CN109816763A - A kind of method for rendering graph - Google Patents
A kind of method for rendering graph Download PDFInfo
- Publication number
- CN109816763A CN109816763A CN201811579824.3A CN201811579824A CN109816763A CN 109816763 A CN109816763 A CN 109816763A CN 201811579824 A CN201811579824 A CN 201811579824A CN 109816763 A CN109816763 A CN 109816763A
- Authority
- CN
- China
- Prior art keywords
- model
- rendering
- group
- frame
- scene
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Abstract
A kind of method for rendering graph, includes the following steps, by scene partitioning at multiple regions landform group, and is screened according to threshold value to all models in scene；The model filtered out is arranged, is merged, the model with same alike result is assigned in a model group, generates multiple model groups with same alike result；It calculates and generates data corresponding with the model group.Method for rendering graph of the invention significantly reduces rendering batch size, saves the CPU operation time, effectively improves rendering performance, program draws effect and program operation frame number.
Description
Technical field
The present invention relates to graph processing technique fields, more particularly to a kind of method for rendering graph.
Background technique
In graph image interactive program, one of the effective ways of abundant user experience be exactly present one it is genuine and believable
The world, be filled with a large amount of various interesting objects and the object rich in various characteristics.From unnumbered thick grass, trees, river
Stream, sundries to various buildings, landscape etc..All these objects can increase the visual experience of user and raising allows use
Immerse substitution sense therein in family.But from the perspective of rendering, so various, a large amount of object is drawn similarly to program
Bring huge pressure.
During figure rendering, it is a relatively slow operation that data, which are submitted to GPU rendering,.An object
A Rendering operations process must at least be executed by being rendered on screen, this operating process is exactly a Draw Call, i.e., and one
A rendering batch.According to the description in " GPU GEM2 " book: " Wloka2003 shows that the CPU of 1GHz is per second can only to be rendered greatly
About 10000-40000 batch (batch).On more modern CPU, estimate this number probably in 30000-120000 per second
Batch." in summary, that is to say, that if we want to draw a scene, each rendering batch is an object in scene,
So, no matter how to operate, and the maximum rendering quantity of object can only achieve 4000 or so and (be to FPS in this scene
General every 1000 to 4000 batch of frame for the system of 30frame/sec).And this operated in saturation will lead to program because
CPU does not have time to handle other tasks and generate exception.Such case is clearly very undesirable, so needing one kind has
The reduction of effect renders batch, promotes the solution of rendering performance to handle this problem.
Summary of the invention
In order to solve the shortcomings of the prior art, the purpose of the present invention is to provide a kind of method for rendering graph, effectively
Saving frequently submit CPU calculation amount caused by rendering batch operation excessive, and simultaneously the GPU free time without operation or frequent switching
The loss of rendering state bring performance, while reducing rendering batch size and effectively improving rendering performance.
To achieve the above object, a kind of method for rendering graph provided by the invention, comprising the following steps:
By scene partitioning at multiple regions landform group, and all models in scene are screened according to threshold value；
The model filtered out is arranged, is merged, the model with same alike result is assigned in a model group, generates tool
There are multiple model groups of same alike result；
It calculates and generates data corresponding with the model group.
Further, the step of all models in scene screen further comprises:
Obtain present frame in rendering in need all objects model.
Further, the step of described pair of model filtered out is arranged, merged further comprises:
Obtain in each region landform point rendering in need valid model list；
The model in list is screened one by one；
Model with same alike result is assigned in a new model group, to generate multiple model groups with same alike result.
Further, described the step of calculating and generating data corresponding with above-mentioned model group, further comprise:
The duplicate attribute instance of model is turned into a attribute.
Further, described the step of calculating and generating data corresponding with above-mentioned model group, further comprise:
The model rendered in the render list of each frame to needs does replacement detection.
Further, the model rendered in the render list of each frame to needs does the step of replacement detection,
Further comprise:
In current render list delete former frame or before it is any formulate frame do not need rendering model.
The implementation of method for rendering graph of the invention significantly reduces rendering batch size, when saving CPU operation
Between, effectively improve rendering performance, program draws effect and program operation frame number.
Other features and advantages of the present invention will be illustrated in the following description, also, partly becomes from specification
It obtains it is clear that understand through the implementation of the invention.
Detailed description of the invention
Attached drawing is used to provide further understanding of the present invention, and constitutes part of specification, and with it is of the invention
Embodiment together, is used to explain the present invention, and is not construed as limiting the invention.In the accompanying drawings:
Fig. 1 is static batch merging method flow chart according to the present invention；
Fig. 2 is the relational graph of the static grouping of batch merging method mesorelief and model grouping according to the present invention；
Fig. 3 is dynamic batch merging method flow chart according to the present invention.
