zapping-misc — General discussion

Re: [Zapping-misc] [PATCH] Conditionally use ISO C99 variadic macro definitions

[Michael S. <msc...@gm...>]

On Thu, 2008-02-21 at 23:19 +0100, Leon Woestenberg wrote:
> the patch below allows to compile against the latest libzvbi public
> header file without warnings using ISO C99, fixing
> build breaks like these:
> 
> gcc -Wall -pedantic -Werror -g -pg -O0 -std=c99
> -I/home/leon/sandbox/libzvbi/install/usr/include -c teletext.c -o
> teletext.o
> In file included from teletext.c:20:
> /home/leon/sandbox/libzvbi/install/usr/include/libzvbi.h:53:28: error:
> ISO C does not permit named variadic macros

Thanks for the patch! However I found my code was also incompatible with
strict ISO C89, so I rewrote the macros without variable arguments.

Regards,
Michael

[Zapping-misc] [PATCH] Conditionally use ISO C99 variadic macro definitions

[Leon W. <le...@ma...>]

Hello Michael,

the patch below allows to compile against the latest libzvbi public
header file without warnings using ISO C99, fixing
build breaks like these:

gcc -Wall -pedantic -Werror -g -pg -O0 -std=c99
-I/home/leon/sandbox/libzvbi/install/usr/include -c teletext.c -o
teletext.o
In file included from teletext.c:20:
/home/leon/sandbox/libzvbi/install/usr/include/libzvbi.h:53:28: error:
ISO C does not permit named variadic macros
/home/leon/sandbox/libzvbi/install/usr/include/libzvbi.h:54:27: error:
ISO C does not permit named variadic macros
/home/leon/sandbox/libzvbi/install/usr/include/libzvbi.h:70:30: error:
ISO C does not permit named variadic macros

gcc --version
gcc (GCC) 4.1.3 20070929 (prerelease) (Ubuntu 4.1.2-16ubuntu2)

Please consider applying this patch. For legal completeness, I hereby
submit it under the LGPLv2+ license.

Regards,

Leon Woestenberg.


diff -u -r1.10 macros.h
--- src/macros.h        19 Feb 2008 00:35:20 -0000      1.10
+++ src/macros.h        21 Feb 2008 22:06:16 -0000
@@ -46,8 +46,13 @@
 #endif
 
 #if (__GNUC__ == 3 && __GNUC_MINOR__ >= 3) || __GNUC__ >= 4
-#  define _vbi_nonnull(args...) nonnull(args)
-#  define _vbi_format(args...) format(args)
+#  if (__STDC_VERSION__ == 199901L)
+#    define _vbi_nonnull(...) nonnull(__VA_ARGS__)
+#    define _vbi_format(...) format(__VA_ARGS__)
+#  else
+#    define _vbi_nonnull(args...) nonnull(args)
+#    define _vbi_format(args...) format(args)
+#  endif
 #else
 #  define _vbi_nonnull(args...) ,
 #  define _vbi_format(args...) ,
@@ -63,7 +68,11 @@
 
 #if __GNUC__ >= 2
 #  define _vbi_inline static __inline__
-#  define _vbi_attribute(args...) __attribute__(args)
+#  if (__STDC_VERSION__ == 199901L)
+#    define _vbi_attribute(...) __attribute__(__VA_ARGS__)
+#  else
+#    define _vbi_attribute(args...) __attribute__(args)
+#  endif
 #else
 #  define _vbi_inline static
 #  define _vbi_attribute(args...)

Re: [Zapping-misc] io-sim.c/gen_teletext_b_row() sets 3 bytes never used in gen_teletext_b()

[Michael S. <msc...@us...>]

On Tue, 2008-02-12 at 16:13 +0100, Leon Woestenberg wrote:
> in using libzvbi to generate a more dynamic Teletext test page and
> stumbled across this:
> 
> io-sim.c/gen_teletext_b_row():
> 	return_buf[0] = 0x55;
> 	return_buf[1] = 0x55;
> 	return_buf[2] = 0x27;
> 
> However, these 3 bytes are ignored by the single caller:
> 
> io-sim.c/gen_teletext_b()
> 	gen_teletext_b_row (sim, buf);
> 	memcpy (&s->data, buf + 3, 42);
> 
> Is there historic reason for the 3 bytes, or did I miss something?

