MPEG stream work

MPEG stream work

[Dmitri Belimov]

[-- Attachment #1: Type: text/plain, Size: 1363 bytes --]

Hi All

Now I have MPEG stream from the saa6752hs MPEG encoder TV card of Beholder M6.

See test video 
http://debian.oshec.org/binary/tmp/mpeg01.dat

This is my script for configure TV card and read data

<script start>

echo "Set Frequency..."
./v4l2-ctl --set-freq=623.25 -d /dev/video0
echo "Set INPUT Id"
./v4l2-ctl --set-input=0 -d /dev/video0
echo "Set Norm"
./v4l2-ctl -s=secam-d -d /dev/video0
echo "Set INPUT Id"
./v4l2-ctl --set-input=0 -d /dev/video1
echo "Set Norm"
./v4l2-ctl -s=secam-d -d /dev/video1
echo "Start MPEG"
echo "Configure MPEG stream"
echo "Set Bitrate mode"
./v4l2-ctl -c video_bitrate_mode=0 -d /dev/video1
echo "Set audio sampling frequency"
./v4l2-ctl -c audio_sampling_frequency=1 -d /dev/video1
echo "Set audio encoding"
./v4l2-ctl -c audio_encoding_layer=1 -d /dev/video1
echo "Set audio bitrate"
./v4l2-ctl -c audio_layer_ii_bitrate=11 -d /dev/video1
echo "Set video bitrate"
./v4l2-ctl -c video_bitrate=7500000 -d /dev/video1
./v4l2-ctl -c video_peak_bitrate=9500000 -d /dev/video1
echo "Set aspect video"
./v4l2-ctl -c video_aspect=1 -d /dev/video1

cat /dev/video1 > test

<script stop>

But I have a trouble. I can't set correct Freq for TV tuner. I send command to tuner
but data from tuner to MPEG encoder is wrong. The encoder send to host stream with "snow window".
Anybody can help me??

With my best regards, Dmitry.

[-- Attachment #2: test.bin --]
[-- Type: application/octet-stream, Size: 512 bytes --]

[-- Attachment #3: Type: text/plain, Size: 164 bytes --]

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: MPEG stream work

[Gert Vervoort]

Dmitri Belimov wrote:
> Hi All
>
>   
Hi Dmitri
> Now I have MPEG stream from the saa6752hs MPEG encoder TV card of Beholder M6.
>
> See test video 
> http://debian.oshec.org/binary/tmp/mpeg01.dat
>
> This is my script for configure TV card and read data
>
> <script start>
>
> echo "Set Frequency..."
> ./v4l2-ctl --set-freq=623.25 -d /dev/video0
> echo "Set INPUT Id"
> ./v4l2-ctl --set-input=0 -d /dev/video0
> echo "Set Norm"
> ./v4l2-ctl -s=secam-d -d /dev/video0
> echo "Set INPUT Id"
> ./v4l2-ctl --set-input=0 -d /dev/video1
> echo "Set Norm"
> ./v4l2-ctl -s=secam-d -d /dev/video1
> echo "Start MPEG"
> echo "Configure MPEG stream"
> echo "Set Bitrate mode"
> ./v4l2-ctl -c video_bitrate_mode=0 -d /dev/video1
> echo "Set audio sampling frequency"
> ./v4l2-ctl -c audio_sampling_frequency=1 -d /dev/video1
> echo "Set audio encoding"
> ./v4l2-ctl -c audio_encoding_layer=1 -d /dev/video1
> echo "Set audio bitrate"
> ./v4l2-ctl -c audio_layer_ii_bitrate=11 -d /dev/video1
> echo "Set video bitrate"
> ./v4l2-ctl -c video_bitrate=7500000 -d /dev/video1
> ./v4l2-ctl -c video_peak_bitrate=9500000 -d /dev/video1
> echo "Set aspect video"
> ./v4l2-ctl -c video_aspect=1 -d /dev/video1
>
> cat /dev/video1 > test
>
> <script stop>
>
> But I have a trouble. I can't set correct Freq for TV tuner. I send command to tuner
> but data from tuner to MPEG encoder is wrong. The encoder send to host stream with "snow window".
> Anybody can help me??
>
>   
When video on the input of the SAA6752 is not valid video (eg a tuner 
which is not properly tuned), then usually it will not work properly (at 
least it will not contain compressed video in the output stream). 
However the "snow window" you see seems to be correct video in eyes of 
the SAA6752 (contains correct syncs), see it has encoded it. The TS 
stream is valid, I can play it back (with xine and ffplay) and see the 
"snow window". The encoder seem to have done a proper job.

Do you also see this "snow window" on the video device of the SAA7143 
(eg xawtv -c /dev/video0)?
That should show the same video signal (but scaled), as is send to the 
SAA6752.
The tuner (CVBS) signal is send to the SAA7134, which will do the analog 
video decoding / analog to digital conversion, the resulting 656 digital 
video signal goes both to the scaler/DMA (visible on /dev/video0) and in 
parallel it also goes to the ITU656 output to which to SAA6752 is connected.

If your TV-card has a CVBS input, you could try what happens if you 
connect an other video source (eg DVD, VCR).

