F-Curve modifier - Builtin-Function - Wrong Sine Amplitude for Angle transformation #42705

New Issue

Dolphin Dream · 2014-11-24T02:51:17+01:00

Dolphin Dream commented

2014-11-24 02:51:17 +01:00

System Information
MacBook Pro
Processor 2.2 GHz Intel Core i7
Memory 8 GB 1333 MHz DDR3
Graphics AMD Radeon HD 6750M 1024 MB
Software OS X 10.8.5 - Mountain Lion

Blender Version
2.72b (9e963ae)

Short description of error
While keyframing a rotation transformation (e.g. rotation about Z axis) using the f-curve modifier (builtin-function) to generate a sine wave, for the default amplitude of 1 the actual amplitude of the generated plot is actually about 57.3 (instead of 1). See the attached screenshot & blend file.

Exact steps for others to reproduce the error
See attached blend file. Add keyframe at frame 0 for the Z rotation for the cube object. In the f-curve editor add a bultin-function modifier (sine) to the z Euler rotation. The generated plot has an amplitude of 57.3ish.

Note:
Btw, I wish this builtin function(s) would be enhanced to allow to define a precise period/frequency, and definitely better documented. Sure, the frequency may be depended on the actual frame rate, but the current "phase multiplier" that alters the frequency is a rather obscure parameter. IMO a simpler parameter set that would allow a sine function of given amplitude and frequency is desirable. Also, it would be helpful if the phase multiplier change could be performed around the mouse location instead of (i presume) zero? I'm having a hard time matching a manually generated sine curve with a generated one by adjusting the phase multiplier and offset independently.

fCurveModifier-WrongSineAmplitude.blend

System Information MacBook Pro Processor 2.2 GHz Intel Core i7 Memory 8 GB 1333 MHz DDR3 Graphics AMD Radeon HD 6750M 1024 MB Software OS X 10.8.5 - Mountain Lion Blender Version 2.72b (9e963ae) **Short description of error** While keyframing a rotation transformation (e.g. rotation about Z axis) using the f-curve modifier (builtin-function) to generate a sine wave, for the default amplitude of 1 the actual amplitude of the generated plot is actually about 57.3 (instead of 1). See the attached screenshot & blend file. **Exact steps for others to reproduce the error** See attached blend file. Add keyframe at frame 0 for the Z rotation for the cube object. In the f-curve editor add a bultin-function modifier (sine) to the z Euler rotation. The generated plot has an amplitude of 57.3ish. Note: Btw, I wish this builtin function(s) would be enhanced to allow to define a precise period/frequency, and definitely better documented. Sure, the frequency may be depended on the actual frame rate, but the current "phase multiplier" that alters the frequency is a rather obscure parameter. IMO a simpler parameter set that would allow a sine function of given amplitude and frequency is desirable. Also, it would be helpful if the phase multiplier change could be performed around the mouse location instead of (i presume) zero? I'm having a hard time matching a manually generated sine curve with a generated one by adjusting the phase multiplier and offset independently. ![Screen_Shot_2014-11-23_at_8.29.53_PM.png](https://archive.blender.org/developer/F125788/Screen_Shot_2014-11-23_at_8.29.53_PM.png) [fCurveModifier-WrongSineAmplitude.blend](https://archive.blender.org/developer/F125789/fCurveModifier-WrongSineAmplitude.blend)

Dolphin Dream commented

2014-11-24 02:51:17 +01:00

Changed status to: 'Open'

Dolphin Dream commented

2014-11-24 02:51:17 +01:00

Added subscriber: @MariusGiurgi

Joshua Leung was assigned by Sergey Sharybin

2014-11-25 15:19:13 +01:00

Sergey Sharybin commented

2014-11-25 15:19:13 +01:00

Added subscribers: @JoshuaLeung, @Sergey

Sergey Sharybin commented

2014-11-25 15:19:13 +01:00

@JoshuaLeung, do you mind having a look here? Afraid you know history of f-curve modifiers better than anyone here :)

Joshua Leung commented

2014-11-26 01:03:57 +01:00

It's a bit of an unfortunate consequence of the way that the way these things work/are set up.

All rotations are internally stored in radians. So a "value" of 1.0 for the property corresponds to 1.0 radians = ~57.3 degrees.

It's only for buttons that these values get translated to degrees, and that's only if the unit conversion code is enabled AND the UI code knows that the property represents a rotation. Unfortunately, the UI code doesn't know about how to handle these sorts of conversions when the type of a property depends on what the owner of that property is attached to (i.e. amplitude property of a FModifier on a rotation FCurve).

A similar situation showed up a few releases back with constraints; I'm not convinced that the solution applied there is necessarily the right solution, as this sort of thing is just going to keep occurring again and again with such interfaces.

It's a bit of an unfortunate consequence of the way that the way these things work/are set up. All rotations are internally stored in radians. So a "value" of 1.0 for the property corresponds to 1.0 radians = ~57.3 degrees. It's only for buttons that these values get translated to degrees, and that's only if the unit conversion code is enabled AND the UI code knows that the property represents a rotation. Unfortunately, the UI code doesn't know about how to handle these sorts of conversions when the type of a property depends on what the owner of that property is attached to (i.e. amplitude property of a FModifier on a rotation FCurve). A similar situation showed up a few releases back with constraints; I'm not convinced that the solution applied there is necessarily the right solution, as this sort of thing is just going to keep occurring again and again with such interfaces.

Julian Eisel commented

2014-11-26 01:13:22 +01:00

Added subscriber: @JulianEisel

Sergey Sharybin commented

2014-12-26 11:09:47 +01:00

Changed status from 'Open' to: 'Archived'

Sergey Sharybin closed this issue

2014-12-26 11:09:47 +01:00

Sergey Sharybin commented

2014-12-26 11:09:47 +01:00

For me it sounds as a TODO related on improving interface in general.

Thanks for the report, but moving ti there now http://wiki.blender.org/index.php/Dev:2.5/Source/Development/Todo/Editors#AnimEditors

For me it sounds as a TODO related on improving interface in general. Thanks for the report, but moving ti there now http://wiki.blender.org/index.php/Dev:2.5/Source/Development/Todo/Editors#AnimEditors

José Contreras commented

2016-12-23 23:26:32 +01:00

Added subscriber: @Esparadrapo