Yeah that's a refactoring relict. Way back the code was part of
signal_teletext() which calculates a complete Teletext wave form
including the clock run-in pulses and framing code.

Michael

[Zapping-misc] io-sim.c/gen_teletext_b_row() sets 3 bytes never used in gen_teletext_b()

[Leon W. <le...@ma...>]

Hello Michael,

in using libzvbi to generate a more dynamic Teletext test page and
stumbled across this:

io-sim.c/gen_teletext_b_row():
	return_buf[0] = 0x55;
	return_buf[1] = 0x55;
	return_buf[2] = 0x27;

However, these 3 bytes are ignored by the single caller:

io-sim.c/gen_teletext_b()
	gen_teletext_b_row (sim, buf);
	memcpy (&s->data, buf + 3, 42);

Is there historic reason for the 3 bytes, or did I miss something?

Regards,

Leon.

Re: [Zapping-misc] zvbi link error: object name conflicts

[<br...@gi...>]

>libtool attempts to find out if there are duplicate
>object file names with a "sort | sort -uc" sequence. Well there aren't
>any on the libtool command line or in the "ar t" output. Is it possible
>the check doesn't work as expected on your box?

Well, whaddya know.  My 'sort' (Coreutils 5.20) was broken.  sort -c
always said the input was unsorted.  It's been that way for 4 years;
I guess nobody ever needed it before.

I got a new 'sort' and Zvbi builds fine.

-- 
Bryan Henderson                                   San Jose, California

Re: [Zapping-misc] zvbi link error: object name conflicts

[Michael S. <msc...@us...>]

On Wed, 2008-01-23 at 16:33 +0000, Bryan Henderson wrote:
> >The make output looks ok to me, except the path above seems to be wrong.
> >The build directory is /mp/hdc4a/zvbi/src:
> >>> make[4]: Entering directory `/mp/hdc4a/zvbi/src'
> >so I would expect
> >.libs/libzvbi.lax/libzvbiinline.a//mp/hdc4a/zvbi/src/./.libs/libzvbiinline.a
> 
> /tmp2 is a symbolic link to /mp/hdc4a.
> 
> >Which object file names in libzvbiinline.a conflict ("ar
> >t /tmp2/zvbi/src/./.libs/libzvbiinline.a" should tell) or why is not
> >apparent.
> 
> 'ar t' shows the following contents:
> 
> libzvbiinline_la-decoder.o
> libzvbiinline_la-exp-gfx.o

Same in my tree (no blank lines).

> That's also the contents of
> /tmp2/zvbi/src/.libs/libzvbi.lax/libzvbiinline.a/ .

Ok.

> I've attached the configure and make dialog after make clean.  It's a
> default 'configure'.  I gathered the above information afterward.

Thanks. (Wasn't mirrored on zapping-misc due to size.)

With gcc 2.95.2 and a /tmp2 symlink I get almost the same output, but
still not the
> copying selected object files to avoid basename conflicts...
and
> libtool: link: ERROR: object name conflicts: [...]
lines. In both cases libtool attempts to find out if there are duplicate
object file names with a "sort | sort -uc" sequence. Well there aren't
any on the libtool command line or in the "ar t" output. Is it possible
the check doesn't work as expected on your box?

mike> echo -e "foo\nbar" | sort | sort -uc && echo ok
ok
mike> echo -e "foo\nbar\nfoo" | sort | sort -uc && echo ok
sort: -:3: disorder: foo
mike> sort --version
sort (coreutils) 5.3.0

Michael

Re: [Zapping-misc] zvbi link error: object name conflicts

[Michael S. <msc...@us...>]

Hi Bryan,

On Mon, 2008-01-21 at 16:38 +0000, Bryan Henderson wrote:
> I'm having some libtool grief.  In creating libvbi.la, libtool manages
> to create the actual .a/.so library, then proceeds to create a directory
> libzvbi.lax/libzvbiinline.a/ and put the contents of libzvbiinline.a
> in it, then fails with the cryptic message:
> 
> libtool: link: ERROR: object name conflicts: .libs/libzvbi.lax/libzvbiinline.a//tmp2/zvbi/src/./.libs/libzvbiinline.a

The make output looks ok to me, except the path above seems to be wrong.
The build directory is /mp/hdc4a/zvbi/src:
> make[4]: Entering directory `/mp/hdc4a/zvbi/src'
so I would expect
.libs/libzvbi.lax/libzvbiinline.a//mp/hdc4a/zvbi/src/./.libs/libzvbiinline.a

Which object file names in libzvbiinline.a conflict ("ar
t /tmp2/zvbi/src/./.libs/libzvbiinline.a" should tell) or why is not
apparent. Perhaps if you make clean and show me the complete make
output, and how you configured.

