lc [l] = Bc [l] 0 = 4c [l] Bc, ^ = ^{RC [21} Bc ₂ 1 -7? C [2] - Vc [2] = Bc [2] 0 = Ac [2] 0 = Æc [3] 0 = The [3] 0 = Bc [3] 1 = 4c [3]. - the colorimetric filter of step d) is completed with the values of Ac, ie: color filter = Rf: Vf: Bf: Af with Rf being defined by the values of the matrix Rc [0]: Vc [0]: Bc [0]: of the matrix Rc [1]: Vc [1]: Bc [ 1]: of the matrix Rc [2]: Vc [2]: Bc [2]: Ac [0], Vf being defined by the values Ac [1], Bf being defined by the values Ac [2], Af being defined by the values 5 of the matrix Rc [3]: Vc [3]: Bc [3]: Ac [3]. 1/3