DO NOT MERGE: Draft for PaintSweepGradient update #362
Conversation
|
See #363 for suggested alternate image. |
|
After we've reviewed and have consensus, we still should wait on merging as we'll need to prepare a separate doc for submission to the MPEG AHG as proposed revision for the new edition. (It's too late to make changes in Amd 2.) |
Are you suggesting to wait with merging until the MPEG AHG has accepted the proposed change, too? |
…ations simplified and clarified
See comment above, updated figures with more clearly separating the color stops. |
|
@justvanrossum Would you have time to review this as well and see if this resolves any ambiguities from our discussion in BlackFoundryCom/black-renderer#17? Thanks in advance. |
No, that's not what I meant. These revisions are in OFF_AMD2_WD, which was proposed content for Amendment 2 of OFF. But it's too late to make further changes to amendment 2. So, we'll need to prepare a separate document for that. |
OFF_AMD2_WD.md
Outdated
| below 0° and equal to or above 360° are not sampled for drawing the rays. | ||
|
|
||
| If the ColorLine's extend mode is reflect or repeat and start and end angle are | ||
| equal, nothing must be drawn. |
There was a problem hiding this comment.
To keep consistent with ISO speak as elsewhere in this doc:
"... nothing shall be drawn."
There was a problem hiding this comment.
I think some items might need to be reordered. Fig 5.31 and Fig 5.32 are back to back, but 5.32 pertains to the text before 5.31. Fig 5.31 should follow Fig 5.30, but the text that pertains to 5.31 comes earlier.
I've added comments at relevant points below with suggestions.
OFF_AMD2_WD.md
Outdated
| is aligned with the start angle, and stop offset 1 is aligned with the end | ||
| angle, independent of which angle is larger. | ||
|
|
||
| NOTE: If the start angle is less than or equal to the end angle, the color line |
There was a problem hiding this comment.
This note should move after fig 5.31.
There was a problem hiding this comment.
In the case of figure 5.30, and if I am not mistaken, in other cases as well, we describe textually what happens, then we illustrate with a figure after that, don't we? In other words, shouldn't we have the note before the figure 5.31 (which is the one for inversion) here as well?
I moved the note, so if I am not mistaken, no we have the following order:
- Figure 5.30 illustrates...
- Figure 5.30,
- NOTE: If the start angle is less than or equal to the end angle...see figure 5.31.
- Figure 5.31
- "Not more than one full...see figure 5.32 for an example"
- Figure 5.32.
OFF_AMD2_WD.md
Outdated
| 230°, using a `ColorLine` with color stops 0 and 1 for yellow and red and pad | ||
| extend mode.** | ||
|
|
||
| Not more than one full rotation is drawn and there is no overlap in drawing for |
There was a problem hiding this comment.
Move the note earlier about stop angle > end angle here. (Perhaps drop the "NOTE:" prefix?) Then move Fig 5.32 to directly follow (and re-number to 5.31).
The this paragraph ("Not more than..." follows that figure, with fig 5.31 (re-numbered to 5.32) after that.
There was a problem hiding this comment.
See above, note moved, but I don't understand the request for renumbering then.
These changes can only go into the new Working Draft of the next edition OFF, and we do need to submit an input contribution summarizing the proposed changes. |
| For sampling and interpolating colors, the color line is aligned to an | ||
| infinitely spiraling circular arc around the center point, with arbitrary | ||
| radius. The position 0° on the arc means the direction of the positive x-axis, | ||
| 360° means one full counter-clockwise rotation. The ColorLine's stop offset 0 |
There was a problem hiding this comment.
"... and 360° means one full counter-clockwise rotation. The color line’s stop offset 0..."
| infinitely spiraling circular arc around the center point, with arbitrary | ||
| radius. The position 0° on the arc means the direction of the positive x-axis, | ||
| 360° means one full counter-clockwise rotation. The ColorLine's stop offset 0 | ||
| is aligned with the start angle, and stop offset 1 is aligned with the end |
There was a problem hiding this comment.
"... is aligned with the start angle and stop offset 1 is aligned..."
(delete the comma)
| is aligned with the start angle, and stop offset 1 is aligned with the end | ||
| angle, independent of which angle is larger. | ||
|
|
||
| Outside the defined interval of the ColorLine, the color value of a position on |
There was a problem hiding this comment.
This paragraph is repeating information provided in the color line section. We don't repeat this explanation when linear or radial gradients are described; we simply say that at some point the color of the color line is used. I think this paragraph can be deleted.
(In the existing spec, some similar details are explained, but that was motivated by the different semantics for sweep gradients that was being used.)
| Lines for more details. In effect, this means that the spiraling circular arc | ||
| can be sampled for colors outside the defined ColorLine interval. | ||
|
|
||
| To draw the sweep gradient, for each position along the circular arc, starting |
There was a problem hiding this comment.
"To draw the sweep gradient, for each position along the circular arc starting from 0° and up to but not including 360°, a ray from the center outward is painted with the color of the color line where the ray intersects the circular arc at that particular angle. Angle positions below 0° and equal to or above 360° are not sampled for drawing the rays."
| @@ -577,7 +577,7 @@ A sweep gradient provides a gradation of colors that sweep around a center | |||
| point. For a given color on a color line, that color projects as a ray from the | |||
There was a problem hiding this comment.