  Gert





--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: MPEG stream work

[hermann pitton]

Am Freitag, den 15.08.2008, 11:59 +1000 schrieb Dmitri Belimov:
> Hi All
> 
> I found problem in v4l2-ctl. This programm can't set correct TV norm. After my hack TV norm was set correct.
> 
> diff -r 42e3970c09aa v4l2-apps/util/v4l2-ctl.cpp
> --- a/v4l2-apps/util/v4l2-ctl.cpp	Sun Jul 27 19:30:46 2008 -0300
> +++ b/v4l2-apps/util/v4l2-ctl.cpp	Fri Aug 15 05:53:38 2008 +1000
> @@ -1572,6 +1572,7 @@
>  	}
>  
>  	if (options[OptSetStandard]) {
> +	  std = 0x320000; // durty hack for SECAM-DK
>  		if (std & (1ULL << 63)) {
>  			vs.index = std & 0xffff;
>  			if (ioctl(fd, VIDIOC_ENUMSTD, &vs) >= 0) {
> 
> I have MPEG stream with CORRECT TV data.
> See link:
> 
> http://debian.oshec.org/binary/tmp/mpeg02.dat
> 
> Yahooooo!
> 
> With my best regards, Dmitry.

Hi Dimitry,

looks fine!

Thanks for all your efforts and excuse to have been thrown in such a
real cruel mess, totally undocumented..

Thanks again, for not giving up!

Cheers,
Hermann




--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: MPEG stream work

[Hans Verkuil]

On Friday 15 August 2008 03:59:54 Dmitri Belimov wrote:
> Hi All
>
> I found problem in v4l2-ctl. This programm can't set correct TV norm.
> After my hack TV norm was set correct.

???

$ v4l2-ctl -s secam-dk
Standard set to 00320000

v4l2-ctl works fine for me!

Are you using the latest v4l2-ctl version from v4l-dvb? I did fix a bug 
in SetStandard some time ago (although I think that bug didn't affect 
this particular situation either).

Regards,

	Hans

>
> diff -r 42e3970c09aa v4l2-apps/util/v4l2-ctl.cpp
> --- a/v4l2-apps/util/v4l2-ctl.cpp	Sun Jul 27 19:30:46 2008 -0300
> +++ b/v4l2-apps/util/v4l2-ctl.cpp	Fri Aug 15 05:53:38 2008 +1000
> @@ -1572,6 +1572,7 @@
>  	}
>
>  	if (options[OptSetStandard]) {
> +	  std = 0x320000; // durty hack for SECAM-DK
>  		if (std & (1ULL << 63)) {
>  			vs.index = std & 0xffff;
>  			if (ioctl(fd, VIDIOC_ENUMSTD, &vs) >= 0) {
>
> I have MPEG stream with CORRECT TV data.
> See link:
>
> http://debian.oshec.org/binary/tmp/mpeg02.dat
>
> Yahooooo!
>
> With my best regards, Dmitry.
>
> --
> video4linux-list mailing list
> Unsubscribe
> mailto:video4linux-list-request@redhat.com?subject=unsubscribe
> https://www.redhat.com/mailman/listinfo/video4linux-list


--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: MPEG stream work

[hermann pitton]

Hans,

Am Freitag, den 15.08.2008, 08:05 +0200 schrieb Hans Verkuil:
> On Friday 15 August 2008 03:59:54 Dmitri Belimov wrote:
> > Hi All
> >
> > I found problem in v4l2-ctl. This programm can't set correct TV norm.
> > After my hack TV norm was set correct.

I'm the first time grumpy here.

> ???
> 
> $ v4l2-ctl -s secam-dk
> Standard set to 00320000
> 
> v4l2-ctl works fine for me!
> 
> Are you using the latest v4l2-ctl version from v4l-dvb? I did fix a bug 
> in SetStandard some time ago (although I think that bug didn't affect 
> this particular situation either).

We don't make jokes, but seriously report what we see.

For Dmitry this is a since months during Odyssey, only caused by us, but
he is expected to work for what he is paid for. An impossible task in
the beginning.

I did spent time too on it, the broken empress encoder first happens,
then vivid sub-norm setting on 2.6.25 was not possible anymore after
ioctl2 conversion on the saa7134, Mauro has hardware from me, and the
next was on 2.6.26 not to be able to set the tuner anymore, another
serious issue, after having a hot fix for the tuner eeprom detection at
least.

Always pointed to it from my side, as far I have hardware, but it made
it straight into released kernels and I'm tired to have to rumble there,
since previously all fuses ready, did not trigger.

Now, after facing such stuff around over months, it also makes it much
more difficult to settle the state of new cards around, not to talk
about ubuntu pleasures, you drop me from the list you might want to
reply in this case, after previously I did include you as maybe
interested?

Why?

