Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[T.Pierret]

Hi,
In order to create a "slidable" area with texts and images content, I first drew the content:

Code: [Select]

CLEAR_COLOR_RGB()
CLEAR(1 ,1 ,1)
CLEAR_TAG( 0 )
...
BITMAP_SOURCE()
BITMAP_LAYOUT()
BITMAP_SIZE()
BEGIN( BITMAPS )
...
CMD_TEXT()
...
Then I defined the area where the touch shall be used for scrolling :

Code: [Select]

COLOR_MASK(0 ,0 ,0 ,0)
TAG( value )
BEGIN(RECTS)
...
Touching the scrolling area, I expected to get the defined tag value. However, if the touch happens on a text or an image, the tag value 0 is obtained.
Is there a way to not get the tag value 0 and obtain the tag of the content area ?

Thanks in advance for any support.
Best regards

[Rudolph]

You can use the same tag value for several objects.
In your case I would just use an rectangle with the background color and the other objects on top of that.

TAG(10)
COLOR_RGB(Background)
..rect..window-size
..bitmaps
COLOR_RGB(Text)
CMD_TEXT()
TAG(0)

I usually do this with sliders to make them more useable by increasing the touch-area.

Oh, annother thing, avoid tag values below 10, for some reasons these do not give reliable touch events.

[BRT Community]

Hello,

Rudolphs suggestion is a suitable solution.
Is there any particular reason you would wish to tag the images and scrolling section with different tag values?

Best Regards,
BRT Community

[avanti]

Oh, annother thing, avoid tag values below 10, for some reasons these do not give reliable touch events.

Is this really true!?  If so, I have a lot of code to revise.

[T.Pierret]

Thanks for those answers.
The purpose is to not tag the content of the area while tagging the whole scrollable area.

I tried the solution proposed by @Rudolph, but it did not work. It is even worst since touching the scroll area itself also returns the 0 tag value 
To be noted : between the content and the scrollable area drawings, there are other tagged items (title bar, menu bar, ...), that overlay the content if this is too big for the display area.

Regards

[Rudolph]

Oh, annother thing, avoid tag values below 10, for some reasons these do not give reliable touch events.

Is this really true!?  If so, I have a lot of code to revise.

I am not quite sure.
I tested this the last time back in January with an EVE3-43G and before that found this behaviour with FT813.
At a tag-value of 1 the value was toggeling between 0 und 1.
At a tag-value of 5 the value was toggeling between 4 und 5.
At a tag-value of 9 the value was toggeling between 8 and 9.
Using other values was stable: 2, 3, 4, 6, 7, 8, 11, 15 and 19.
It was reproduceable, I had seen this before and I got a bug-report regarding this.

I just setup a new test with an EVE3-50G and the value is stable.
It just is freaking stable - as it should be.
No idea why, could not explain it back then either, just reading REG_TOUCH_TAG has a rather low potential for failure.

So well, ignore this, sorry.

[darkjezter]

No matter where you touch the display, I'm pretty sure you should get a tag value of some kind.  When you clear the display, if you've set the bit to clear the tag, the the entire display will return the tag value determined by CLEAR_TAG which is zero by default.  As you draw, tag values are replaced in the region drawn by the current value of TAG (1 to 255 according to the doc, though Rudolph's example also uses zero).  Thus, if you draw your scroll region last, then any touches in it should report that tag value.  The only exception to this will be when using TAG_MASK, which you don't appear to be.

If you've tried Rudolph's suggestion and as a result are seeing tags of zero for your scrollable region, then it sounds like TAG(0) is supported in spite of the documentation, and you've either failed to set tag again before drawing your scrollable area, or drawn something over top of the scrollable area with TAG still set to zero.

[T.Pierret]

Hi,

You are right, the behaviour is not the one I explained. However I often observe a tag value of 0 at the beginning of a sliding movement. Here is some recorded samples of the case:

Event	Time	Smoothstep	Read	Tag	Pressure
"Touch pressed"	18516	172 203	172  203	0	957
"Touch moved"	18521	172 202	473  148	0	1318
"Touch moved"	18526	172 201	501  148	0	1021
"Touch moved"	18531	172 200	534  140	0	899
"Touch moved"	18536	172 199	509  154	16	865
"Touch moved"	18793	237 190	568  150	16	416
"Touch moved"	19078	302 185	568  169	16	479
"Touch moved"	19335	342 187	546  198	16	477
"Touch moved"	19592	375 194	545  234	16	452
"Touch moved"	19849	406 204	565  260	16	564
"Touch moved"	20106	430 215	558  277	16	553
"Touch moved"	20363	449 227	551  293	16	535
"Touch moved"	20620	465 240	547  312	16	529
"Touch moved"	20877	477 256	543  342	16	607
"Touch moved"	21134	487 273	540  363	16	547
"Touch moved"	21391	497 289	548  373	16	556
"Touch released"	21648	505 302	32768 32768	16	32767

where the scrollable area (tag value=16) is defined as (80,80) -> (799,479).

Why the first tag values are 0 ? Yet the touch coordinates are well in the 16-tagged area.

Supplementary: why the 1st touch coordinates may be so different of the rest ?

Thanks in advance for any support.
Regards

[BRT Community]

Hello,

Can you provide an example display list of how you are tagging the different objects on your screen?
I would like to investigate this behaviour further.

Best Regards,
BRT Community

[BRT Community]

Hello,

The REG_TOUCH_TAGs default value is 0, however this normally indicates that no touch is detected. It may be that you are reading the register before the frame has fully scanned out.