Michael

[Zapping-misc] zvbi link error: object name conflicts

[<br...@gi...>]

I'm having some libtool grief.  In creating libvbi.la, libtool manages
to create the actual .a/.so library, then proceeds to create a directory
libzvbi.lax/libzvbiinline.a/ and put the contents of libzvbiinline.a
in it, then fails with the cryptic message:

libtool: link: ERROR: object name conflicts: .libs/libzvbi.lax/libzvbiinline.a//tmp2/zvbi/src/./.libs/libzvbiinline.a

I note that this string is the concatenation of a directory name and an
absolute file name (/tmp2/zvbi is my zvbi source tree).

I've never seen libtool do this .lax thing and I don't know what any of this
has to do with building libzvbi.la, so I'm lost.  Any ideas on how to fix or
work around this?

This is the zvbi 0.2.26 package.

Here are the make messages:

make  all-recursive
make[1]: Entering directory `/mp/hdc4a/zvbi'
Making all in m4
make[2]: Entering directory `/mp/hdc4a/zvbi/m4'
make[2]: Nothing to be done for `all'.
make[2]: Leaving directory `/mp/hdc4a/zvbi/m4'
Making all in src
make[2]: Entering directory `/mp/hdc4a/zvbi/src'
make  all-recursive
make[3]: Entering directory `/mp/hdc4a/zvbi/src'
Making all in dvb
make[4]: Entering directory `/mp/hdc4a/zvbi/src/dvb'
make[4]: Nothing to be done for `all'.
make[4]: Leaving directory `/mp/hdc4a/zvbi/src/dvb'
make[4]: Entering directory `/mp/hdc4a/zvbi/src'
/bin/sh ../libtool --tag=CC --mode=link gcc  -g -O2   -o libzvbi.la -rpath /home/bryanh/zvbi/lib -version-info 11:0:11 bit_slicer.lo cache.lo caption.lo conv.lo dvb_mux.lo dvb_demux.lo exp-html.lo exp-templ.lo exp-txt.lo exp-vtx.lo export.lo hamm.lo idl_demux.lo io.lo io-bktr.lo io-dvb.lo io-sim.lo io-v4l.lo io-v4l2.lo io-v4l2k.lo lang.lo misc.lo packet.lo teletext.lo page_table.lo pfc_demux.lo proxy-client.lo raw_decoder.lo sampling_par.lo search.lo ure.lo sliced_filter.lo tables.lo trigger.lo vbi.lo vps.lo wss.lo xds_demux.lo proxy-msg.lo libzvbiinline.la -lpthread  -lm   -lpthread  -lm 
rm -fr  .libs/libzvbi.lax .libs/libzvbi.so .libs/libzvbi.so.0 .libs/libzvbi.so.0.11.0
gcc -shared  .libs/bit_slicer.o .libs/cache.o .libs/caption.o .libs/conv.o .libs/dvb_mux.o .libs/dvb_demux.o .libs/exp-html.o .libs/exp-templ.o .libs/exp-txt.o .libs/exp-vtx.o .libs/export.o .libs/hamm.o .libs/idl_demux.o .libs/io.o .libs/io-bktr.o .libs/io-dvb.o .libs/io-sim.o .libs/io-v4l.o .libs/io-v4l2.o .libs/io-v4l2k.o .libs/lang.o .libs/misc.o .libs/packet.o .libs/teletext.o .libs/page_table.o .libs/pfc_demux.o .libs/proxy-client.o .libs/raw_decoder.o .libs/sampling_par.o .libs/search.o .libs/ure.o .libs/sliced_filter.o .libs/tables.o .libs/trigger.o .libs/vbi.o .libs/vps.o .libs/wss.o .libs/xds_demux.o .libs/proxy-msg.o -Wl,--whole-archive ./.libs/libzvbiinline.a -Wl,--no-whole-archive  -lpthread -lm  -Wl,-soname -Wl,libzvbi.so.0 -o .libs/libzvbi.so.0.11.0
(cd .libs && rm -f libzvbi.so.0 && ln -s libzvbi.so.0.11.0 libzvbi.so.0)
(cd .libs && rm -f libzvbi.so && ln -s libzvbi.so.0.11.0 libzvbi.so)
rm -fr .libs/libzvbi.lax
mkdir .libs/libzvbi.lax
rm -fr .libs/libzvbi.lax/libzvbiinline.a
mkdir .libs/libzvbi.lax/libzvbiinline.a
(cd .libs/libzvbi.lax/libzvbiinline.a && ar x /tmp2/zvbi/src/./.libs/libzvbiinline.a)
libtool: link: ERROR: object name conflicts: .libs/libzvbi.lax/libzvbiinline.a//tmp2/zvbi/src/./.libs/libzvbiinline.a
make[4]: *** [libzvbi.la] Error 1
make[4]: Leaving directory `/mp/hdc4a/zvbi/src'
make[3]: *** [all-recursive] Error 1
make[3]: Leaving directory `/mp/hdc4a/zvbi/src'
make[2]: *** [all] Error 2
make[2]: Leaving directory `/mp/hdc4a/zvbi/src'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/mp/hdc4a/zvbi'
make: *** [all] Error 2


