Opened 18 years ago
Closed 18 years ago
Last modified 18 years ago
#4215 closed defect (fixed)
pip colors not right after changeset 14692
| Reported by: | skamithi | Owned by: | skamithi |
|---|---|---|---|
| Priority: | minor | Milestone: | 0.21 |
| Component: | mythtv | Version: | head |
| Severity: | medium | Keywords: | |
| Cc: | Ticket locked: | no |
Description
spoke to danielk about the issue. PiP compositing is wrong. patch just exposed the bug. Working on a bug fix.
--excerpt from daniel's email--
the pip compositing, VideoOutput::ShowPip() is the problem. In particular, this code is completely bogus:
for (int i = 0; i < piph; i++)
{
memcpy(frame->buf + (i + yoff) * frame->width + xoff,
pipbuf + i * pipw, pipw);
}
// Copy U & V (half plane chroma values)
xoff /= 2;
yoff /= 2;
unsigned char *uptr = frame->buf + frame->width * frame->height;
unsigned char *vptr = frame->buf + frame->width * frame->height *
5 / 4;
int vidw = frame->width / 2;
This should be using frame->buf, frame->pitches[], and frame->offsets[], and not looking at the width and height at all to set this up.
As is this code was already not working before [14692] on ATI and with the proprietary drivers nor with some Intel drivers, and would cause segfaults with VIA hardware. The problem is just more visible now
Attachments (2)
Change History (5)
by , 18 years ago
| Attachment: | 4215_fix.diff added |
|---|
comment:2 by , 18 years ago
| Resolution: | → fixed |
|---|---|
| Status: | new → closed |
comment:3 by , 18 years ago
Replying to Roo <roo.watt@gmail.com>:
Was this fixed in [15303]?
No it was not. still had the same problem after updating > [15303].

attempt to implemented suggested changes.works fine on i945 and fx5200 before and after [14692]