Lines 306 and 307 (note about repeat/mirror modes being special for sweep) need to be deleted.
| of 230°.](images/colr_conic_gradient_start_stop_angles.png) | ||
|
|
||
| **Figure 5.30 A sweep gradient with start angle of 110° and an end angle of | ||
| 230°, using a `ColorLine` with color stops 0 and 1 for yellow and red and pad |
There was a problem hiding this comment.
"... using a color line..."
OFF_AMD2_WD.md
Outdated
| @@ -1873,6 +1899,10 @@ mapping data; see 5.7.11.4 for details. | |||
| Angles are expressed in counter-clockwise degrees from the direction of the | |||
| positive x-axis in the design grid. | |||
|
|
|||
| NOTE: A bias of 1.0 is used in the representation of start and end angles of | |||
| sweep gradients. This is not done, however, for angles in rotate and skew | |||
| transforms (5.7.11.2.5.11 and 5.7.11.2.5.12). | |||
There was a problem hiding this comment.
it would be nice to explain why this +1.0 bias is only applied for sweep gradient angles and not for other rotate or skew angles
There was a problem hiding this comment.
Suggestion:
NOTE: To allow for representation of +360°, a bias of 1.0 is used in the representation of start and end angles of sweep gradients. ...
There was a problem hiding this comment.
To allow for representation of +360°, ...
👍
An example would also be nice
There was a problem hiding this comment.
Just to confirm.
A normal (i.e. non biased) F2Dot14-encoded-fraction-of-half-circle angle (as used in PaintRotate, PaintSkewGradient, etc.) allows to encode -360 <= v < 360 degrees, thus it can't represent a positive full +360 angle.
Whereas the proposed "biased" F2Dot14 angle used in PaintSweepGradient would allow to encode angles in degrees (-360 + 180) <= v < (360 + 180) => -180 <= v < 540.
E.g. an F2Dot14 value of 0x0 means 0 degrees for the non-biased angle, or 180 degrees for the biased angle.
Or a F2Dot14 value of -16384 (-0x4000) means -180 degrees for the non-biased angle, or 0 degrees for the biased angle.
There was a problem hiding this comment.
yet another example, 225 degree angle represented using the biased F2Dot14 angle:
>>> v = 4096
>>> bits = 14
>>> angle = (v / (1 << bits) + 1.0) * 180
>>> print(angle)
225.0
>>> v2 = (angle / 180 - 1.0) * (1 << 14)
>>> assert v2 == vThere was a problem hiding this comment.
Here's my draft revision (as porting to OT 1.9.1):
NOTE: To allow for a representation of +360°, a bias of 1.0 is used in the representation of stand and end angles of sweep gradients. For example, an F2DOT14 value of -2.0 (0x8000) represents -180°; an F2DOT14 value of 0.0 (0x0000) represents +180°; an F2DOT14 value of 0.25 (0x1000) represents +225°; an F2DOT14 value of 1.0 (0x4000) represents +360°. However, a bias is not used for representation of angles in rotate or skew transforms (5.7.11.2.5.11, 5.7.11.2.5.12).
There was a problem hiding this comment.
Though I wonder if it's really worth adding the extra complexity of having two kinds of F2Dot14 angles, a biased one for Sweep, and the non-biased one for Rotate and Skew. Why is +360° special, and why does it only matter for Sweep and not, say, for Rotate? What angle intervals can't be encoded using the regular angle without bias? If one wants to do a full rotation 0..360 one can always write it as -360..0 and achieve the same visual result, I think.
There was a problem hiding this comment.
Another option could be that we change the type of the angles from Fixed2Dot14 to a new Fixed4Dot12 for example, always a 16-bit type but it would have slightly smaller fractional part, which would still allow up to 3 decimal digits when converted to/from float, and it could allow to encode values between -16.0 <= v < 15.9998, which in degrees means -2880.0° <= angle < +2879.95605°
OFF_AMD2_WD.md
Outdated
| @@ -1873,6 +1899,10 @@ mapping data; see 5.7.11.4 for details. | |||
| Angles are expressed in counter-clockwise degrees from the direction of the | |||
| positive x-axis in the design grid. | |||
|
|
|||
| NOTE: A bias of 1.0 is used in the representation of start and end angles of | |||
| sweep gradients. This is not done, however, for angles in rotate and skew | |||
| transforms (5.7.11.2.5.11 and 5.7.11.2.5.12). | |||
There was a problem hiding this comment.
Here's my draft revision (as porting to OT 1.9.1):
NOTE: To allow for a representation of +360°, a bias of 1.0 is used in the representation of stand and end angles of sweep gradients. For example, an F2DOT14 value of -2.0 (0x8000) represents -180°; an F2DOT14 value of 0.0 (0x0000) represents +180°; an F2DOT14 value of 0.25 (0x1000) represents +225°; an F2DOT14 value of 1.0 (0x4000) represents +360°. However, a bias is not used for representation of angles in rotate or skew transforms (5.7.11.2.5.11, 5.7.11.2.5.12).
|
@PeterConstable Note sure why GitHub does not let me comment directly in the thread above, but just wanted to point you to the typo in the edited version:
|
No description provided.