<!-- Some styling for better description lists --><style type='text/css'>dt { font-weight: bold;float: left;display:inline;margin-right: 1em} dd { display:block; margin-left: 2em}</style>

   pa: anyone with a dvb-s2 card aligned to Hotbird 13E in here?
   Ormu: <u>pa</u>: i have a standalone receiver pointed to 13E
   <br> does anyone here have experience on DVB cards on Windows?
   <br> i have Technotrend TT-Budget C-1501, which shows lots of artifacts when watched using ProgDVB on Windows. On Linux, it works fine (with VLC and Kaffeine). However, if I boot to Linux, then reboot to Windows, the artifacts are gone
   <br> so I suppose the Linux driver loads a properly working firmware or configuration to the card, but the Windows driver does not
   <br> or could this be something related to the motherboard settings?
   <br> the Windows driver is the latest version provided by Technotrend, and it's a BDA driver
   pa: Ormu, oh okay.. no i need someone who can use linux apps with the card..
   <br> Ormu, i have the same problem
   <br> on windoze progdvb shows artifacts
   <br> but they persist iirc
   <br> but i have a different card
   xdarklight_: <u>crope</u>: hello :) are you aware of any issues with those Astrometa AMDVB-T2 v2 devices and DVB-C? I get -EINVAL when vlc calls the DMX_SET_PES_FILTER ioctl: https://paste.kde.org/prl8axlvs#line-19
   <br> (v2 as in r828d and MN88473)
   merbanan: <u>xdarklight_</u>: DMX_SET_PES_FILTER should work, I added support for PID filters
   <br> <u>xdarklight</u>: what kernel/driver do you use ?
   xdarklight: <u>merbanan</u>: 4.6.0 and dvb_usb_rtl28xxu
   merbanan: <u>xdarklight</u>: I'll check
   <br> <u>xdarklight</u>: can you receive or does it not work ?=
   xdarklight: (I'm using vlc which works fine on my other em28xx based card) vlc does not even try to open the frontend -&gt; I'm not receiving anything
   <br> <u>merbanan</u>: please note that this is the special case with a slave demod (might be relevant)
   <br> <u>merbanan</u>: sorry, but I have to go for half an hour - I'll be back in 30 min
   merbanan: <u>xdarklight</u>: yeah, my changes are not in 4.6 :/
   <br> you might need to compile newer drivers
   crope: <u>merbanan</u>: your pid filter patch didnt applied and told that you. you have just forgot to send new patch
   merbanan: <u>crope</u>: look at that :/ my bad
   crope: <u>xdarklight</u>: can you test it using vlc?
   <br> $ dvbv5-zap "Yle TV1 HD" --adapter=0 --frontend=1 --record
   <br> $ vlc --dvb-budget-mode dvb://
   xdarklight: <u>merbanan</u>: ah, found your patch - however, latest comment is that you want to rebase it: https://patchwork.linuxtv.org/patch/32030/
   <br> I just rebased it manually, let's see
   <br> <u>crope</u>: let me test that also - I just have to figure out the channels.conf format for dvbv5-zap
   merbanan: <u>xdarklight</u>: my patch will not magically make stuff work, it will just enable pid filter support
   xdarklight: <u>crope</u>: hmm: https://paste.kde.org/pqt1be26b
   <br> (here's my channels.conf / test.conf: https://paste.kde.org/purvmopdr )
   merbanan: tuning to 122000000 Hz ?
   xdarklight: <u>merbanan</u>: well, provider says symbol rate = 6900, 122MHz, QAM 256 is where channel block 02 lives :)
   <br> ah, a channel at 442MHz seems to work better (QAM 64 though)
   merbanan: I think I know what the problem might be
   <br> I have only tested the demod + tuner at 290MHz at the lowest
   <br> 122MHz probably uses the wrong tuner settings for DVB-C
   xdarklight: <u>crope</u>: 422MHz QAM64 channel: https://paste.kde.org/pftev9qyt - but vlc reports: [00007fe0a4000e28] dvb access error: FrontEndOpen: opening device failed: Device or resource busy
   <br> <u>merbanan</u>: in which part of the puzzle (assuming you're speaking of mn88473)?
   merbanan: <u>xdarklight</u>: just to be sure, do you have digital channels at 122MHz ?
   xdarklight: <u>merbanan</u>: yes - and my em28xx based card receives the programs there just fine
   merbanan: astrometa has a R828D tuner and mn88473 demod
   xdarklight: ah, we're one level further down now - that's where R828D comes into play :)
   merbanan: <u>xdarklight</u>: I have never had digital tv working at that low frequency with the R828D, thus not tested
   <br> I am 90% sure the tuner settings are wrong for that scenario
   xdarklight: ok, I have lots of channels below 290MHz
   <br> let's see where it breaks
   merbanan: do you get lock at 300MHz ?
