Re: plotted circle is regular decagon on opengl graphics

[Rik]

On 08/11/2016 12:25 PM, address@hidden wrote:

Subject:

Re: plotted circle is regular decagon on opengl graphics

From:

Daniel J Sebald <address@hidden>

Date:

08/11/2016 12:25 PM

To:

Andreas Weber <address@hidden>, Tatsuro MATSUOKA <address@hidden>, "address@hidden" <address@hidden>

List-Post:

<mailto:address@hidden>

Precedence:

list

MIME-Version:

1.0

References:

<address@hidden> <address@hidden>

In-Reply-To:

<address@hidden>

Message-ID:

<address@hidden>

Content-Type:

multipart/mixed; boundary="------------020400010901030207090104"

Message:

6

On 08/11/2016 03:12 AM, Andreas Weber wrote:

Hi Tatsuro,

Am 11.08.2016 um 08:17 schrieb Tatsuro MATSUOKA:

Graphics tool kit: qt and fltk
plot (1:1, 'o', 'markersize', 24)
print -depsc circle.eps
plotted circle is regular decagon (ugly).

The code to draw the markers with OpenGL is in gl-render.cc
opengl_renderer::make_marker_list

     case 'o':
       {
         double ang_step = M_PI / 5;

         glBegin ((filled ? GL_POLYGON : GL_LINE_LOOP));
         for (double ang = 0; ang < (2*M_PI); ang += ang_step)
           glVertex2d (sz*cos (ang)/2, sz*sin (ang)/2);
         glEnd ();
       }

As you can see the decagon is hardcoded. AFAIK there is no "draw circle"
command in OpenGL so the only way to get a smoother circle would be to
increase the numer of segments.

I'm surprised there is no type of primitive for arcs in OpenGL (or its utility library GLU) so that hardware/CPU optimization on graphics card drivers could be used.

Yes, it is surprising since it is an important basic shape.  However, try Google for "draw circle OpenGL" and you will find that everyone is constructing circles from line segments.

The performance cost is the cos/sin computations, not the underlying driver.  The driver essentially sweeps out the same arc no matter how fine the angular resolution is (i.e., it may draw more polygon lines, but the individual polygon lines are shorter so don't need as many computations to complete).

I'm happy to have something different in place.  The procedure here (http://slabode.exofire.net/circle_draw.shtml) uses only multiplications and additions and is possibly faster than either the trig approach or the sqrt approach.  Probably easy to benchmark all three options since the only thing that needs testing is the creation of the vertex points, not the actual calls to OpenGL.

--Rik