kraken_engine/kraken
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fixed access violations

Browse Source 
Implemented a divide-and-conquer approach to speeding up reverb engine, without increasing latency using multiple, different sized FFT's.  Algorithm optimized to ensure that the same amount of processing is done each frame.  Reverb now working in real-time.
This commit is contained in:
Kearwood Gilbert
2013-02-16 01:35:05 -08:00
parent eee1dc28e0
commit 7e0459653d
6 changed files with 42 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
KREngine/kraken/KRScene.h
View File

@@ -67,7 +67,7 @@ public:
void renderFrame(float deltaTime, int width, int height);
void render(KRCamera *pCamera, std::map<KRAABB, int> &visibleBounds, const KRViewport &viewport, KRNode::RenderPass renderPass, bool new_frame);
void render(KROctreeNode *pOctreeNode, std::map<KRAABB, int> &visibleBounds, KRCamera *pCamera, std::vector<KRLight *> lights, const KRViewport &viewport, KRNode::RenderPass renderPass, std::vector<KROctreeNode *> &remainingOctrees, std::vector<KROctreeNode *> &remainingOctreesTestResults, std::vector<KROctreeNode *> &remainingOctreesTestResultsOnly, bool bOcclusionResultsPass, bool bOcclusionTestResultsOnly);
void render(KROctreeNode *pOctreeNode, std::map<KRAABB, int> &visibleBounds, KRCamera *pCamera, std::vector<KRLight *> &lights, const KRViewport &viewport, KRNode::RenderPass renderPass, std::vector<KROctreeNode *> &remainingOctrees, std::vector<KROctreeNode *> &remainingOctreesTestResults, std::vector<KROctreeNode *> &remainingOctreesTestResultsOnly, bool bOcclusionResultsPass, bool bOcclusionTestResultsOnly);
void notify_sceneGraphCreate(KRNode *pNode);