-- 
Bryan Henderson                                   San Jose, California

[Zapping-misc] compile error zapping-0.10cvs6 "simd.c:219: error: shift must be an immediate"

[Seb B. <sil...@ho...>]

Hello,

on compiling zapping-0.10cvs6 on my AMD64-machine with Fedora 8 installed, =
I get the compile-error "simd.c:219: error: shift must be an immediate".
Thought it might be of interest as I remember reading that it was considere=
d fixed in an earlier version.=20

Greetings, Seb.

P.S.: Here some of the output of make:

if /bin/sh ../libtool --tag=3DCC --mode=3Dcompile gcc -DHAVE_CONFIG_H -I. -=
I. -I.. -I..   -O3 -fomit-frame-pointer -msse -DSIMD=3DCPU_FEATURE_SSE_INT =
-g -O2 -MT libsimd_SSE_la-simd.lo -MD -MP -MF ".deps/libsimd_SSE_la-simd.Tp=
o" -c -o libsimd_SSE_la-simd.lo `test -f 'simd.c' || echo './'`simd.c; \
        then mv -f ".deps/libsimd_SSE_la-simd.Tpo" ".deps/libsimd_SSE_la-si=
md.Plo"; else rm -f ".deps/libsimd_SSE_la-simd.Tpo"; exit 1; fi
 gcc -DHAVE_CONFIG_H -I. -I. -I.. -I.. -O3 -fomit-frame-pointer -msse -DSIM=
D=3DCPU_FEATURE_SSE_INT -g -O2 -MT libsimd_SSE_la-simd.lo -MD -MP -MF .deps=
/libsimd_SSE_la-simd.Tpo -c simd.c  -fPIC -DPIC -o .libs/libsimd_SSE_la-sim=
d.o
/bin/sh ../libtool --tag=3DCC --mode=3Dlink gcc  -g -O2   -o libsimd_SSE.la=
   libsimd_SSE_la-simd.lo  -lpthread=20
ar cru .libs/libsimd_SSE.a .libs/libsimd_SSE_la-simd.o
ranlib .libs/libsimd_SSE.a
creating libsimd_SSE.la
(cd .libs && rm -f libsimd_SSE.la && ln -s ../libsimd_SSE.la libsimd_SSE.la=
)
if /bin/sh ../libtool --tag=3DCC --mode=3Dcompile gcc -DHAVE_CONFIG_H -I. -=
I. -I.. -I..   -O3 -fomit-frame-pointer -msse2 -DSIMD=3DCPU_FEATURE_SSE2 -g=
 -O2 -MT libsimd_SSE2_la-simd.lo -MD -MP -MF ".deps/libsimd_SSE2_la-simd.Tp=
o" -c -o libsimd_SSE2_la-simd.lo `test -f 'simd.c' || echo './'`simd.c; \
        then mv -f ".deps/libsimd_SSE2_la-simd.Tpo" ".deps/libsimd_SSE2_la-=
simd.Plo"; else rm -f ".deps/libsimd_SSE2_la-simd.Tpo"; exit 1; fi
 gcc -DHAVE_CONFIG_H -I. -I. -I.. -I.. -O3 -fomit-frame-pointer -msse2 -DSI=
MD=3DCPU_FEATURE_SSE2 -g -O2 -MT libsimd_SSE2_la-simd.lo -MD -MP -MF .deps/=
libsimd_SSE2_la-simd.Tpo -c simd.c  -fPIC -DPIC -o .libs/libsimd_SSE2_la-si=
md.o
simd.c: In function 'test_SSE2':
simd.c:219: error: shift must be an immediate
simd.c:220: error: shift must be an immediate
simd.c:221: error: shift must be an immediate
make[2]: *** [libsimd_SSE2_la-simd.lo] Error 1
make[2]: Leaving directory `/home/seb/Downloads/Zapping/zapping-0.10cvs6/te=
st'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/seb/Downloads/Zapping/zapping-0.10cvs6'
make: *** [all] Error 2
[seb@JB-R23 zapping-0.10cvs6]$=20
_________________________________________________________________
Make distant family not so distant with Windows Vista=AE + Windows Live=99.
http://www.microsoft.com/windows/digitallife/keepintouch.mspx?ocid=3DTXT_TA=
GLM_CPC_VideoChat_distantfamily_012008=