Specific embodiment
Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings, it should be understood that preferred reality described herein
Apply example only for the purpose of illustrating and explaining the present invention and is not intended to limit the present invention.
In rendering program, the cycle time and rendering batch size that CPU expends show linear relationship, rendering one
10,000 triangles of ten thousand identical triangles and primary rendering, the performance gap of the two be can not ignore.So having
Attribute (such as material, texture, vertex or only spatial position, size, the phase of rotation difference according to certain rending models of effect
With model etc.) reasonably merge some models into a rendering batch, it, will in the case where having no need to change rendering state
Originally the object for needing multiple rendering batches to draw, merging are uniformly processed (such as: 100 of same material in a batch
Tree, only spatial position, scaling, direction are different, we can be merged into a model, original by doing to the vertex after merging
The data-bias of affiliated model calculates, and 100 trees need 100 rendering batches originally in this way, and 1 rendering batch is reduced to after merging
It is secondary).In this way, not only can solve the submission bring CPU operation of a large amount of rendering batch originally, but also GPU can be reduced because every
State switching bring redundancy between a rendering batch can be lost.
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, below by specific embodiment
The present invention is described in further detail.
Embodiment one
Static batch merges
Fig. 1 is the flow chart that the static batch that the present embodiment is related to merges, below with reference to Fig. 1, to static batch of the invention
Merging method is described in detail.
Firstly, in step 101, by a scene partitioning at multiple regions landform group.In reality, usual human eye is can not
See behind and more than the object outside certain field range, so equally in the scene, we can be according to program or user
Set a reasonable controllable field range critical value.The field range critical value represents maximum visual in current scene
A threshold values i.e. field range critical value is arranged in a step 101 in distance or consolidation scope distance；According to the threshold values by one
Scene partitioning is at multiple regions landform group.Thereby, it is possible to acquire 1-N region landform grouping (region landform group 1, region
Landform group 2, region landform group 3, region landform group N ...).As shown in Fig. 2, in the present embodiment, a scene is divided into 4 areas
Domain landform group, but the number of region landform group is without being limited thereto.The size of each region landform grouping can be identical, can also mutually not
It is identical.
In step 102, screening is executed to each region landform group and is calculated.In a step 102, it obtains in each region landform grouping
Rendering in need valid model list；The model in list is screened one by one, and will have same alike result (such as material,
Texture, vertex etc.) model be assigned in a new model group, to generate multiple model group (models with same alike result
Group 1, model group 2, model group 3, model group N ...).If not the model of same alike result can ignore without generate model
Group.As a result, in the present embodiment, multiple models with same alike result feature that entire landform group 1 screens can be obtained
Group.
In step 103, the multiple model groups filtered out are ranked up and it is merged.In step 103, according to
It needs, such as according to the distance between model perhaps vertex quantity multiple models or list close to the distance of "current" model group
A model is ranked up；To after sequence multiple models or single model merge.In the present embodiment, it is contemplated that rendering
The problem of cutting, together by the close model sequence of distance in single grouping.Since in rendering program, video camera and human eye
Equally there is certain visual range and hiding relation, will not be drawn out beyond the visual field or shielded model.If will
The model nearest apart from video camera is together with model combination relatively far away from, then certain cutting out property of video camera will be slatterned
Can, for another example fruit sets certain merging rules, such as vertex quantity, if then the current model needs for having 100 same alike results
Merge, merging in regular critical value, merges the model that a part of maximum limitation is closed on, remerge remaining model, equally
Video camera can also effectively be improved and cut performance.
In step 104, calculates and generate data corresponding with above-mentioned model group.Specifically, by calculating these models
Duplicate attribute instance turns on a attribute (such as material, texture, vertex), and slightly discrepant data pass through corresponding offset
A data (vertex position, scaling in such as space, rotation) is merged into after calculating.Pass through optimal feelings after calculating in this way
Condition is original calculation amount for needing N number of rendering batch can be reduced to 1 rendering batch.
In step 105, repeat step 103 and step 104 until until entire scene joint account terminates.When
After the union operation of above-mentioned model group 1, can be operated with circulating repetition remaining model group 2, model group N ... and it is remaining its
He does same union operation at region landform group, until entire scene joint account terminates.
