On 10/9/06, Hans Verkuil <[EMAIL PROTECTED]
> wrote:
Hans,
On Saturday, I finally had some time and I did a lot of testing with my card. Before the +4 patch I had a good signal on tuner 1 and a weak signal on tuner 2. After the patch both tuners were the same quality but both had a distortion patten in every channel that varied depending on the channel #.
There have been a few complaints about the Samsung image quality, even
with the 'Low Noise Amplifier' fix applied (as it the case with
2.6.18/ivtv-0.8).
There are a few complicating issues here which I will try to explain.
First of all, make sure you have a decent signal. The antenna signal is
split into two for the PVR500 (two tuners after all), so this implies a
corresponding drop in signal strength. So if the signal isn't very
strong to begin with, then you will need a antenna amplifier.
But the other problem is that I suspect that while channels with a low
signal strength are now shown properly (or at least as good as the
Windows driver can produce), channels with a high signal strength may
now be worse than before.
The datasheet for the PAL variant of the Samsung tuner actually has a
flow chart showing when to set or clear the LNA bit, but the NTSC
datasheet has no mention of this. When I worked on this tuner I asked
someone who had the PAL tuner to take a look at a certain tuner
register to see how it fit with the flow chart, but it made little
sense to me at the time, so I ignored this flow chart completely.
But based on the feedback of the last couple of days I may need to look
at this again. But I need help with this!
First I need confirmation that the setting of the LNA bit does in fact
make some channels worse. So if you are using 2.6.18 and have these
problems, then please go to drivers/media/video/tuner- types.c in the
kernel sources, search for 'Low Noise' and you'll see the PAL and NTSC
Samsung tuner entries. Remove the '+4' from your tuner, recompile,
install and check if the problematic channels are now improved. Please
let me know whether or not that is the case!
Assuming that removing the LNA does in fact solve this problem, then I
need help from someone who can compile kernels and do some tests and
preferably can do some testing/programming without too much help from
me.
I tested different values added in tuner_samsung_tcpn_2121p30a_ntsc_ranges, an antenna amplifier and a couple of splitters and different cable routing. I found out although changing the values from +0x04 to +0x14, + 0x24, +0x34,+0x44 and +0x54 did appear to change the interference pattern what did the best was to attenuate the signal using splitters. I knew my signal was a little strong as my cable modem says that my channel power is almost +6dBmV which I am told is supposed to be around 0dBmV. Anyways after trying many combinations and looking at the channels on both tuners I found I got the best result with a 3 way splitter with the input signal on the a 7dB output leg and the output signal (fed directly to the card) on a 3.5dB output leg. If I put either cable on the input the distortion was visible on most channels although it was less than without the splitter. With the setup using the output legs only both tuners have a good picture on most of the channels however there still is a visible distortion pattern around channel 45 +- 3 channels. This is with the default values of +4 in drivers/media/video/tuner- types.c.
John
_______________________________________________ ivtv-users mailing list [email protected] http://ivtvdriver.org/mailman/listinfo/ivtv-users