Re: [Zapping-misc] VBI read error

[Jonathan S. <js...@vo...>]

Kudos.  It works now.

Thanks!

Jonathan

Michael Schimek wrote:
> On Tue, 2007-09-18 at 10:08 -0500, Jonathan Smith wrote:
>   
>> I'm using zvbi-ntsc-cc to read from bt878 /dev/vbi? devices on fedora.  
>> This works sometimes.  Other times, I get the following errors:
>>     
>>> VBI read error: 5, Input/output error (ignored)
>>> VBI read error: 22, Invalid argument (ignored)
>>> VBI read error: 22, Invalid argument (ignored)
>>>       
>> It either works or errors.  If I get the errors, I restart the program 
>> until i don't get errors.
>>     
>
> I think it's fixed now. Can you confirm?
>
> cd /tmp
> export CVSROOT=:pserver:ano...@za...:/cvsroot/zapping
> cvs login # no password, just press enter
> cvs co vbi
> cd vbi
> ./autogen.sh --prefix=/usr
> make
> contrib/zvbi-ntsc-cc -c
>
> Michael
>
>
>   

Re: [Zapping-misc] significant cc data loss on one channel

[Michael S. <msc...@us...>]

For the archive, this was picture data and bursts of static interrupting
the transmission, origin unknown, occasional transmissions of null bytes
without odd parity on line 284, and CC signals with improperly low
amplitude (low signal to noise ratio). I added a new bit slicer which
seems to work better with noisy signals. Thanks to Rich for testing and
samples.

Michael

Re: [Zapping-misc] VBI read error

[Michael S. <msc...@us...>]

On Tue, 2007-09-18 at 10:08 -0500, Jonathan Smith wrote:
> I'm using zvbi-ntsc-cc to read from bt878 /dev/vbi? devices on fedora.  
> This works sometimes.  Other times, I get the following errors:
> > VBI read error: 5, Input/output error (ignored)
> > VBI read error: 22, Invalid argument (ignored)
> > VBI read error: 22, Invalid argument (ignored)
> It either works or errors.  If I get the errors, I restart the program 
> until i don't get errors.

I think it's fixed now. Can you confirm?

cd /tmp
export CVSROOT=:pserver:ano...@za...:/cvsroot/zapping
cvs login # no password, just press enter
cvs co vbi
cd vbi
./autogen.sh --prefix=/usr
make
contrib/zvbi-ntsc-cc -c

Michael

Re: [Zapping-misc] VBI read error

[Michael S. <msc...@us...>]

Hi Jonathan,

> I'm using zvbi-ntsc-cc to read from bt878 /dev/vbi? devices on fedora. 
> This works sometimes.  Other times, I get the following errors:
> > VBI read error: 5, Input/output error (ignored)
> > VBI read error: 22, Invalid argument (ignored)
> > VBI read error: 22, Invalid argument (ignored)
> It either works or errors.  If I get the errors, I restart the program 
> until i don't get errors.

The bttv driver returns EIO on hardware problems. E.g. a timeout while
waiting for the DMA to complete, a DMA fifo overflow, or the RISC
program controlling the DMA derailed. You may find a more detailed error
message in the syslog.

For apps like zvbi-ntsc-cc which just ignore errors libzvbi attempts to
restart capturing, but it seems to be doing something wrong.
Sometimes(?) the bttv driver just keeps returning EINVAL on any further
attempts to capture data. Perhaps reopening the device is really the
only way to put it back on track? I'll investigate.