Hermann








> Regards,
> 
> 	Hans
> 
> >
> > diff -r 42e3970c09aa v4l2-apps/util/v4l2-ctl.cpp
> > --- a/v4l2-apps/util/v4l2-ctl.cpp	Sun Jul 27 19:30:46 2008 -0300
> > +++ b/v4l2-apps/util/v4l2-ctl.cpp	Fri Aug 15 05:53:38 2008 +1000
> > @@ -1572,6 +1572,7 @@
> >  	}
> >
> >  	if (options[OptSetStandard]) {
> > +	  std = 0x320000; // durty hack for SECAM-DK
> >  		if (std & (1ULL << 63)) {
> >  			vs.index = std & 0xffff;
> >  			if (ioctl(fd, VIDIOC_ENUMSTD, &vs) >= 0) {
> >
> > I have MPEG stream with CORRECT TV data.
> > See link:
> >
> > http://debian.oshec.org/binary/tmp/mpeg02.dat
> >
> > Yahooooo!
> >
> > With my best regards, Dmitry.
> >


--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: MPEG stream work

[Hans Verkuil]

Hermann,

I did not remove you from the CC list. You were not on it when I 
received the message. I think it is because of the way Mauro's address 
is being formatted by Dmitri's mailer.

Dmitri, I think you have Mauro's mail address formatted something like:

Chehab <mchehab@infradead.org>, Mauro

When that is send out some ISPs get confused and it ends up as e.g.:

Chehab <mchehab@infradead.org>, Mauro@xs4all.nl

In my case my xs4all ISP thinks 'Mauro' refers to a local user and adds 
the local domain to it. I also suspect it probably strips off one or 
more additional CC addresses, which is why Hermann's email disappeared 
in the mail that I received.

The correct format is: Mauro Carvalho Chehab <mchehab@infradead.org> (no 
comma in between). Can you change it to prevent this unfortunate event 
from happening?

Hermann, if I ever gave the impression that I thought this was a joke, 
then I apologize profusely since I know very well that Dmitri has a 
difficult task to take on. Having a driver in such a poor state is 
definitely no joking matter. I am also helping out Dmitri as well as I 
can, but sadly I do not have as much time as I would like to have due 
to a range of other commitments.

Regards,

	Hans

On Saturday 16 August 2008 02:44:03 hermann pitton wrote:
> Hans,
>
> Am Freitag, den 15.08.2008, 08:05 +0200 schrieb Hans Verkuil:
> > On Friday 15 August 2008 03:59:54 Dmitri Belimov wrote:
> > > Hi All
> > >
> > > I found problem in v4l2-ctl. This programm can't set correct TV
> > > norm. After my hack TV norm was set correct.
>
> I'm the first time grumpy here.
>
> > ???
> >
> > $ v4l2-ctl -s secam-dk
> > Standard set to 00320000
> >
> > v4l2-ctl works fine for me!
> >
> > Are you using the latest v4l2-ctl version from v4l-dvb? I did fix a
> > bug in SetStandard some time ago (although I think that bug didn't
> > affect this particular situation either).
>
> We don't make jokes, but seriously report what we see.
>
> For Dmitry this is a since months during Odyssey, only caused by us,
> but he is expected to work for what he is paid for. An impossible
> task in the beginning.
>
> I did spent time too on it, the broken empress encoder first happens,
> then vivid sub-norm setting on 2.6.25 was not possible anymore after
> ioctl2 conversion on the saa7134, Mauro has hardware from me, and the
> next was on 2.6.26 not to be able to set the tuner anymore, another
> serious issue, after having a hot fix for the tuner eeprom detection
> at least.
>
> Always pointed to it from my side, as far I have hardware, but it
> made it straight into released kernels and I'm tired to have to
> rumble there, since previously all fuses ready, did not trigger.
>
> Now, after facing such stuff around over months, it also makes it
> much more difficult to settle the state of new cards around, not to
> talk about ubuntu pleasures, you drop me from the list you might want
> to reply in this case, after previously I did include you as maybe
> interested?
>
> Why?
>
> Hermann
>
> > Regards,
> >
> > 	Hans
> >
> > > diff -r 42e3970c09aa v4l2-apps/util/v4l2-ctl.cpp
> > > --- a/v4l2-apps/util/v4l2-ctl.cpp	Sun Jul 27 19:30:46 2008 -0300
> > > +++ b/v4l2-apps/util/v4l2-ctl.cpp	Fri Aug 15 05:53:38 2008 +1000
> > > @@ -1572,6 +1572,7 @@
> > >  	}
> > >
> > >  	if (options[OptSetStandard]) {
> > > +	  std = 0x320000; // durty hack for SECAM-DK
> > >  		if (std & (1ULL << 63)) {
> > >  			vs.index = std & 0xffff;
> > >  			if (ioctl(fd, VIDIOC_ENUMSTD, &vs) >= 0) {
> > >
> > > I have MPEG stream with CORRECT TV data.
> > > See link:
> > >
> > > http://debian.oshec.org/binary/tmp/mpeg02.dat
> > >
> > > Yahooooo!
> > >
> > > With my best regards, Dmitry.


