Tvheadend

No idea, the used ioctls are in src/input/mpegts/linuxdvb/linuxdvb_frontend.c - linuxdvb_frontend_monitor(). You may hack this code and provide your results back.

If you use "Force old status", then the same ioctls like dvbsnoop should be used, so it's not clear what's different. I would also report this to the driver developer (probably hw vendor in this case). But guessing from the other user reports - dd drivers are simply different (don't follow 100% the linux's dvbapi).

Also, you may check tvh's log for linuxdvb lines (errors/warnings).

I have the same issue when I use the dd drivers. The proposed solution is to use the media build (also pointed to at their web site). The down side is that you get to build a whole new universe of kernel stuff and it uses (at least 3 days ago) the dd driver 0.9.18. The good side is, the snr values work. Also there are claims that the original dd drivers will not work side by side with other drivers for different cards, whereas the media build driver set while allow to work with TT... something.

Jaroslav Kysela wrote:

No idea, the used ioctls are in src/input/mpegts/linuxdvb/linuxdvb_frontend.c - linuxdvb_frontend_monitor(). You may hack this code and provide your results back.

Thank you for your advice, I found this
https://tvheadend.org/issues/1868

and I tried to hack linuxdvb_frontend.c

746 - mmi->tii_stats.signal = u16;
746 + mmi->tii_stats.signal = u16*100000/0xffff;

763 mmi->tii_stats.snr = u16;
763 + mmi->tii_stats.snr = u16*10000/0xffff;

but not solved the problem. I don't know what can I "hack" too...

Also, you may check tvh's log for linuxdvb lines (errors/warnings).

I don't found tvheadend's log, maybe I need to start debug?

Robin Mitra wrote:

I have the same issue when I use the dd drivers.

Can you solved the problem?

The proposed solution is to use the media build (also pointed to at their web site). The down side is that you get to build a whole new universe of kernel stuff and it uses (at least 3 days ago) the dd driver 0.9.18.

I installed dddriver 0.9.22 (https://github.com/DigitalDevices/dddvb/releases) and I didn't find there any other media link. What do you think about? Do I need to install something too?

The good side is, the snr values work. Also there are claims that the original dd drivers will not work side by side with other drivers for different cards, whereas the media build driver set while allow to work with TT... something.
Yes, I experienced it. I have an DVB-T tuner (Zarlink MT352 chip) and it doesn't work together DD Cine card.

Thank you.

the media build drivers are pointed to here:
http://support.digital-devices.eu/knowledgebase.php?article=124

Note that the 'original' dd drivers work except for the signal values, and maybe they prevent you from using other tuners.
That seems more a general linux legacy that you can't easily build such a driver without building a myriad other things you don't know of/are not interested in.

BTW, did you notice that ddbridge drivers dislike msi ?
Interestingly quite the opposite of the tbs drivers.

s. steve wrote:

I will try your recommended media_experimental packages but I don't know I can use it for DVB-C2/T2 or only S2 card...

I removed DD's driver but it was some complicated first I needed to use rmmod ddbridge, dvb-core and cxd2099 then needed to remove "extra" dir from /lib/modules/kernel_version/

After, I installed media_build_experimental what you recommend here http://support.digital-devices.eu/knowledgebase.php?article=124. Very interest, tvheadend can recognize my old DVB-T tuner now. But this old DVB-T tuner show correct SNR and Signal Strength values and DD Cine tuners can't show any values...

I don't know what can I do that ecah tuner card can show correct values...

As far as I understand the DD linux drivers (any package) don't support signal strength and SNR for DVB-C and DVB-S cards. You can read up all about them at http://www.vdr-portal.de/board18-vdr-hardware/board102-dvb-karten/124665-aktuelle-treiber-für-octopus-ddbridge-cines2-ngene-ddbridge-duoflex-s2-duoflex-ct-cinect-max-s8-sowie-tt-s2-6400-teil-3/, mostly in German, though.

That isn't really 'readable' as it doesn't get to the point any time soon.
Anyway, AFAIK the values are read from the tuner module, (stv090x) aren't they?
However, every vendor is free to choose any reference point for those values acquired from the two registers. so it should be possible, if the ref point is known.
Also if you build the media_build drivers, you do get a value that seems to change somehow with your signal. So I guess they're just using a different tuner module?

I also seem to remember that the windows drivers changed the signal reporting sometime (long ago) from arbitary to NorDig (http://www.nordig.org/pdf/NorDig-Unified_Test_specification_ver_2.2.2.pdf)
Dunno if that is:
a) true,
b) useful,
c) actually the case for linux drivers too.