On Mon, 2009-06-08 at 13:47 +0200, Hans Verkuil wrote: > OK, I had some time, so here are a few comments: > > > From: Eduardo Valentin <eduardo.valentin@xxxxxxxxx> > > + <row><entry spanname="descr">Configures the pre-emphasis value for > > broadcasting. > > +A pre-emphasis filter is applied to the broadcast to accentuate the high > > audio frequencies. > > +Depending on the region, a time constant of either 50 or 75 useconds is > > used. Possible values > > +are:</entry> > > + </row><row> > > + <entrytbl spanname="descr" cols="2"> > > + <tbody valign="top"> > > + <row> > > + > > <entry><constant>V4L2_FMTX_PREEMPHASIS_DISABLED</constant> </entry> > > + <entry>No pre-emphasis is applied.</entry> > > + </row> > > + <row> > > + > > <entry><constant>V4L2_FMTX_PREEMPHASIS_50_uS</constant> </entry> > > + <entry>A pre-emphasis of 50 uS is used.</entry> > > + </row> > > + <row> > > + > > <entry><constant>V4L2_FMTX_PREEMPHASIS_75_uS</constant> </entry> > > + <entry>A pre-emphasis of 75 uS is used.</entry> > > + </row> > > + </tbody> > > + </entrytbl> > > + > > + </row> > > Do you know when to use which pre-emphasis? I depends on what de-emphasis filter the receivers are using. This is usually a national or regional regulatory decision imposed upon the broadcasters and hence receiver manufacturers. The choice made by regulators probably depends on allowable field strength (chosen by regulators) for the broadcast area and the expected propagation conditions and man-made noise sources (which dominate in VHF). Preemphasis boosts the normally smaller magnitude, high frequency part of the FM signal spectrum. The receiver can then attenuate the received noise and knock down the higher frequency part of the FM signal back down to normal levels with a deemphasis filter. The end result is an imporved SNR for the high frequency part of the FM signal spectrum. The CX25843 datasheet actually has an OK picture. > There is a similar MPEG > control but I've never been able to find out when to use pre-emphasis and > which mode should be used. I'd appreciate it if you can point me to some > documentation on this issue. And perhaps that info should also be added to > this doc. If you have audio that you know has a lot of high frequency content (televised music concert), you may wish to use pre-emphahsis filtering before compression, and not use it for something with mostly low freqs (a talk show). You of course have to have the pre-emphasis filters available and the ability to enable/disable them. The preemphasis parameters in MPEG are to describe the preemphasis added to the baseband audio, or the deepmhasis that must occur after the audio is decompressed back to baseband. The idea is that digitzation and compression processing are going to introduce noise that will be significant compared to the high frequency components of the audio. A preemphahsis filter amplifies these before sampling. That way, after reconstruction on the other end, the high frequency audio can be deemphasized with a complementary filter. MPEG can carry metadata specifying a 50/15 us or a CCITT J.17 filter characteristic. For 44.1 ksps, 50us and 15 us correspond to filter corner freqs of 1/(50 us * 2 * pi) = 3.183 kHz 1/(15 us * 2 * pi) = 10.61 kHz with a 10 dB change in gain between the two corners (I think). It looks like the CCITT J.17 curve corresponds to the pre-emphasis filter used by NICAM. There's a picture of the J.17 preemphasis curve here in section 4.1: http://tallyho.bc.nu/~steve/nicam.html The CX25843 data sheet also has a picture of a NICAM de-emphasis filter characteristic. (Someone should really pull out the standards and verify all that.) > > + <row> > > + <entry > > spanname="id"><constant>V4L2_CID_TUNE_ANTENNA_CAPACITOR</constant> </entry> > > + <entry>integer</entry> > > + </row> > > + <row><entry spanname="descr">This selects the value of antenna tuning > > capacitor > > +manually or automatically if set to zero. Unit, range and step are > > driver-specific.</entry> > > Same question: is there a reason why the unit is driver-specific? Because for a peaking control, an absolute unit doesn't matter. It would probably be too much work to figure out what they Though for a capacitor, one would expect a unit to be some fraction of a Farad. If I'm guessing right, this value is used to tune an impedance matching circuit for maximum power transfer to the antenna. While tweaking this, I assume one is inspecting an output somewhere else for a peak or minimum level. This really seems like something you usually want to happen automatically with a control loop anyway (which is handled by the 0 case). Regards, Andy > > + </row> > > + <row><entry></entry></row> > > + </tbody> > > + </tgroup> > > + </table> > > + </section> > > </section> > > > > <!-- > > > > Regards, > > Hans -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