--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: MPEG stream work

[hermann pitton]

Hi Hans,

Am Samstag, den 16.08.2008, 11:37 +0200 schrieb Hans Verkuil:
> Hermann,
> 
> I did not remove you from the CC list. You were not on it when I 
> received the message. I think it is because of the way Mauro's address 
> is being formatted by Dmitri's mailer.
> 
> Dmitri, I think you have Mauro's mail address formatted something like:
> 
> Chehab <mchehab@infradead.org>, Mauro
> 
> When that is send out some ISPs get confused and it ends up as e.g.:
> 
> Chehab <mchehab@infradead.org>, Mauro@xs4all.nl
> 
> In my case my xs4all ISP thinks 'Mauro' refers to a local user and adds 
> the local domain to it. I also suspect it probably strips off one or 
> more additional CC addresses, which is why Hermann's email disappeared 
> in the mail that I received.
> 
> The correct format is: Mauro Carvalho Chehab <mchehab@infradead.org> (no 
> comma in between). Can you change it to prevent this unfortunate event 
> from happening?
> 
> Hermann, if I ever gave the impression that I thought this was a joke, 
> then I apologize profusely since I know very well that Dmitri has a 
> difficult task to take on. Having a driver in such a poor state is 
> definitely no joking matter. I am also helping out Dmitri as well as I 
> can, but sadly I do not have as much time as I would like to have due 
> to a range of other commitments.
> 
> Regards,
> 
> 	Hans

Sorry! I obviously got it wrong and better should have slept a night
over it.

I previously tried to have your opinion about the FQ1236 MK4 and why we
don't have the tda9887 for it in tuner-types.c and did _not_ post
directly to you as well. You likely never saw it. It was added for ivtv
once.

Trying to make it better this time, but being stripped from the
responses, intrigued me into you might have enough with other stuff and
prefer to live without more reports.

This would be your free choice as well, but luckily even the opposite is
the case.

I have two of my older machines with enough PCI slots up and running
again, but I'm not through with all tests yet and also face some other
minor oddities with them. (Fans and need to crimp some new network
cables)

For sure is for now, that on a 2.6.18 a TCL2002MB-1F on my old
FlyVideo3000, card=2 has no problems with changing freq. and standard
with Gerd's old v4lctl during xawtv/tvtime/mplayer are running.

The even better is, v4l2-ctl compiled on so far latest on a other x86
machine, has no problems too and is even fully backward compatible.

Why the FMD1216ME hybrid MK3 immediately loses picture and sound after
an ioctl is touched during a running capture, not using any empress for
now, and if it is the same for the FM1216ME/I MK3s and when it starts, I
don't know yet.

Since Dmitry has a solution now and is not blocked by utilities, sorry
for the noise again.

I report back if I have more.

Cheers,
Hermann

> On Saturday 16 August 2008 02:44:03 hermann pitton wrote:
> > Hans,
> >
> > Am Freitag, den 15.08.2008, 08:05 +0200 schrieb Hans Verkuil:
> > > On Friday 15 August 2008 03:59:54 Dmitri Belimov wrote:
> > > > Hi All
> > > >
> > > > I found problem in v4l2-ctl. This programm can't set correct TV
> > > > norm. After my hack TV norm was set correct.
> >
> > I'm the first time grumpy here.
> >
> > > ???
> > >
> > > $ v4l2-ctl -s secam-dk
> > > Standard set to 00320000
> > >
> > > v4l2-ctl works fine for me!
> > >
> > > Are you using the latest v4l2-ctl version from v4l-dvb? I did fix a
> > > bug in SetStandard some time ago (although I think that bug didn't
> > > affect this particular situation either).
> >
> > We don't make jokes, but seriously report what we see.
> >
> > For Dmitry this is a since months during Odyssey, only caused by us,
> > but he is expected to work for what he is paid for. An impossible
> > task in the beginning.
> >
> > I did spent time too on it, the broken empress encoder first happens,
> > then vivid sub-norm setting on 2.6.25 was not possible anymore after
> > ioctl2 conversion on the saa7134, Mauro has hardware from me, and the
> > next was on 2.6.26 not to be able to set the tuner anymore, another
> > serious issue, after having a hot fix for the tuner eeprom detection
> > at least.
> >
> > Always pointed to it from my side, as far I have hardware, but it
> > made it straight into released kernels and I'm tired to have to
> > rumble there, since previously all fuses ready, did not trigger.
> >
> > Now, after facing such stuff around over months, it also makes it
> > much more difficult to settle the state of new cards around, not to
> > talk about ubuntu pleasures, you drop me from the list you might want
> > to reply in this case, after previously I did include you as maybe
> > interested?
> >
> > Why?
> >
> > Hermann
> >
> > > Regards,
> > >
> > > 	Hans
> > >
> > > > diff -r 42e3970c09aa v4l2-apps/util/v4l2-ctl.cpp
> > > > --- a/v4l2-apps/util/v4l2-ctl.cpp	Sun Jul 27 19:30:46 2008 -0300
> > > > +++ b/v4l2-apps/util/v4l2-ctl.cpp	Fri Aug 15 05:53:38 2008 +1000
> > > > @@ -1572,6 +1572,7 @@
> > > >  	}
> > > >
> > > >  	if (options[OptSetStandard]) {
> > > > +	  std = 0x320000; // durty hack for SECAM-DK
> > > >  		if (std & (1ULL << 63)) {
> > > >  			vs.index = std & 0xffff;
> > > >  			if (ioctl(fd, VIDIOC_ENUMSTD, &vs) >= 0) {
> > > >
> > > > I have MPEG stream with CORRECT TV data.
> > > > See link:
> > > >
> > > > http://debian.oshec.org/binary/tmp/mpeg02.dat
> > > >
> > > > Yahooooo!
> > > >
> > > > With my best regards, Dmitry.
> 
> 

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: MPEG stream work

[Dmitri Belimov]

Hi Hans

> > I found problem in v4l2-ctl. This programm can't set correct TV
> > norm. After my hack TV norm was set correct.
> 
> ???
> 
> $ v4l2-ctl -s secam-dk
> Standard set to 00320000
> 
> v4l2-ctl works fine for me!
> 
> Are you using the latest v4l2-ctl version from v4l-dvb? I did fix a
> bug in SetStandard some time ago (although I think that bug didn't
> affect this particular situation either).