> If so, how could I solve this so that the 
> program re-syncs itself?

Well if you want to dive into the ugly mess of ntsc-cc.c, the easy way
would be to call vbi_capture_delete(cap) and cap =
vbi_capture_v4l2_new(...) after vbi_capture_read() failed.

Michael

[Zapping-misc] VBI read error

[Jonathan S. <js...@vo...>]

Hi,

I'm using zvbi-ntsc-cc to read from bt878 /dev/vbi? devices on fedora.  
This works sometimes.  Other times, I get the following errors:
> VBI read error: 5, Input/output error (ignored)
> VBI read error: 22, Invalid argument (ignored)
> VBI read error: 22, Invalid argument (ignored)
It either works or errors.  If I get the errors, I restart the program 
until i don't get errors.

Is the bt878 driver messing things up, or is it a message 
synchronization problem?  If so, how could I solve this so that the 
program re-syncs itself?

Thank you,

Jonathan

Re: [Zapping-misc] Building with the dvb_mux interface

[Michael S. <msc...@us...>]

On Mon, 2007-07-30 at 21:24 +0200, Leon Woestenberg wrote:
> I have been trying to test and work with the preliminary DVB muxer 
> interface. The "test/decode.c" application uses it, but not through the 
> public header file.

test/capture.c.

> How can I export the (experimental) dvb_mux interface in my builds of 
> libzvbi and in the header file?
>
> I saw VBI_BEGIN_DECLS and VBI_END_DECLS, and  the LIBZVBI_HDRS
variable 
> in src/Makefile, and I assume I need to perform some maintainer mode 
> magic, but what?

The libzvbi.h generator in src/Makefile.am scans the files in
LIBZVBI_HEADERS and copies the lines between /* Public */ and /* Private
*/. VBI_BEGIN/END_DECLS is future stuff.

Incidentally the dvb_mux is finished now and will go into 0.2.26.

Michael

Re: [Zapping-misc] significant cc data loss on one channel

[Rich K. <ric...@ap...>]

One other thing I'm noticing.  There seems to be sensitivity tied not only
to that channel (San Diego Fox 6), but also to specific programs.  I don't
know if it's half of the programs, but certainly some programs are riddled
with parity errors and garbled captions, while other programs are fairly
clear.as good as programs on other channels.  For example, today, an episode
of Frasier gave me very good caption data, and then the next program started
"That 70's Show", and now I got over 3000 parity errors in the first 5
minutes of that show.

 

I don't know what that means, but again, it's not just the program.  It's a
combination of certain programs on that specific channel.  And if I watch
the captions on a TV, they are perfect.

 

--

 

Rich Kadel

Appeligo, Inc.

(858) 433-1747

 <http://www.appeligo.com/> www.appeligo.com

 

From: Rich Kadel [mailto:ric...@ap...] 
Sent: Wednesday, August 15, 2007 12:36 PM
To: 'zap...@li...'
Subject: significant cc data loss on one channel

 

Currently using 0.2.25 (just upgraded from minor variation on 0.2.22, but it
made no difference to upgrade)

 

I've recently noticed that I get very poor results when capturing closed
captions on at least one channel.  Michael may remember helping me with
something similar about a year ago.  I'm in San Diego, using Time Warner
Cable (NTSC).  Last year the History Channel (channel 56) had an incorrect
clock run-in (off by 1/2 cycle).  This is NOT the problem I'm experiencing
now.

 

Specifically, on channel 6 (Fox), I get numerous parity errors on the order
of 100s every minute.  Occasionally I will get a clear sentence or two, and
then captions will drop out again for a few minutes.  Using osc, the
waveform appears well defined, and appears to be timed correctly compared
with all other channels.  

 

I see nothing obviously wrong except that I have noticed occasional (a few
times a minute, to the naked eye) bursts of "noise" (significant high
frequency waves that appear for a split second and then go away).  That may
or may not cause the problem.

 

I have the same poor results in two completely different locations, so I
don't think it's my tuner card.

 

The video captured from the card looks pretty clean.

 

The captions appear fine on a standard TV, so zvbi appears to be sensitive
to something TV's are not.

 

I'm willing to try some diagnostics or experiments, if given some guidance.

 

Thanks,

Rich

 

--

 

Rich Kadel

Appeligo, Inc.