Embodiment two
Dynamic batch merges
Embodiment one provides a kind of method that preparatory static rendering batch merges, and effectively reduces the rendering batch in scene
Quantity, still, in some cases, we be possible to also to need it is agile, dynamically increase in real time, decompose, deleting some scenes
In object, so, the present invention also provides second for dynamic batch merge technical method.
Embodiment two is different from embodiment one according to presetting field range by user or consolidation scope numerical value is pre-
The mode of shaped area is first divided, but only that obtaining the wash with watercolours in the video camera space that present procedure is presented on the display device
Range is contaminated as combined region computer capacity.
Firstly, obtaining model list in present frame in step 301.In step 301, present procedure render list is obtained
In it is in need present frame draw object.
In step 302, model list is carried out to execute screening calculating.It can use and matching mould identical in embodiment one
The mode of type screens the model in current rendering frame, is grouped, merges to qualified model, ineligible
Model do not operate, keep Rendering operations do not change.New screening conditions can also be used, such as: the material of specified limitation, line
Reason etc..
In step 303, the model group filtered out is ranked up and it is merged.Can using in embodiment one
Identical processing.Simultaneously, it is contemplated that be real-time operation, save render performance under the premise of, can also increase it is some relative to
Merge merging rule more direct, stringent but more efficient for rule in embodiment one, such as: the limitation on vertex, face number
Or whether it is necessary for same model etc. requirement.
In step 304, the model rendered in the render list of each frame to needs does replacement detection.Specifically, such as
They are done union operation, then delete this by the model for not needing rendering in previous frame but needing to render in present frame
It is not included in current render list in frame or in specified frame number (the 1-N frame as before) but in the mould being merged before
Type.Circulate operation repeatedly just meets the demand scheme of model dynamic batch merging.
Those of ordinary skill in the art will appreciate that: the foregoing is only a preferred embodiment of the present invention, and does not have to
In the limitation present invention, although the present invention is described in detail referring to the foregoing embodiments, for those skilled in the art
For, still can to foregoing embodiments record technical solution modify, or to part of technical characteristic into
Row equivalent replacement.All within the spirits and principles of the present invention, any modification, equivalent replacement, improvement and so on should all include
Within protection scope of the present invention.
Claims (6)
1. a kind of method for rendering graph, comprising the following steps:
By scene partitioning at multiple regions landform group, and all models in scene are screened according to threshold value；
The model filtered out is arranged, is merged, the model with same alike result is assigned in a model group, generates tool
There are multiple model groups of same alike result；
It calculates and generates data corresponding with the model group.
2. image rendering method according to claim 1, which is characterized in that it is described according to threshold value to all moulds in scene
The step of type is screened further comprises:
Obtain present frame in rendering in need all objects model.
3. rendering image rendering method according to claim 1, which is characterized in that the described pair of model filtered out is arranged
The step of column, merging, further comprise:
Obtain in each region landform point rendering in need valid model list；
The model in list is screened one by one；
Model with same alike result is assigned in a new model group, to generate multiple model groups with same alike result.
4. image rendering method according to claim 1, which is characterized in that the calculating simultaneously generates and above-mentioned model group pair
The step of data answered, further comprise:
The duplicate attribute instance of model is turned into a attribute.
5. image rendering method according to claim 1, which is characterized in that the calculating simultaneously generates and above-mentioned model group pair
The step of data answered, further comprise:
The model rendered in the render list of each frame to needs does replacement detection.