This is my error. I run

v4l2-ctl -s=secam-dk -d /dev/video0

It response set TV norm to 0x0b000

When I run

v4l2-ctl -s secam-dk -d /dev/video0

All is OK.

With my best regards, Dmitry.
 
> Regards,
> 
> 	Hans
> 
> >
> > diff -r 42e3970c09aa v4l2-apps/util/v4l2-ctl.cpp
> > --- a/v4l2-apps/util/v4l2-ctl.cpp	Sun Jul 27 19:30:46 2008
> > -0300 +++ b/v4l2-apps/util/v4l2-ctl.cpp	Fri Aug 15 05:53:38
> > 2008 +1000 @@ -1572,6 +1572,7 @@
> >  	}
> >
> >  	if (options[OptSetStandard]) {
> > +	  std = 0x320000; // durty hack for SECAM-DK
> >  		if (std & (1ULL << 63)) {
> >  			vs.index = std & 0xffff;
> >  			if (ioctl(fd, VIDIOC_ENUMSTD, &vs) >= 0) {
> >
> > I have MPEG stream with CORRECT TV data.
> > See link:
> >
> > http://debian.oshec.org/binary/tmp/mpeg02.dat
> >
> > Yahooooo!
> >
> > With my best regards, Dmitry.
> >
> > --
> > video4linux-list mailing list
> > Unsubscribe
> > mailto:video4linux-list-request@redhat.com?subject=unsubscribe
> > https://www.redhat.com/mailman/listinfo/video4linux-list
> 
> 

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: MPEG stream work

[hermann pitton]

Hi,

Am Samstag, den 16.08.2008, 13:47 +1000 schrieb Dmitri Belimov:
> Hi Hans
> 
> > > I found problem in v4l2-ctl. This programm can't set correct TV
> > > norm. After my hack TV norm was set correct.
> > 
> > ???
> > 
> > $ v4l2-ctl -s secam-dk
> > Standard set to 00320000
> > 
> > v4l2-ctl works fine for me!
> > 
> > Are you using the latest v4l2-ctl version from v4l-dvb? I did fix a
> > bug in SetStandard some time ago (although I think that bug didn't
> > affect this particular situation either).
> 
> This is my error. I run
> 
> v4l2-ctl -s=secam-dk -d /dev/video0
> 
> It response set TV norm to 0x0b000
> 
> When I run
> 
> v4l2-ctl -s secam-dk -d /dev/video0
> 
> All is OK.
> 
> With my best regards, Dmitry.

yes, sorry.

This does explain at least the NTSC-M seen after it and lowers confusion
on this, after all, amazing and successful story.

However, using v4l2-ctl for setting the correct norm still results in
losing picture and sound immediately on all FMD1216ME/I MK3 and all
tda8275a tuners in this machine here.

The same happens with old "v4lctl setnorm PAL-BG".
The tda8275a is known for not playing nicely with all apps, the FMD1216
is famous for other issues, but not those.

Eventually I might have a chance in the evening to play with some very
old tuners to look out for any difference, for now I still say, I can't
set the norm neither with v4l2-ctl nor v4lctl on all four cards I have
here and this was my latest contribution to help to debug it.

Cheers,
Hermann

>  
> > Regards,
> > 
> > 	Hans
> > 
> > >
> > > diff -r 42e3970c09aa v4l2-apps/util/v4l2-ctl.cpp
> > > --- a/v4l2-apps/util/v4l2-ctl.cpp	Sun Jul 27 19:30:46 2008
> > > -0300 +++ b/v4l2-apps/util/v4l2-ctl.cpp	Fri Aug 15 05:53:38
> > > 2008 +1000 @@ -1572,6 +1572,7 @@
> > >  	}
> > >
> > >  	if (options[OptSetStandard]) {
> > > +	  std = 0x320000; // durty hack for SECAM-DK
> > >  		if (std & (1ULL << 63)) {
> > >  			vs.index = std & 0xffff;
> > >  			if (ioctl(fd, VIDIOC_ENUMSTD, &vs) >= 0) {
> > >
> > > I have MPEG stream with CORRECT TV data.
> > > See link:
> > >
> > > http://debian.oshec.org/binary/tmp/mpeg02.dat
> > >
> > > Yahooooo!
> > >
> > > With my best regards, Dmitry.
> > >


--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: MPEG stream work

[Hans Verkuil]