   <br> and working reception?
   xdarklight: 274MHz still seems to lock
   crope: ERROR    command SYMBOL_RATE (8) not found during store
   <br> ERROR    command INNER_FEC (9) not found during store
   <br> you are missing those 2 required parameters
   xdarklight: <u>crope</u>: but they're there: https://paste.kde.org/pmlxuiipw
   <br> or do I have the wrong format?
   crope: my live cable frquencies are between 266000000 - 378000000 and all are working. got many bug raports that higher frequencies are not working
   <br> [Eurosport] SERVICE_ID = 905 VIDEO_PID = 1103 AUDIO_PID = 3098 4083 4086 4087 PID_06 = 6016 FREQUENCY = 370000000 MODULATION = QAM/128 INVERSION = OFF SYMBOL_RATE = 6875000 INNER_FEC = NONE DELIVERY_SYSTEM = DVBC/ANNEX_A
   xdarklight: my provider uses freqs between 122MHz and 442MHz
   merbanan: <u>xdarklight</u>: do you have the ability to run the astrometa card with the windows drivers ?
   <br> and sniff the usb traffic
   xdarklight: <u>merbanan</u>: ah, one step closer: frontend0 is the mn88473 demod, not frontend1
   <br> with the correct frontend I can also tune below 174MHz (which is the lower limit in rtl2832.c)
   merbanan: <u>xdarklight</u>: but does it work ?
   crope: frequency limits are tuner specific
   <br> but many times those are just set ~170-860 MHz in demod driver as in real life DVB-T frequency allocations are on that
   xdarklight: <u>merbanan</u>: not yet, this is the first time I see the dvb-demod-mn88473-01.fw firmware message
   <br> <u>crope</u>: is your webserver down currently? http://palosaari.fi/linux/v4l-dvb/firmware/MN88473/01/
   crope: <u>xdarklight</u>: works for me
   merbanan: works here also
   xdarklight: indeed, works via VPN to my server
   <br> yay, LOCK :)
   <br> ok, so with dvbv5-zap it's working fine now
   crope: [DVBC] FREQUENCY = 274000000 MODULATION = QAM/256 INVERSION = OFF SYMBOL_RATE = 6875000 INNER_FEC = NONE DELIVERY_SYSTEM = DVBC/ANNEX_A
   <br> just add single entry to scan file and do dvbv5-scan. it creates channel file for you
   <br> usually single working mux is enough as all the rest neighbour muxes are got from that working one NIT tables
   xdarklight: <u>crope</u>: yes, got it - dvbv5-zap is working fine now, but vlc still reports "cannot setup TS demultiplexer: Invalid argument"
   merbanan: <u>xdarklight</u>: I posted a patch on the linuxtv mailinglist, please test it if you still think the error is the pid filter stuff
   xdarklight: <u>merbanan</u>: I had to rebase your patch manually as it didn't apply (that's why it's not upstream yet)
   merbanan: I just sent a patch, please test the patch if you test the code at all
   xdarklight: <u>merbanan</u>: https://patchwork.linuxtv.org/patch/34358/ &lt;-this one seems to be missing the rtl2832_slave_ts_ctrl bit, this is my rebased version: https://paste.kde.org/pa79sqlen
   merbanan: looks like my rebased patch also
   xdarklight: <u>merbanan</u>: dev-&gt;slave_ts = 1; &lt;-in your patch that line is inside rtl2832_pid_filter_ctrl (which looks suspicious to me)
   <br> in your original patch that line was in rtl2832_enable_slave_ts, which is now replaced by rtl2832_slave_ts_ctrl (which not only allows enabling, but also disabling slave TS')
   merbanan: yeah, that is not correct :/
   pa: Hi! anyone with a dvb-s2 card aligned to Hotbird 13E in here at this time? :-)
   merbanan: <u>xdarklight</u>: do you have the energy to update/test the patch and send it upstream ?
   xdarklight: <u>merbanan</u>: sure, just maybe not tonight (will continue tomorrow)
   merbanan: <u>xdarklight</u>: the rtl2832_slave_ts_ctrl() should handle the dev-&gt;slave_ts parameter on it's own
   ***: Rondom has quit IRC (Remote host closed the connection)
   merbanan: ie the rtl2832_init() hunk should be removed
   xdarklight: <u>merbanan</u>: any reason for removing the rtl2832_init hunk? I know that bool defaults to false, but this way it's easier to read (personal preference though)
   merbanan: <u>xdarklight</u>: do whatever is more clear, I just mean that rtl2832_slave_ts_ctrl() should control it also and not just only enable like my original patch did
   xdarklight: okay
   crope: just add flag and set its value inside rtl2832_slave_ts_ctrl(). then select used PID filter according to that flag
   <br> when flag is set use slave ts pid filter and when flag is not se use normal pid-filter
   xdarklight: <u>crope</u>: yep, merbanan already prepared that - this is my updated version which hooks into rtl2832_slave_ts_ctrl https://paste.kde.org/pa79sqlen
   <br> I'll test it tomorrow morning
   crope: now it uses always normal pid filter, so if you are using pid filters for mn88473 slave demod it does not work as wrong filter is used
   pa: it's funny how in all IRC there's nobody else using dvb-s2 :D
   iive: <u>pa</u>: you are asking your question in a wrong way. E.g. somebody might have dvb-s2 and dish with a motor, that is currently not pointing to that satellite.