Can you provide an example display list of how you are tagging the different objects on your screen?
I would like to investigate this behaviour further.

Best Regards,
BRT Community

[Rudolph]

though Rudolph's example also uses zero).

As zero is an invalid value and also the default CLEAR_TAG value, I just use TAG(0)
for objects I do not need touch events for.

[darkjezter]

"Why the first tag values are 0 ?"

I got this one.  Reported tag values only update once per video frame (due to the scanout mentioned above)  So touch will almost always be reported first, but expect the corresponding tag update to lag until the next video frame interval.

In my use of this, I've implemented a sticky tag system.  Touch events are tracked separately from tags, and reported tag values of zero are more or less ignored.  Once a non-zero tag comes in, I latch it internally, remembering its value until touch is released.  New tag values besides the one selected are ignored (only activate one control per touch event), and I also make use of this mechanism to keep a button activated until touch is released, or the finger drags outside of the control (based on position, not tag) plus a margin.  All done without additional drawing commands to the display, and also allowing overlapping release areas between adjacent controls.

Most of this was made necessary by the fact that we're using a resistive touchscreen, and that we need the screens to be usable with work gloves on.

[T.Pierret]

Hi,

Here is the simplest display list where the trouble can be observed. It's dynamically built. Hence it is extracted from our mockup code. Small errors may appear in coordinates (hope not).

Code: [Select]

CLEAR_COLOR_RGB(10, 10, 60)
CLEAR(1 ,1 ,1)
CLEAR_TAG( 0 )
TAG( CONTENT_TAG_ID )                     // Start drawing content
SAVE_CONTEXT()
draw_image(KITE_ICON_ID, 88, 88, 64, 64)
RESTORE_CONTEXT()
SAVE_CONTEXT()
COLOR_RGB(255, 255, 255)
CMD_TEXT(140, 96, 26, 0, "Blablabla")
CMD_TEXT(252, 136, 26, 0, "Blabla: blabla bla bla")
RESTORE_CONTEXT()
SAVE_CONTEXT()
draw_image(SAIL_ICON_ID, 752, 200, 64, 64)
RESTORE_CONTEXT()
TAG( 0 )                                  // End of the content
SAVE_CONTEXT()
BEGIN(RECTS) );
COLOR_RGB(80, 80, 80)                     // Title bar background
VERTEX2F(0 * 16, 0 * 16)
VERTEX2F(79 * 16, 799 * 16)
COLOR_RGB(60, 60, 60)                     // Menu bar background
VERTEX2F(0 * 16, 80 * 16)
VERTEX2F(79 * 16, 479 * 16)
COLOR_RGB(0, 0, 0)                        // Selection area background
VERTEX2F(0 * 16, y * 16)
VERTEX2F(79 * 16, 79 * 16)
END()
RESTORE_CONTEXT()
SAVE_CONTEXT()               // Status led
draw_image(STATUS_GREEN_ID, 704, 16, 40, 40);
RESTORE_CONTEXT()
SAVE_CONTEXT()                            // Icons
TAG( HOME_TAG_ID )                        // Home icon
draw_image(HOME_ICON_ID, 8, 8, 64, 64 );
RESTORE_CONTEXT()
SAVE_CONTEXT()
TAG( BACK_TAG_ID )                        // Back icon
draw_image(BACK_ICON_ID, 88, 8, 64, 64);
TAG( SETTINGS_TAG_ID )                    // Settings icon
draw_image(SETTINGS_ICON_ID, 712, 8, 64, 64);
TAG( mpCurrentState->tags[i] )            // Function icon
draw_image(FUNCTION_ICON_ID, 8, 88, 64, 64);
RESTORE_CONTEXT()
COLOR_RGB(240, 240, 240)                  // Title
CMD_TEXT(400, 8, 31, OPT_CENTERX, "Tests")
COLOR_MASK(0 ,0 ,0 ,0)                    // Content area
TAG(CONTENT_TAG_ID)
BEGIN(RECTS)
VERTEX2F(80 * 16, 80 * 16)
VERTEX2F(799 * 16, 479 * 16)
END()
where the draw_image function acts as this:

Code: [Select]

draw_image( id, x, y, w, h ):
  BITMAP_SOURCE( images[id].ram_address )
  BITMAP_LAYOUT( images[id].pImage->format, images[id].pImage->stride, images[id].pImage->height )
  BITMAP_SIZE( NEAREST, BORDER, BORDER, images[id].pImage->width, images[id].pImage->height )
  BEGIN( BITMAPS )
  if ( images[id].pImage->width != w )
  {
    factor = images[id].pImage->width * 256 / w;
    BITMAP_TRANSFORM_A(factor)
  }
  if ( images[id].pImage->height != h )
  {
    factor = images[id].pImage->height * 256 / h;
    BITMAP_TRANSFORM_E(factor)
  }
  VERTEX2F( x * 16, y * 16 )
  END()
All the icons are ARGB2 images (white drawing on transparent background), but the status led image is an ARGB555 image.

Thanks in advance for your support
Best regards

[T.Pierret]

Hi,

As stated in the http://www.brtcommunity.com/index.php?topic=143.15, the trouble is solved by using data from the REG_TOUCH_TAG_XY register instead of the REG_TOUCH_SCREEN_XY register. The tag value now appears as expected by the real touch.

Thanks for your support.

Regards.