Hi Hermann,

On Saturday 16 August 2008 10:30:12 hermann pitton wrote:
> Hi,
>
> Am Samstag, den 16.08.2008, 13:47 +1000 schrieb Dmitri Belimov:
> > Hi Hans
> >
> > > > I found problem in v4l2-ctl. This programm can't set correct TV
> > > > norm. After my hack TV norm was set correct.
> > >
> > > ???
> > >
> > > $ v4l2-ctl -s secam-dk
> > > Standard set to 00320000
> > >
> > > v4l2-ctl works fine for me!
> > >
> > > Are you using the latest v4l2-ctl version from v4l-dvb? I did fix
> > > a bug in SetStandard some time ago (although I think that bug
> > > didn't affect this particular situation either).
> >
> > This is my error. I run
> >
> > v4l2-ctl -s=secam-dk -d /dev/video0
> >
> > It response set TV norm to 0x0b000
> >
> > When I run
> >
> > v4l2-ctl -s secam-dk -d /dev/video0
> >
> > All is OK.
> >
> > With my best regards, Dmitry.
>
> yes, sorry.
>
> This does explain at least the NTSC-M seen after it and lowers
> confusion on this, after all, amazing and successful story.
>
> However, using v4l2-ctl for setting the correct norm still results in
> losing picture and sound immediately on all FMD1216ME/I MK3 and all
> tda8275a tuners in this machine here.

Are you changing the norm while a capture is in progress, or before 
starting a capture? And if you capture, do you capture from the MPEG 
device or from the framegrabber?

At least for the ivtv/cx18 driver changing the norm in midcapture will 
not work at all since the cx2341x chip usually just loses sync with the 
video encoder chip. So trying to change the norm while a capture is in 
progress will result in an EBUSY error for ivtv and cx18.

Of course, if this worked fine in the past for the saa7134 driver, then 
clearly something broke.

Regards,

	Hans

>
> The same happens with old "v4lctl setnorm PAL-BG".
> The tda8275a is known for not playing nicely with all apps, the
> FMD1216 is famous for other issues, but not those.
>
> Eventually I might have a chance in the evening to play with some
> very old tuners to look out for any difference, for now I still say,
> I can't set the norm neither with v4l2-ctl nor v4lctl on all four
> cards I have here and this was my latest contribution to help to
> debug it.
>
> Cheers,
> Hermann
>
> > > Regards,
> > >
> > > 	Hans
> > >
> > > > diff -r 42e3970c09aa v4l2-apps/util/v4l2-ctl.cpp
> > > > --- a/v4l2-apps/util/v4l2-ctl.cpp	Sun Jul 27 19:30:46 2008
> > > > -0300 +++ b/v4l2-apps/util/v4l2-ctl.cpp	Fri Aug 15 05:53:38
> > > > 2008 +1000 @@ -1572,6 +1572,7 @@
> > > >  	}
> > > >
> > > >  	if (options[OptSetStandard]) {
> > > > +	  std = 0x320000; // durty hack for SECAM-DK
> > > >  		if (std & (1ULL << 63)) {
> > > >  			vs.index = std & 0xffff;
> > > >  			if (ioctl(fd, VIDIOC_ENUMSTD, &vs) >= 0) {
> > > >
> > > > I have MPEG stream with CORRECT TV data.
> > > > See link:
> > > >
> > > > http://debian.oshec.org/binary/tmp/mpeg02.dat
> > > >
> > > > Yahooooo!
> > > >
> > > > With my best regards, Dmitry.
>
> --
> video4linux-list mailing list
> Unsubscribe
> mailto:video4linux-list-request@redhat.com?subject=unsubscribe
> https://www.redhat.com/mailman/listinfo/video4linux-list


--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: MPEG stream work

[Hans Verkuil]

On Saturday 16 August 2008 05:47:20 Dmitri Belimov wrote:
> Hi Hans
>
> > > I found problem in v4l2-ctl. This programm can't set correct TV
> > > norm. After my hack TV norm was set correct.
> >
> > ???
> >
> > $ v4l2-ctl -s secam-dk
> > Standard set to 00320000
> >
> > v4l2-ctl works fine for me!
> >
> > Are you using the latest v4l2-ctl version from v4l-dvb? I did fix a
> > bug in SetStandard some time ago (although I think that bug didn't
> > affect this particular situation either).
>
> This is my error. I run
>
> v4l2-ctl -s=secam-dk -d /dev/video0
>
> It response set TV norm to 0x0b000
>
> When I run
>
> v4l2-ctl -s secam-dk -d /dev/video0
>
> All is OK.

Ah, thank you. I was getting worried that there was another bug in 
v4l2-ctl.

