Smalltalk › Squeak › Squeak - Dev

About Morphic-cmm.442

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

5 messages Options

Chris Muller-3

About Morphic-cmm.442

This is a little fix from Henry Johansen which, for me, improves the
rendering of rotated Morphs with respect to translucency. Today when
Morphs are rotated they are always drawn with the Form paint rule, but
Henriks change allows Form blend for 32-bit displays.

It seems to work, and improves the look of my applications. You can
see the difference easily by rotating a simple RectangleMorph that is
alpha-colored..

However, as there are certainly more-qualified experts who may wish to
comment on the subject first, I thought I would start it at the Inbox.

- Chris

Bert Freudenberg

Re: About Morphic-cmm.442

On 14.06.2010, at 05:28, Chris Muller wrote:

> This is a little fix from Henry Johansen which, for me, improves the
> rendering of rotated Morphs with respect to translucency. Today when
> Morphs are rotated they are always drawn with the Form paint rule, but
> Henriks change allows Form blend for 32-bit displays.
>
> It seems to work, and improves the look of my applications. You can
> see the difference easily by rotating a simple RectangleMorph that is
> alpha-colored..
>
> However, as there are certainly more-qualified experts who may wish to
> comment on the subject first, I thought I would start it at the Inbox.
>
> - Chris

Etoys depends on primary colors being preserved under rotation. It uses both rotation and color tests a lot.

When I rotate a sketch that had not had translucency before, it would have none after rotating. With your patch, its edges get fuzzy:

vs

So this solution is not general enough. If there was translucency before, it should be preserved under rotation, yes. But if there was none, it should not be introduced. I think there is a way to have your cake and eat it, too, though I'm not quite sure how atm.

- Bert -

PastedGraphic-2.png (4K) Download Attachment

PastedGraphic-3.png (4K) Download Attachment

Chris Muller-3

Re: About Morphic-cmm.442

Hmm, I noticed that the "fuzzy edges" only seem to be there only when
it's rotation is not a multiple of 90-degrees.

Another observation: That when it is exactly 0-degrees (straight up),
the translucent color is true, but at all other angles, it is somewhat
"darkened". I'm not sure why, but there is some "shadow" logic in
that same method..

Also, I don't notice any fuzzy edges on standard morphs (Polygon and
Rectangle), regardless of angle, just the SketchMorph like you pointed
out.

Is your objection about visual-quality of those rotated sketches or
are you saying those fuzzy edges could affect color-testing logic in
eToy programs? It sounds like you are saying the latter..

It would indeed be nice if we can have our cake and eat it too..

- Chris

2010/6/14 Bert Freudenberg <[hidden email]>:

> On 14.06.2010, at 05:28, Chris Muller wrote:
>
>> This is a little fix from Henry Johansen which, for me, improves the
>> rendering of rotated Morphs with respect to translucency. Today when
>> Morphs are rotated they are always drawn with the Form paint rule, but
>> Henriks change allows Form blend for 32-bit displays.
>>
>> It seems to work, and improves the look of my applications. You can
>> see the difference easily by rotating a simple RectangleMorph that is
>> alpha-colored..
>>
>> However, as there are certainly more-qualified experts who may wish to
>> comment on the subject first, I thought I would start it at the Inbox.
>>
>> - Chris
>
> Etoys depends on primary colors being preserved under rotation. It uses both rotation and color tests a lot.
>
> When I rotate a sketch that had not had translucency before, it would have none after rotating. With your patch, its edges get fuzzy:
>
>
> vs
>
>
> So this solution is not general enough. If there was translucency before, it should be preserved under rotation, yes. But if there was none, it should not be introduced. I think there is a way to have your cake and eat it, too, though I'm not quite sure how atm.
>
> - Bert -
>
>
>
>
>
>

Bert Freudenberg

Re: About Morphic-cmm.442

On 14.06.2010, at 22:59, Chris Muller wrote:

> Hmm, I noticed that the "fuzzy edges" only seem to be there only when
> it's rotation is not a multiple of 90-degrees.

Right.

