UpsilonNumworks/Upsilon
1
0
Fork 0
mirror of https://github.com/UpsilonNumworks/Upsilon.git synced 2026-01-19 00:37:25 +01:00
Code Issues Packages Projects Releases Wiki Activity

[shared] InteractiveCurveViewRange::isOrthonormal

Browse Source 
The old method isOrthonormal has been split into two :
  - shouldBeNormalized tests whether the range is close enough to a
    normal range
  - isOrthonormal tests whether the range is strictly orthonormal, but
    takes into account imprecisions of floating-point arithmetic
This commit is contained in:
Gabriel Ozouf
2021-01-04 14:41:32 +01:00
committed by LeaNumworks
parent 89107da1ae
commit 89c50509d8
3 changed files with 42 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

44
apps/shared/interactive_curve_view_range.cpp
View File

@@ -157,8 +157,13 @@ void InteractiveCurveViewRange::normalize(bool forceChangeY) {
m_yRange.setMin(newYMin, k_lowerMaxFloat, k_upperMaxFloat);
MemoizedCurveViewRange::protectedSetYMax(newYMax, k_lowerMaxFloat, k_upperMaxFloat);
/* The range should be close to orthonormal, unless it has been clipped because the maximum bounds have been reached. */
assert(isOrthonormal() || xMin() <= - k_lowerMaxFloat || xMax() >= k_lowerMaxFloat || yMin() <= - k_lowerMaxFloat || yMax() >= k_lowerMaxFloat);
/* The range should be close to orthonormal, unless :
* - it has been clipped because the maximum bounds have been reached.
* - the the bounds are too close and of too large a magnitude, leading to
* a drastic loss of significance. */
assert(isOrthonormal()
|| xMin() <= - k_lowerMaxFloat || xMax() >= k_lowerMaxFloat || yMin() <= - k_lowerMaxFloat || yMax() >= k_lowerMaxFloat
|| normalizationSignificantBits() <= 0);
setZoomNormalize(isOrthonormal());
}
@@ -193,7 +198,7 @@ void InteractiveCurveViewRange::setDefault() {
m_yRange.setMin(roundLimit(m_delegate->addMargin(yMin(), yRange, true , true), yRange, true), k_lowerMaxFloat, k_upperMaxFloat);
MemoizedCurveViewRange::protectedSetYMax(roundLimit(m_delegate->addMargin(yMax(), yRange, true , false), yRange, false), k_lowerMaxFloat, k_upperMaxFloat);
if (m_delegate->defaultRangeIsNormalized() || isOrthonormal(k_orthonormalTolerance)) {
if (m_delegate->defaultRangeIsNormalized() || shouldBeNormalized()) {
/* Normalize the axes, so that a polar circle is displayed as a circle.
* If we are displaying cartesian functions with a default range, we want
* the X bounds untouched. */
@@ -276,8 +281,27 @@ void InteractiveCurveViewRange::panToMakePointVisible(float x, float y, float to
setZoomNormalize(isOrthonormal());
}
bool InteractiveCurveViewRange::isOrthonormal(float tolerance) const {
if (tolerance == 0.f) {
bool InteractiveCurveViewRange::shouldBeNormalized() const {
float ratio = (yMax() - yMin()) / (xMax() - xMin());
return ratio >= NormalYXRatio() / k_orthonormalTolerance && ratio <= NormalYXRatio() * k_orthonormalTolerance;
}
bool InteractiveCurveViewRange::isOrthonormal() const {
float significantBits = normalizationSignificantBits();
if (significantBits <= 0) {
return false;
}
float ratio = (yMax() - yMin()) / (xMax() - xMin());
/* The last N (= 23 - significantBits) bits of "ratio" mantissa have become
* insignificant. "tolerance" is the difference between ratio with those N
* bits set to 1, and ratio with those N bits set to 0 ; i.e. a measure of
* the interval in which numbers are indistinguishable from ratio with this
* level of precision. */
float tolerance = std::pow(2.f, IEEE754<float>::exponent(ratio) - significantBits);
return ratio - tolerance <= NormalYXRatio() && ratio + tolerance >= NormalYXRatio();
}
int InteractiveCurveViewRange::normalizationSignificantBits() const {
float xr = std::fabs(xMin()) > std::fabs(xMax()) ? xMax() / xMin() : xMin() / xMax();
float yr = std::fabs(yMin()) > std::fabs(yMax()) ? yMax() / yMin() : yMin() / yMax();
/* The subtraction x - y induces a loss of significance of -log2(1-x/y)
@@ -285,9 +309,11 @@ bool InteractiveCurveViewRange::isOrthonormal(float tolerance) const {
* the ratio of the normalized range will deviate from the Normal ratio. We
* add an extra two lost bits to account for loss of precision from other
* sources. */
tolerance = std::pow(2.f, - std::log2(std::min(1.f - xr, 1.f - yr)) - 23.f + 2.f);
}
float ratio = (yMax() - yMin()) / (xMax() - xMin());
return ratio <= NormalYXRatio() + tolerance && ratio >= NormalYXRatio() - tolerance;
float loss = std::log2(std::min(1.f - xr, 1.f - yr));
if (loss > 0.f) {
loss = 0.f;
}
return std::floor(loss + 23.f - 2.f);
}
}