Regards,

	Hans

>
> With my best regards, Dmitry.
>
> > Regards,
> >
> > 	Hans
> >
> > > diff -r 42e3970c09aa v4l2-apps/util/v4l2-ctl.cpp
> > > --- a/v4l2-apps/util/v4l2-ctl.cpp	Sun Jul 27 19:30:46 2008
> > > -0300 +++ b/v4l2-apps/util/v4l2-ctl.cpp	Fri Aug 15 05:53:38
> > > 2008 +1000 @@ -1572,6 +1572,7 @@
> > >  	}
> > >
> > >  	if (options[OptSetStandard]) {
> > > +	  std = 0x320000; // durty hack for SECAM-DK
> > >  		if (std & (1ULL << 63)) {
> > >  			vs.index = std & 0xffff;
> > >  			if (ioctl(fd, VIDIOC_ENUMSTD, &vs) >= 0) {
> > >
> > > I have MPEG stream with CORRECT TV data.
> > > See link:
> > >
> > > http://debian.oshec.org/binary/tmp/mpeg02.dat
> > >
> > > Yahooooo!
> > >
> > > With my best regards, Dmitry.
> > >
> > > --
> > > video4linux-list mailing list
> > > Unsubscribe
> > > mailto:video4linux-list-request@redhat.com?subject=unsubscribe
> > > https://www.redhat.com/mailman/listinfo/video4linux-list
>
> --
> video4linux-list mailing list
> Unsubscribe
> mailto:video4linux-list-request@redhat.com?subject=unsubscribe
> https://www.redhat.com/mailman/listinfo/video4linux-list


--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: MPEG stream work

[hermann pitton]

Hi,

Am Donnerstag, den 14.08.2008, 09:33 +1000 schrieb Dmitri Belimov: 
> Hi All
> 
> Now I have MPEG stream from the saa6752hs MPEG encoder TV card of Beholder M6.
> 
> See test video 
> http://debian.oshec.org/binary/tmp/mpeg01.dat
> 
> This is my script for configure TV card and read data
> 
> <script start>
> 
> echo "Set Frequency..."
> ./v4l2-ctl --set-freq=623.25 -d /dev/video0
> echo "Set INPUT Id"
> ./v4l2-ctl --set-input=0 -d /dev/video0
> echo "Set Norm"
> ./v4l2-ctl -s=secam-d -d /dev/video0
> echo "Set INPUT Id"
> ./v4l2-ctl --set-input=0 -d /dev/video1
> echo "Set Norm"
> ./v4l2-ctl -s=secam-d -d /dev/video1
> echo "Start MPEG"
> echo "Configure MPEG stream"
> echo "Set Bitrate mode"
> ./v4l2-ctl -c video_bitrate_mode=0 -d /dev/video1
> echo "Set audio sampling frequency"
> ./v4l2-ctl -c audio_sampling_frequency=1 -d /dev/video1
> echo "Set audio encoding"
> ./v4l2-ctl -c audio_encoding_layer=1 -d /dev/video1
> echo "Set audio bitrate"
> ./v4l2-ctl -c audio_layer_ii_bitrate=11 -d /dev/video1
> echo "Set video bitrate"
> ./v4l2-ctl -c video_bitrate=7500000 -d /dev/video1
> ./v4l2-ctl -c video_peak_bitrate=9500000 -d /dev/video1
> echo "Set aspect video"
> ./v4l2-ctl -c video_aspect=1 -d /dev/video1
> 
> cat /dev/video1 > test
> 
> <script stop>
> 
> But I have a trouble. I can't set correct Freq for TV tuner. I send command to tuner
> but data from tuner to MPEG encoder is wrong. The encoder send to host stream with "snow window".
> Anybody can help me??
> 

hmm, looks like black snow. The tda9887 on the FM1216ME/I H-3 goes off
if the tuner is not used by an application.

Enable saa7134-empress debug and tda9887 debug=2.

I have no analog audio out on that card, so I use saa7134-alsa and sox. 
sox -c 2 -s -w -r 32000 -t ossdsp /dev/dsp4 -t ossdsp -w -r 48000 /dev/dsp

With Hans' "qv4l2" from v4l2-apps/util on /dev/video3, your video0, I
have sound, can switch channels and standards.

If I force SECAM-DK i get of course distorted sound and the empress
loses the video signal, but this is all like expected on PAL-BG.

You can also enable saa7134 audio_debug=1, but you should have video and
sound on /dev/video0 ? I let "tvtime" run on it.