(858) 433-1747

www.appeligo.com <http://www.appeligo.com/> 

 

[Zapping-misc] significant cc data loss on one channel

[Rich K. <ric...@ap...>]

Currently using 0.2.25 (just upgraded from minor variation on 0.2.22, but it
made no difference to upgrade)

 

I've recently noticed that I get very poor results when capturing closed
captions on at least one channel.  Michael may remember helping me with
something similar about a year ago.  I'm in San Diego, using Time Warner
Cable (NTSC).  Last year the History Channel (channel 56) had an incorrect
clock run-in (off by 1/2 cycle).  This is NOT the problem I'm experiencing
now.

 

Specifically, on channel 6 (Fox), I get numerous parity errors on the order
of 100s every minute.  Occasionally I will get a clear sentence or two, and
then captions will drop out again for a few minutes.  Using osc, the
waveform appears well defined, and appears to be timed correctly compared
with all other channels.  

 

I see nothing obviously wrong except that I have noticed occasional (a few
times a minute, to the naked eye) bursts of "noise" (significant high
frequency waves that appear for a split second and then go away).  That may
or may not cause the problem.

 

I have the same poor results in two completely different locations, so I
don't think it's my tuner card.

 

The video captured from the card looks pretty clean.

 

The captions appear fine on a standard TV, so zvbi appears to be sensitive
to something TV's are not.

 

I'm willing to try some diagnostics or experiments, if given some guidance.

 

Thanks,

Rich

 

--

 

Rich Kadel

Appeligo, Inc.

(858) 433-1747

www.appeligo.com <http://www.appeligo.com/> 

 

[Zapping-misc] Building with the dvb_mux interface

[Leon W. <le...@ma...>]

Hello Michael,

I have been trying to test and work with the preliminary DVB muxer 
interface. The "test/decode.c" application uses it, but not through the 
public header file.

How can I export the (experimental) dvb_mux interface in my builds of 
libzvbi and in the header file?

I saw VBI_BEGIN_DECLS and VBI_END_DECLS, and  the LIBZVBI_HDRS variable 
in src/Makefile, and I assume I need to perform some maintainer mode 
magic, but what?

Thanks,

Leon

Re: [Zapping-misc] start line check

[Michael S. <msc...@gm...>]

On Mon, 2007-07-23 at 21:54 +0200, Leon Woestenberg wrote:
> What is the purpose of the conditions "0 != sp->start[]" in
> raw_decoder.c:

Drivers may not provide line numbers:
http://v4l2spec.bytesex.org/spec/x5490.htm

Michael

[Zapping-misc] start line check

[Leon W. <le...@ma...>]

Hello,



What is the purpose of the conditions "0 != sp->start[]" in
raw_decoder.c:

                        /* Positive match, output decoded line. */

                        sliced->id = job->id;
                        sliced->line = 0;

                        if (i >= (unsigned int) sp->count[0]) {
                                if (sp->synchronous
                                    && 0 != sp->start[1])
                                        sliced->line = sp->start[1]
                                                + i - sp->count[0];
                        } else {
                                if (sp->synchronous
                                    && 0 != sp->start[0])
                                        sliced->line = sp->start[0] + i;
                        }

Thanks, best regards,

Leon Woestenberg.

[Zapping-misc] How to make external teletext decoder using linux with bt878 ?

[Kuriawan S. <ksu...@ya...>]

Hello,

Anybody here ever try making linux box as external teletext decoder
Input from bt878 card and then decode teletext and send to rs232 serial port ?
Do you guys know howto stream /dev/vbi0 to /dev/ttyS0 ?
I think the data is packets 31 (not show on TV set)
The aplication receive the data via rs232.

regard,

Kurniawan.


       
---------------------------------
Got a little couch potato? 
Check out fun summer activities for kids.

