Files
OpenSpace/modules/touch/src/directinputsolver.cpp
Mikael Pettersson 4e75b161db Feature/internal touchhandling (#1038)
* Removal of dead code and compiler warnings

* Added basic internal touch

This commit only adds the description-shell of the touch implementation

* Added callbacks and first WIP of internal touch

Makes use of the TouchInput/TouchInputs class in the TouchModule.
Internally we cache the TouchInputs as an input deque and utilizes it
for motion-vectors.
This commit has bugs and issues, which will be worked upon.

* Happy new year!

Bumped year on branch-local files

* Improvements to internal touch

Almost reached feature-parity with tuio-handled touch events

- Added most of the touch-logic to touchinteraction
- Added helper functions to new TouchInput/TouchInputs classes

* Naming changes to touch interface

* Translate TUIO to TouchInput

This commit translates TUIO messages to an internal TouchInput structure
while still trying to keep feature parity.
Removed TUIO-dependencies from many files.
Changed behavior on tuioear to lock-swap its content.

* Minor cleanup and fixes

- Should fix touch roll
- Simplified some functions

* Build fix

* Use internal touch in webgui

- Added consume-logic to touch callbacks
- Constrained touch-input to either webgui or 3D application as mouse is
  - This fixes some flaws with previous implementation,
    such as ghost inputs

- Initialize touchmodule through init-functions rather than constructor

* Cleanup of comments

* Simplified touch classes

Added timestamp through constructor meaning no more sprinkled timestamps
Renamed TouchInputs to TouchInputHolder for clarity
Added helper functions to the Holder to see if it holds an input
Remade addInput as tryAddInput which return true on successful insertion
+ other cleanup

* Code style cleanup and tweaks

Removed avoidable zero-comparison for code clarity
Cleanup of code style

* Added comments to DirectInputSolver

Clarifying the use of the DirectInputSolver.

* Changes for coding style
Change SGCT version to make it checkout-able

* Clarify magic bitmask

* const -> constexpr const for magic bitmasks

Co-authored-by: Alexander Bock <mail@alexanderbock.eu>
2020-01-13 08:27:13 +01:00