saa7134[3]/empress: video signal acquired
tda9887 5-0043: configure for: SECAM-DK
tda9887 5-0043: writing: b=0x14 c=0x70 e=0x4b
tda9887 5-0043: write: byte B 0x14
tda9887 5-0043:   B0   video mode      : sound trap
tda9887 5-0043:   B1   auto mute fm    : no
tda9887 5-0043:   B2   carrier mode    : QSS
tda9887 5-0043:   B3-4 tv sound/radio  : FM/TV
tda9887 5-0043:   B5   force mute audio: no
tda9887 5-0043:   B6   output port 1   : low (active)
tda9887 5-0043:   B7   output port 2   : low (active)
tda9887 5-0043: write: byte C 0x70
tda9887 5-0043:   C0-4 top adjustment  : 0 dB
tda9887 5-0043:   C5-6 de-emphasis     : 50
tda9887 5-0043:   C7   audio gain      : 0
tda9887 5-0043: write: byte E 0x4b
tda9887 5-0043:   E0-1 sound carrier   : 6.5 MHz / AM
tda9887 5-0043:   E6   l pll gating   : 36
tda9887 5-0043:   E2-4 video if        : 38.9 MHz
tda9887 5-0043:   E5   tuner gain      : normal
tda9887 5-0043:   E7   vif agc output  : pin3+pin22 port
tda9887 5-0043: --
saa7134[3]/empress: no video signal
tda9887 5-0043: configure for: PAL-BGHN
tda9887 5-0043: writing: b=0x14 c=0x70 e=0x49
tda9887 5-0043: write: byte B 0x14
tda9887 5-0043:   B0   video mode      : sound trap
tda9887 5-0043:   B1   auto mute fm    : no
tda9887 5-0043:   B2   carrier mode    : QSS
tda9887 5-0043:   B3-4 tv sound/radio  : FM/TV
tda9887 5-0043:   B5   force mute audio: no
tda9887 5-0043:   B6   output port 1   : low (active)
tda9887 5-0043:   B7   output port 2   : low (active)
tda9887 5-0043: write: byte C 0x70
tda9887 5-0043:   C0-4 top adjustment  : 0 dB
tda9887 5-0043:   C5-6 de-emphasis     : 50
tda9887 5-0043:   C7   audio gain      : 0
tda9887 5-0043: write: byte E 0x49
tda9887 5-0043:   E0-1 sound carrier   : 5.5 MHz
tda9887 5-0043:   E6   l pll gating   : 36
tda9887 5-0043:   E2-4 video if        : 38.9 MHz
tda9887 5-0043:   E5   tuner gain      : normal
tda9887 5-0043:   E7   vif agc output  : pin3+pin22 port
tda9887 5-0043: --
tda9887 5-0043: configure for: PAL-BGHN
tda9887 5-0043: writing: b=0x14 c=0x70 e=0x49
tda9887 5-0043: write: byte B 0x14
tda9887 5-0043:   B0   video mode      : sound trap
tda9887 5-0043:   B1   auto mute fm    : no
tda9887 5-0043:   B2   carrier mode    : QSS
tda9887 5-0043:   B3-4 tv sound/radio  : FM/TV
tda9887 5-0043:   B5   force mute audio: no
tda9887 5-0043:   B6   output port 1   : low (active)
tda9887 5-0043:   B7   output port 2   : low (active)
tda9887 5-0043: write: byte C 0x70
tda9887 5-0043:   C0-4 top adjustment  : 0 dB
tda9887 5-0043:   C5-6 de-emphasis     : 50
tda9887 5-0043:   C7   audio gain      : 0
tda9887 5-0043: write: byte E 0x49
tda9887 5-0043:   E0-1 sound carrier   : 5.5 MHz
tda9887 5-0043:   E6   l pll gating   : 36
tda9887 5-0043:   E2-4 video if        : 38.9 MHz
tda9887 5-0043:   E5   tuner gain      : normal
tda9887 5-0043:   E7   vif agc output  : pin3+pin22 port
tda9887 5-0043: --
saa7134[3]/empress: video signal acquired

Only if I open /dev/video4 empress with qv4l2 too, then I can't change
the frequency anymore, changing standard still works on /dev/video3 and
both never worked on video4, since unsupported input "Television" on
CCIR   comes back.

I also notice that mute doesn't work anymore with qv4l2 on both empress
devices. It worked previously even from /dev/video4 like the other
controls still do. Without empress enabled mute is OK.

BUT, using v4l2-ctl currently causes that any tuned in channel is
immediately lost! Noise for video and audio.

And the tda9887 always shows

tda9887 5-0043: configure for: NTSC-M
tda9887 5-0043: writing: b=0x34 c=0x30 e=0x44
tda9887 5-0043: write: byte B 0x34
tda9887 5-0043:   B0   video mode      : sound trap
tda9887 5-0043:   B1   auto mute fm    : no
tda9887 5-0043:   B2   carrier mode    : QSS
tda9887 5-0043:   B3-4 tv sound/radio  : FM/TV
tda9887 5-0043:   B5   force mute audio: yes
tda9887 5-0043:   B6   output port 1   : low (active)
tda9887 5-0043:   B7   output port 2   : low (active)
tda9887 5-0043: write: byte C 0x30
tda9887 5-0043:   C0-4 top adjustment  : 0 dB
tda9887 5-0043:   C5-6 de-emphasis     : no
tda9887 5-0043:   C7   audio gain      : 0
tda9887 5-0043: write: byte E 0x44
tda9887 5-0043:   E0-1 sound carrier   : 4.5 MHz
tda9887 5-0043:   E6   l pll gating   : 36
tda9887 5-0043:   E2-4 video if        : 45.75 MHz
tda9887 5-0043:   E5   tuner gain      : normal
tda9887 5-0043:   E7   vif agc output  : pin3+pin22 port
tda9887 5-0043: --

This seems to be a bug.

If you see the same, try to use qv4l2.

hg head
changeset:   8647:405d3e926ffd
tag:         tip
parent:      8641:deeb5dfb37a1
parent:      8646:2bade2ed7ac8
user:        Mauro Carvalho Chehab <mchehab@infradead.org>
date:        Sat Aug 09 09:21:15 2008 -0300
summary:     merge: http://www.linuxtv.org/hg/~stoth/v4l-dvb

Some 2.6.24 is booted and Hans' latest improvements are in.

Since I can't switch yet from DVB-T to the encoder, likely can't test
much more.

Cheers,
Hermann









--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list