   <br> <u>pa</u>: you should actually ask your real question...
   <br> (btw, i don't have any dvb-s)
   pa: hm.. that's right
   <br> iive, well the real question is:
   <br> i have a skystar s2 card, with the liplianin drivers, which works wonderfully (e.g., even better than on windows with proprietary drivers)
   <br> BUT
   <br> there is ONE (only one) specific transponder on Hotbird that i cannot scan
   <br> i can scan and tune on windows
   <br> but not on linux with liplianin drivers
   <br> other DVB-S2 transponders work
   <br> this one somehow doesnt
   <br> and the problem is that it is not even found
   <br> now, since other transponders with the same symbol rate and codec and whatnot works just fine
   <br> i wanted to find someone with perhaps another dvb-s2 card able to tune that transponder, scan it with kaffeine, and hand me over the sqlite database containing the binaries of the channel
   <br> and see if with that one i can tune the channel
   iive: by database, you mean the list of transponders?
   pa: iive, kinda long story.thats why i wasnt asking the right question
   <br> iive, no, kaffeine actually stores something more into sqlite
   <br> some binary stuff that the satellite transmits
   <br> i don't fully know the specs, so i donno what that is
   iive: network information table?
   pa: possibly..
   <br> but i know the transponder params
   iive: what are the params of the missing transponder?
   pa: like frequency, polarization, symbol rate, FEC, etc
   <br> so
   <br> moment
   <br> i dig it up
   <br> iive, S2 12475500 H 29900000  3/4 AUTO   8PSK
   <br> i tried also to step around that frequency (between 12475 and 12476), to change the rolloff
   <br> but nothing
   iive: have you tried with stepping the symbol rate?
   pa: but for example i have no problem in scanning and tuning S2 12437000 H 29900000  3/4 AUTO   8PSK
   <br> hm not
   <br> but it's documented 29900000 iirc
   iive: the symbol rate could be off by a little
   pa: hm
   <br> i can try
   <br> thanks
   <br> although.. iirc these were the params spitted by progdvb too
   <br> 12475500 for example
   <br> (it was documented 12476)
   <br> but anyway
   <br> i will try to step the symbol rate :)
   <br> weird thing is: only kaffeine seems to have no problems with dvb-s2
   <br> everything else struggles
   <br> (mplayer, vlc, etc)
   <br> iive, so that's what happens
   <br> http://codepad.org/Q2FZDkWM
   <br> stepping the symbol rate
   <br> it can tune around the symbol rate i wrote
   <br> but it has problems finding the channels
   <br> i also tried w_scan (i think)
   <br> it also didnt find the channels
   <br> and actually kaffeine is the only soft who was able to at least tune 12437 (the other S2 transponder i was mentioning)
   iive: i guess tuning succeeded mean it was able to get a lock.
   <br> the timeout however means that it probably gets only broken data.
   <br> try with another fec, or set it on auto.
   pa: i tried 2/3
   <br> i can try on AUTO (is it possible?)
   iive: try and see
   <br> btw, you may try to use the "working" SR as base and do some stepping at the frequency.
   <br> frequency and SR are controller by clocks and they could be off. That's why tuners and demodulator have automatic fine-tuning.
   <br> however testing for perfect match takes time, and sometime the driver doesn't give enough time to the hardware.
   pa: so i tried to AUTO the FEC. apparently it's supported. but the result is the same
   <br> so i tried to use the working SR and stepping the frequency already
   <br> but do you mean to step both?
   <br> together?
   iive: keep the "working" SR
   pa: ok, then i did the test already
   <br> is there maybe some timeout i can change in the driver?
   <br> i mean i have the source so i guess i could try
   <br> btw i tried also the driver for the SS-S2 that is in the 4.4? kernel
   <br> and that can't tune ANY S2 transponder O_o
   iive: for that maybe the real developers could answer :}
   pa: :-)
   <br> well so, after all those tests, my last hope was to make someone else tune the channel with kaffeine, and use that database to try to tune it here :-)
   iive: btw, that frequency is not at the edge of the supported freq, is it?
   <br> i mean, you probably have transponders with higher and with lower freq.