269 lines
10 KiB
C++

/*****************************************************************************************
* *
* OpenSpace *
* *
* Copyright (c) 2014-2019 *
* *
* Permission is hereby granted, free of charge, to any person obtaining a copy of this *
* software and associated documentation files (the "Software"), to deal in the Software *
* without restriction, including without limitation the rights to use, copy, modify, *
* merge, publish, distribute, sublicense, and/or sell copies of the Software, and to *
* permit persons to whom the Software is furnished to do so, subject to the following *
* conditions: *
* *
* The above copyright notice and this permission notice shall be included in all copies *
* or substantial portions of the Software. *
* *
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, *
* INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A *
* PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT *
* HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF *
* CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE *
* OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. *
****************************************************************************************/
#include <modules/touch/include/touchinteraction.h>
#include <openspace/scene/scenegraphnode.h>
#include <openspace/util/camera.h>
namespace {
// Used in the LM algorithm
struct FunctionData {
std::vector<glm::dvec3> selectedPoints;
std::vector<glm::dvec2> screenPoints;
int nDOF;
const openspace::Camera* camera;
openspace::SceneGraphNode* node;
LMstat stats;
};
} // namespace
namespace openspace {
DirectInputSolver::DirectInputSolver() {
levmarq_init(&_lmstat);
}
// project back a 3D point in model view to clip space [-1,1] coordinates on the view
// plane
glm::dvec2 castToNDC(const glm::dvec3& vec, Camera& camera, SceneGraphNode* node) {
glm::dvec3 posInCamSpace = glm::inverse(camera.rotationQuaternion()) *
(node->worldRotationMatrix() * vec +
(node->worldPosition() - camera.positionVec3()));
glm::dvec4 clipspace = camera.projectionMatrix() * glm::dvec4(posInCamSpace, 1.0);
return (glm::dvec2(clipspace) / clipspace.w);
}
// Returns the screen point s(xi,par) dependent the transform M(par) and object point xi
double distToMinimize(double* par, int x, void* fdata, LMstat* lmstat) {
FunctionData* ptr = reinterpret_cast<FunctionData*>(fdata);
// Apply transform to camera and find the new screen point of the updated camera state
// { vec2 globalRot, zoom, roll, vec2 localRot }
double q[6] = { 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 };
for (int i = 0; i < ptr->nDOF; ++i) {
q[i] = par[i];
}
using namespace glm;
// Create variables from current state
dvec3 camPos = ptr->camera->positionVec3();
dvec3 centerPos = ptr->node->worldPosition();
dvec3 directionToCenter = normalize(centerPos - camPos);
dvec3 lookUp = ptr->camera->lookUpVectorWorldSpace();
dvec3 camDirection = ptr->camera->viewDirectionWorldSpace();
// Make a representation of the rotation quaternion with local and global
// rotations
dmat4 lookAtMat = lookAt(
dvec3(0, 0, 0),
directionToCenter,
// To avoid problem with lookup in up direction
normalize(camDirection + lookUp)
);
dquat globalCamRot = normalize(quat_cast(inverse(lookAtMat)));
dquat localCamRot = inverse(globalCamRot) * ptr->camera->rotationQuaternion();
{
// Roll
dquat rollRot = angleAxis(q[3], dvec3(0.0, 0.0, 1.0));
localCamRot = localCamRot * rollRot;
}
{
// Panning (local rotation)
dvec3 eulerAngles(q[5], q[4], 0);
dquat panRot = dquat(eulerAngles);
localCamRot = localCamRot * panRot;
}
{
// Orbit (global rotation)
dvec3 eulerAngles(q[1], q[0], 0);
dquat rotationDiffCamSpace = dquat(eulerAngles);
dvec3 centerToCamera = camPos - centerPos;
dquat rotationDiffWorldSpace =
globalCamRot * rotationDiffCamSpace * inverse(globalCamRot);
dvec3 rotationDiffVec3 = centerToCamera * rotationDiffWorldSpace - centerToCamera;
camPos += rotationDiffVec3;
centerToCamera = camPos - centerPos;
directionToCenter = normalize(-centerToCamera);
dvec3 lookUpWhenFacingCenter =
globalCamRot * dvec3(ptr->camera->lookUpVectorCameraSpace());
lookAtMat = lookAt(
dvec3(0, 0, 0),
directionToCenter,
lookUpWhenFacingCenter
);
globalCamRot = normalize(quat_cast(inverse(lookAtMat)));
}
{ // Zooming
camPos += directionToCenter * q[2];
}
// Update the camera state
Camera cam = *(ptr->camera);
cam.setPositionVec3(camPos);
cam.setRotation(globalCamRot * localCamRot);
// we now have a new position and orientation of camera, project surfacePoint to
// the new screen to get distance to minimize
glm::dvec2 newScreenPoint = castToNDC(
ptr->selectedPoints.at(x),
cam,
ptr->node
);
lmstat->pos.push_back(newScreenPoint);
return glm::length(ptr->screenPoints.at(x) - newScreenPoint);
}
// Gradient of distToMinimize w.r.t par (using forward difference)
void gradient(double* g, double* par, int x, void* fdata, LMstat* lmstat) {
FunctionData* ptr = reinterpret_cast<FunctionData*>(fdata);
double f0 = distToMinimize(par, x, fdata, lmstat);
// scale value to find minimum step size h, dependant on planet size
double scale = log10(ptr->node->boundingSphere());
std::vector<double> dPar(ptr->nDOF, 0.0);
dPar.assign(par, par + ptr->nDOF);
for (int i = 0; i < ptr->nDOF; ++i) {
// Initial values
double h = 1e-8;
double lastG = 1;
dPar.at(i) += h;
double f1 = distToMinimize(dPar.data(), x, fdata, lmstat);
dPar.at(i) = par[i];
// Iterative process to find the minimum step h that gives a good gradient
for (int j = 0; j < 100; ++j) {
if ((f1 - f0) != 0 && lastG == 0) { // found minimum step size h
// scale up to get a good initial guess value
h *= scale * scale * scale;
// clamp min step size to a fraction of the incoming parameter
if (i == 2) {
double epsilon = 1e-3;
// make sure incoming parameter is larger than 0
h = std::max(std::max(std::abs(dPar.at(i)), epsilon) * 0.001, h);
}
else if (ptr->nDOF == 2) {
h = std::max(std::abs(dPar.at(i)) * 0.001, h);
}
// calculate f1 with good h for finite difference
dPar[i] += h;
f1 = distToMinimize(dPar.data(), x, fdata, lmstat);
dPar[i] = par[i];
break;
}
else if ((f1 - f0) != 0 && lastG != 0) {
// h too big
h /= scale;
}
else if ((f1 - f0) == 0) {
// h too small
h *= scale;
}
lastG = f1 - f0;
dPar.at(i) += h;
f1 = distToMinimize(dPar.data(), x, fdata, lmstat);
dPar.at(i) = par[i];
}
g[i] = (f1 - f0) / h;
}
if (ptr->nDOF == 2) {
// normalize on 1 finger case to allow for horizontal/vertical movement
for (int i = 0; i < 2; ++i) {
g[i] = g[i] / std::abs(g[i]);
}
}
else if (ptr->nDOF == 6) {
for (int i = 0; i < ptr->nDOF; ++i) {
// lock to only pan and zoom on 3 finger case, no roll/orbit
g[i] = (i == 2) ? g[i] : g[i] / std::abs(g[i]);
}
}
}
bool DirectInputSolver::solve(const std::vector<TouchInputHolder>& list,
const std::vector<SelectedBody>& selectedBodies,
std::vector<double>* parameters, const Camera& camera)
{
int nFingers = std::min(static_cast<int>(list.size()), 3);
_nDof = std::min(nFingers * 2, 6);
// Parse input data to be used in the LM algorithm
std::vector<glm::dvec3> selectedPoints;
std::vector<glm::dvec2> screenPoints;
for (int i = 0; i < nFingers; ++i) {
const SelectedBody& sb = selectedBodies.at(i);
selectedPoints.push_back(sb.coordinates);
screenPoints.emplace_back(
2.0 * (list[i].latestInput().x - 0.5),
-2.0 * (list[i].latestInput().y - 0.5)
);
}
FunctionData fData = {
selectedPoints,
screenPoints,
_nDof,
&camera,
selectedBodies.at(0).node,
_lmstat
};
void* dataPtr = reinterpret_cast<void*>(&fData);
bool result = levmarq(
_nDof,
parameters->data(),
static_cast<int>(screenPoints.size()),
nullptr,
distToMinimize,
gradient,
dataPtr,
&_lmstat
);
return result;
}
int DirectInputSolver::nDof() const {
return _nDof;
}
const LMstat& DirectInputSolver::levMarqStat() {
return _lmstat;
}
void DirectInputSolver::setLevMarqVerbosity(bool verbose) {
_lmstat.verbose = verbose;
}
} // openspace namespace