[Zapping-misc] Compile Error: rgb2rgb.c In function '_tv_sbggr_to_rgb_MMX'

[Kropp, H. <wir...@st...>]

I experience the following error during make:
-----------------------------------------------------------------------
gcc -DHAVE_CONFIG_H -I. -I. -I.. -I.. -O3 -fomit-frame-pointer --param
max-inline-insns-single=50000 --param large-function-insns=150000
--param inline-unit-growth=30000 --param large-function-growth=300000
-mmmx -DSIMD=CPU_FEATURE_MMX -g -O2 -MT libtv_MMX_la-rgb2rgb.lo -MD -MP
-MF .deps/libtv_MMX_la-rgb2rgb.Tpo -c rgb2rgb.c  -fPIC -DPIC
-o .libs/libtv_MMX_la-rgb2rgb.o
rgb2rgb.c: In function `_tv_sbggr_to_rgb_MMX':
rgb2rgb.c:809: internal compiler error: in trunc_int_for_mode, at
explow.c:54
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
*** Error code 1

Stop in /usr/home/hkropp/Desktop/zapping-0.10cvs6/libtv.
*** Error code 1

Stop in /usr/home/hkropp/Desktop/zapping-0.10cvs6/libtv.
*** Error code 1

Stop in /usr/home/hkropp/Desktop/zapping-0.10cvs6/libtv.
*** Error code 1

Stop in /usr/home/hkropp/Desktop/zapping-0.10cvs6.
*** Error code 1

Stop in /usr/home/hkropp/Desktop/zapping-0.10cvs6.
----------------------------------------------------------------------


I am using the following compiler:
------------------------------------------------
Using built-in specs.
Configured with: FreeBSD/i386 system compiler
Thread model: posix
gcc version 3.4.6 [FreeBSD] 20060305
------------------------------------------------

Output ./configure:

        Optional packages
        -------------------------------------------------------------
        Video devices                       BKTR XVideo 
        Audio devices                       OSS 
        PAM                                 not needed
        DPMS                                yes
        DGA                                 yes
        VidMode                             yes
        XVideo                              yes
        Xinerama                            XFree86
        SIMD                                MMX SSE 
        libzvbi                             yes
        libpng                              no
        librte                              no
        liblirc_client                      not available
        libjpeg                             no

        Plugins: screenshot teletext subtitle



I fetched the latest source with CVS, configured with
'--with-gnu-ld=yes' and still having the compile error.

Any suggestions?

Thanks,
Henning Kropp

[Zapping-misc] Picture Tuning

[Gregory M. <ra...@ic...>]

Hi,
I have Zapping 0.10cvs6 on Fedora core 6 and I would just like to know 
how or if it is possible to adjust the color settings for the picture? 
Like the color saturation or contrast for example?

Regards,
Greg

[Zapping-misc] Choosing a x display

[Daniel R. <dr...@dt...>]

zapping -d is for debugging.

I want zapping --display :0.1 to make the window appear on my second
xorg screen.

Is this possible? How can this be done?  mplayer does it....

 zapping --remote :0.1 doesn't work either; and I'd rather not turn off
overlay anyway.

I suspect that I'm completely misunderstanding the way this type of
thing works, since xawtv doesn't mention the display in the man page
either...

anyone?

[Zapping-misc] configure: error: could not find DocBook XML DTD V4.1.2 in XML catalog

[Kropp, H. <wir...@st...>]

Hi!

Cant configure zapping for build process, because checking of DocBook
XML DTD fails. Although I have the latest version of DocBook installed!
How (where) can I link to the installed DocBook? Where (What) is that
"XML catalog"? Can I build zapping without Documentation?

Thanks in advanced!


#./configure --prefix=/usr/X11R6
..
checking for scrollkeeper-config... /usr/local/bin/scrollkeeper-config
checking for xsltproc... /usr/local/bin/xsltproc
checking for XML catalog (/etc/xml/catalog)... not found
checking for xmlcatalog... /usr/local/bin/xmlcatalog
checking for DocBook XML DTD V4.1.2 in XML catalog... not found
configure: error: could not find DocBook XML DTD V4.1.2 in XML catalog

# pkg_info | grep docbook
docbook-4.1_2       V4.1 of the DocBook DTD, designed for technical
documentati
docbook-4.2         V4.2 of the DocBook DTD, designed for technical
documentati
docbook-sk-4.1.2_3  XML version of the DocBook DTD version controlled
for Scrol
docbook-xml-4.2_1   XML version of the DocBook DTD
docbook-xml-4.3     DocBook/XML DTD V4.3, designed for technical
documentation
docbook-xml-4.4     DocBook/XML DTD V4.4, designed for technical
documentation
docbook-xsl-1.71.1_2 XSL DocBook stylesheets
sdocbook-xml-1.1,1  "Simplified" DocBook XML DTD

# uname -a 
FreeBSD 6.2-RELEASE FreeBSD 6.2-RELEASE #9: Wed Jan 17 23:34:39 CET 2007
i386