> Another observation: That when it is exactly 0-degrees (straight up),
> the translucent color is true, but at all other angles, it is somewhat
> "darkened". I'm not sure why, but there is some "shadow" logic in
> that same method..
>
> Also, I don't notice any fuzzy edges on standard morphs (Polygon and
> Rectangle), regardless of angle, just the SketchMorph like you pointed
> out.
>
> Is your objection about visual-quality of those rotated sketches or
> are you saying those fuzzy edges could affect color-testing logic in
> eToy programs? It sounds like you are saying the latter..

The latter.

- Bert -

> It would indeed be nice if we can have our cake and eat it too..
>
> - Chris
>
>
> 2010/6/14 Bert Freudenberg <[hidden email]>:
>> On 14.06.2010, at 05:28, Chris Muller wrote:
>>
>>> This is a little fix from Henry Johansen which, for me, improves the
>>> rendering of rotated Morphs with respect to translucency. Today when
>>> Morphs are rotated they are always drawn with the Form paint rule, but
>>> Henriks change allows Form blend for 32-bit displays.
>>>
>>> It seems to work, and improves the look of my applications. You can
>>> see the difference easily by rotating a simple RectangleMorph that is
>>> alpha-colored..
>>>
>>> However, as there are certainly more-qualified experts who may wish to
>>> comment on the subject first, I thought I would start it at the Inbox.
>>>
>>> - Chris
>>
>> Etoys depends on primary colors being preserved under rotation. It uses both rotation and color tests a lot.
>>
>> When I rotate a sketch that had not had translucency before, it would have none after rotating. With your patch, its edges get fuzzy:
>>
>>
>> vs
>>
>>
>> So this solution is not general enough. If there was translucency before, it should be preserved under rotation, yes. But if there was none, it should not be introduced. I think there is a way to have your cake and eat it, too, though I'm not quite sure how atm.
>>
>> - Bert -
>>
>>
>>
>>
>>
>>
>

Karl Ramberg

Re: About Morphic-cmm.442

Related to this issue:

http://bugs.squeak.org/view.php?id=2241

Karl

On Mon, Jun 14, 2010 at 11:10 PM, Bert Freudenberg <[hidden email]> wrote:

On 14.06.2010, at 22:59, Chris Muller wrote:

> Hmm, I noticed that the "fuzzy edges" only seem to be there only when
> it's rotation is not a multiple of 90-degrees.

Right.

> Another observation: That when it is exactly 0-degrees (straight up),
> the translucent color is true, but at all other angles, it is somewhat
> "darkened". I'm not sure why, but there is some "shadow" logic in
> that same method..
>
> Also, I don't notice any fuzzy edges on standard morphs (Polygon and
> Rectangle), regardless of angle, just the SketchMorph like you pointed
> out.
>
> Is your objection about visual-quality of those rotated sketches or
> are you saying those fuzzy edges could affect color-testing logic in
> eToy programs? It sounds like you are saying the latter..

The latter.

- Bert -

> It would indeed be nice if we can have our cake and eat it too..
>
> - Chris
>
>
> 2010/6/14 Bert Freudenberg <[hidden email]>:
>> On 14.06.2010, at 05:28, Chris Muller wrote:
>>
>>> This is a little fix from Henry Johansen which, for me, improves the
>>> rendering of rotated Morphs with respect to translucency. Today when
>>> Morphs are rotated they are always drawn with the Form paint rule, but
>>> Henriks change allows Form blend for 32-bit displays.
>>>
>>> It seems to work, and improves the look of my applications. You can
>>> see the difference easily by rotating a simple RectangleMorph that is
>>> alpha-colored..
>>>
>>> However, as there are certainly more-qualified experts who may wish to
>>> comment on the subject first, I thought I would start it at the Inbox.
>>>
>>> - Chris
>>
>> Etoys depends on primary colors being preserved under rotation. It uses both rotation and color tests a lot.
>>
>> When I rotate a sketch that had not had translucency before, it would have none after rotating. With your patch, its edges get fuzzy:
>>
>>
>> vs
>>
>>
>> So this solution is not general enough. If there was translucency before, it should be preserved under rotation, yes. But if there was none, it should not be introduced. I think there is a way to have your cake and eat it, too, though I'm not quite sure how atm.
>>
>> - Bert -
>>
>>
>>
>>
>>
>>
>