6. rendering image rendering method according to claim 5, which is characterized in that described in the render list of each frame
The step of doing replacement detection to the model that needs render further comprises:
In current render list delete former frame or before it is any formulate frame do not need rendering model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811579824.3A CN109816763A (en) | 2018-12-24 | 2018-12-24 | A kind of method for rendering graph |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811579824.3A CN109816763A (en) | 2018-12-24 | 2018-12-24 | A kind of method for rendering graph |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109816763A true CN109816763A (en) | 2019-05-28 |
Family
ID=66602323
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811579824.3A Withdrawn CN109816763A (en) | 2018-12-24 | 2018-12-24 | A kind of method for rendering graph |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109816763A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111063032A (en) * | 2019-12-26 | 2020-04-24 | 北京像素软件科技股份有限公司 | Model rendering method and system and electronic device |
| CN111744182A (en) * | 2020-07-03 | 2020-10-09 | 广州要玩娱乐网络技术股份有限公司 | Batch rendering method and device of game scenes and server |
| CN111798361A (en) * | 2019-09-20 | 2020-10-20 | 厦门雅基软件有限公司 | Rendering method, rendering device, electronic equipment and computer-readable storage medium |
| CN111882636A (en) * | 2020-07-24 | 2020-11-03 | 上海米哈游天命科技有限公司 | Picture rendering method, device, equipment and medium |
| CN112423111A (en) * | 2020-11-05 | 2021-02-26 | 上海哔哩哔哩科技有限公司 | Graphic engine and graphic processing method suitable for player |
| CN113426130A (en) * | 2021-07-01 | 2021-09-24 | 网易（杭州）网络有限公司 | Batch processing method and device for models |
| CN115272549A (en) * | 2022-08-02 | 2022-11-01 | 动森数字科技(重庆)有限公司 | Method and device for storing, rendering and scheduling super-large digital scene |
-
2018
- 2018-12-24 CN CN201811579824.3A patent/CN109816763A/en not_active Withdrawn
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111798361A (en) * | 2019-09-20 | 2020-10-20 | 厦门雅基软件有限公司 | Rendering method, rendering device, electronic equipment and computer-readable storage medium |
| CN111063032A (en) * | 2019-12-26 | 2020-04-24 | 北京像素软件科技股份有限公司 | Model rendering method and system and electronic device |
| CN111063032B (en) * | 2019-12-26 | 2024-02-23 | 北京像素软件科技股份有限公司 | Model rendering method, system and electronic device |
| CN111744182A (en) * | 2020-07-03 | 2020-10-09 | 广州要玩娱乐网络技术股份有限公司 | Batch rendering method and device of game scenes and server |
| CN111882636A (en) * | 2020-07-24 | 2020-11-03 | 上海米哈游天命科技有限公司 | Picture rendering method, device, equipment and medium |
| CN112423111A (en) * | 2020-11-05 | 2021-02-26 | 上海哔哩哔哩科技有限公司 | Graphic engine and graphic processing method suitable for player |
| CN113426130A (en) * | 2021-07-01 | 2021-09-24 | 网易（杭州）网络有限公司 | Batch processing method and device for models |
| CN115272549A (en) * | 2022-08-02 | 2022-11-01 | 动森数字科技(重庆)有限公司 | Method and device for storing, rendering and scheduling super-large digital scene |
| CN115272549B (en) * | 2022-08-02 | 2023-12-15 | 动森数字科技(重庆)有限公司 | Storage and rendering scheduling method and device for oversized digital scene |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109816763A (en) | A kind of method for rendering graph | |
| US10964105B2 (en) | Rendering views of a scene in a graphics processing unit | |
| US7439982B2 (en) | Optimized scene graph change-based mixed media rendering | |
| JP7146091B2 (en) | Information embedding method in video, computer equipment and computer program | |
| US20130300741A1 (en) | Adaptive mesh refinement | |
| US9041773B2 (en) | Conversion of 2-dimensional image data into 3-dimensional image data | |
| CN106658139B (en) | Focus control method and device | |
| US9734551B1 (en) | Providing depth-of-field renderings | |
| DE112009002383T5 (en) | Graphics processing using culling on groups of vertices | |
| CN106339983A (en) | Blurring animation realization method through Gaussian blurring and blurring animation realization device thereof | |
| US20160111129A1 (en) | Image edits propagation to underlying video sequence via dense motion fields | |
| Cabral et al. | Structure‐Preserving Reshape for Textured Architectural Scenes | |
| DE112013003714T5 (en) | Stochastic depth buffer compression by means of generalized plane coding | |
| CN107679150A (en) | Magnanimity three-dimensional data fast dispatch method | |
| CN104506867B (en) | Sample point self-adapted offset parameter method of estimation and device | |
| CN115861510A (en) | Object rendering method, device, electronic equipment, storage medium and program product | |
| KR101667306B1 (en) | Low-Complexity Topological Derivative-Based Image Segmentation Method and System | |
| CN104517273A (en) | Image super-resolution processing method and apparatus | |
| CN105243652B (en) | The method and device of image noise reduction | |
| EP2821997A1 (en) | Method and device for editing a video sequence | |
| US9947114B2 (en) | Modifying gradation in an image frame including applying a weighting to a previously processed portion of the image frame | |
| CN116711301A (en) | Virtual viewpoint generating, rendering and decoding methods and devices, equipment and storage medium | |
| Buades et al. | Separable soft shadow mapping | |
| CN108876912A (en) | Three-dimensional scenic physics renders method and its system | |
| CN115686727B (en) | Method for realizing synthesis rendering based on wlroots |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190528 |
|
| WW01 | Invention patent application withdrawn after publication |