Mailing List archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[linux-dvb] Re: AW: Re: AW: again: IS simultaneous recording of Video + AC3 "nearly" dropoutfree possible ?



Dr. Werner Fink writes:
 > > Well that's no biggie in my opinion, i would do playback with the HW+ Card
 > > or burn it to DVD and let my Yamakawa do the playback :-)
 > > But of course it would be nice to have AC-3 playback in VDR too.
 > > When reading the opeinion that the DVB Cards are not able to playback AC3,
 > > then i sometimes ask myself if this is true, did you follow the *same*
 > > discussion about the D-Box 1 ???
 > > Now due to a New Firmware ( Dr. Overflow alias Uli ) it CAN passthrough raw
 > > SP-Dif, so it WORKS ...
 > > I wonder if the firmware guys could teach this to the DSP (?!?) in the DVB
 > > Card too ...
 > 
 > AFAIK ... after asking around ... the AV7110/AV7111 has a BUG in hardware
 > which means that is is possible but not usable.  Maybe someone knows
 > more about that?


Never heard about this "BUG".
AFAIK, the AV7110 only has a PCM mode for SP/DIF.


 
 > > The other possibilities for you would be to playback AC3 via Soundcard with
 > > SP-DIF ( Like SBLive or Terratec 512i ) or via Hollywood+ MPEG2 Decoder Card
 > > SP-DIF ...
 > 
 > Yes I want to use a Dxr3 decoder card which is similar to the Hollywood+ MPEG2
 > decoder card (-> dxr3.sourceforge.org). The only problem is: There is no code
 > usable to redirect or bypass the sound to this card or its mapped device
 > within the DVB driver.  See driver/dvb at line 5482:
 > 
 >                 case AUDIO_SET_BYPASS_MODE:
 >                         return -EINVAL;
 > 
 > ... which isn't very useful, is it?


And what sense would it make to put anything in there?
This mode is for hardware which has a DVB chip and an AC3 decoder on
the same card/component where the AC3 chip is controlled by and 
directly connected to the DVB hardware. Now guess which manufacturer 
has such hardware.
Supporting a separate AC3 card in the same driver makes no sense
at all.
 

 > > Thank you for those news, i hope you get VDR stable with the new API so we
 > > can see all your NEW ideas becaome real :-)
 > 
 > ... and to make one thing more clear, the most questions about
 > DVB from customers of SuSE is _with_ VDR _and_ CDROM writer or one
 > of the new DVD writers.  Therefore OSD should not gone away ;^)


Of course, if SuSE needs it ...



 > PS:
 > IMHO the rare SDRAM/DRAM on the DVB card shouldn't be used for things
 > which could be done in normal system RAM in a similar matter
 > (which means with the same quality).  For an example see the newest
 > CVS shot of DVB driver in driver/dvb_demux.c  If I'm correct here(?)
 > it should be possible to define NOBUFS only if USE_OSD is not
 > defined ... and the memory leak within dvb_demux.c is IMHO
 > fixable by doing the vmalloc() only once if needed and vfree()
 > the buffer if the module is unloaded or ... if this isn't usable
 > (maybe ts_feed and section_feed are used at the same time) two
 > different buffers should be used, nevertheless IMHO the allocation
 > should be done only once and the disallocation before module unloading.


How many times did I say now that saving DRAM will do us no good
at all in having more OSD?

If you really looked at the code you will also notice that the 
buffers were never used. 
They are not needed in this implementation because initial buffering 
of section and TS feeds has to be, has always been, and will be done
on the card. Another buffering layer would only mean that the data
is copied again without need.


Ralph 


-- 
Info:
To unsubscribe send a mail to listar@linuxtv.org with "unsubscribe linux-dvb" as subject.



Home | Main Index | Thread Index