[Rejected] Signal Strength for more TV cards. (1 Viewer)

[elliottmc]

Hi Framug,

There has been some discussion of this issue in IRC.

As it stands at the moment, I suspect your patch is good, but possibly only for the cards which you have. This was not explicitly discussed, so this bit is mainly my opinion.

Different vendors report different raw values for signal strength, so there is no universal way of doing this.

The team have a request for you, so that eventually your code (modified as necessary) will make it into SVN.

Would you be able to add logging code to the channel scan, so that a log file is produced containing details such as card manufacturer, driver version and signal strength during scan? This should be done without your other code changes in the first instance.

Presumably the next step will be to modify your patch to take care of cards from more different vendors.

I do promise, the team is listening and appreciates your help. It just doesn't make sense to use your patch as it is, since it is clear that the numbers reported for some cards will make no sense.

Best wishes,

Mark

 

[framug]

Hi elliottmc,

I assure you my patch (wich implies modifications in many classes, as you can see) will work all the time, without breaking something for users.
I took an extreme care to dev it (no hardcoded values, respect all MP quality code, etc..) and, I really tested all possible cases.

For the signal calculation, I changed a bit my first proposal, since the "3" "-97" reported.
What you have to know :
- mDB (1/1000 DB) is always converted as DB then, the -100000 to +100000 become -100 to +100 rank.
- Positive signal is reported as it in all case.
- Negative signal is reported by two ways, now :
* You don't check Nova T-500 like for the cards ==> the -97 value will became +97.
* You check Nova T-500 like for the cards ==> the -97 value will became +3.

I don't know if this expression exist in english : "J'en mettrai ma main au feu" translation "I'm ready to put my hand in the fire".
It means I'm ABSOLUTELY sure it will work all the cases.
Devs, please give me a signal value sample wich can't work with that.

Of course, I could dev what you ask for but, it will take many work and time, and, I must go away for some days.
Then, it will not be ready before 24th july.

This is asked only because you think there are cases wich can't work ?
Please, give me a sample not working before.
I thought MP devs wasn't so frightened.

If the patch doesn't work, I promise you to provide another lot of patchs wich will revert all the modifications I made and this, no matter what SVN number will be.
I can provide you all .dll and .exe you want for test but, you have to know that, you have to delete your MP database before using it.
Then, you will loose all your params (channels, etc) or, maybe are you able to just delete "card" table ?
You could share it with others devs who have others TV cards types and, you will see it can't break something.

Also, I thought SVN build was made for feedback users testing, isn't it ?
Do you bet with me how many negative percentage feedback users will be ?

What do you think about ?

Franck.

 

[ronilse]

Hi,
Nope we don't try to break SVN, only "tested" stuff should be in with some control & changes like this will most probably need separate tests from all users that can to get a wide spread results from as many cards manufacturer as possible & wrong reported values/possible bugs be corrected & all this before a new release.
I've not had time to test this patch yet(just to much work in real life)

Regards
Roy

 

[elliottmc]

Hi Framug,

I think the basic problem (as I understand from discussions in IRC) is that not all cards report signal strenght in mDb, even if they say they do.

So two cards from different vendors will report different signal levels on the same channel/same aerial. This is the problem that needs dealing with.

Without knowing the range of signal strengths, a proper fix is impossible.

Which types of card have you tested your fix with ?

Best wishes,

Mark

 

[framug]

Roy : I understand then, maybe could it be possible to create a specific and unique SVN build wich will be intended only to have a users report for signal, then.
Just tell them that it is ONLY for that and, they NEED to delete their database before.

And, I also have a real life and busy work right now then not really more time that already spent.
Cause I could see what amazing work was made by MP devs for providing the MP code, I tried to be nice also and, offer my help as I can.
The goal was for help MP devs a bit and, a better users experience.
Now, sorry but, I have no more clue to help you more than I did.


elliottmc : Normally, there is no other possibility than report signal in mDb or in Db.
And even if it was the case (I'm pretty sure not) then, it will be a numeric value in all cases.

Finally, what is better :
Leave MP like it is today and, having many users who complain because MP report a no signal for their TV cards then, they can't use it or, having a signal strength value for all cards ?
It's up to you.

The last improvement of my code I could see could be to transform the bool value in the database by a INT, the checkbox by a listbox, for better evolutivity but, I'm pretty sure it wil not be necessary.

I tested with theses cards and sorry, again, I can't do more because I'm only a developer in the real life, not a card seller :
Analog :
- Hauppauge PVR-350.
- Hauppauge PVR-500.
DVB-T :
- NOVA-T500 model 84XXX
- Pinnacle DUAL DVB-T PRO PCI (doesn't work but, not for the same reason, explained on the #1 post-).

Bye.

Edit : I forgot to explain :
I didn't invent the -100 000 to +100 000 values.
Take a look inside remarks in the BDAMEDIA file, you will see it's explained.
Also, if vendors doesn't report values correctly, they can fix it by providing newer drivers wich will provide correct value.
In this case, it will not be a MP fault, it will be a vendor fault.
Today, the no signal value is a MP fault, sometimes.

 

[tourettes]

The signal values are completely non-comparable even within one manufacturer's cards Signal Strength/Quality Utility - Hauppauge UK Forum

There are few possible ways to sort this out:

• Used only signal present (yes/no) and present thesignal quality level to user (this should be in %)
• Gather data from ALL DVB cards and build a database how to convert those values to dB
• Do nothing and use the current way (few cards at least will report something useful)

 

[tourettes]

Also, if vendors doesn't report values correctly, they can fix it by providing newer drivers wich will provide correct value.
In this case, it will not be a MP fault, it will be a vendor fault.

Unfortunately we cannot rely on that as the issue on manufacturer's side won't be ever solved.

 

[rtv]

The signal values are completely non-comparable even within one manufacturer's cards Signal Strength/Quality Utility - Hauppauge UK Forum

There are few possible ways to sort this out:

• Used only signal present (yes/no) and present thesignal quality level to user (this should be in %)
• Gather data from ALL DVB cards and build a database how to convert those values to dB
• Do nothing and use the current way (few cards at least will report something useful)

Nice link - for those that might be to lazy to read the docs:
This is just Hauppauge - ONE of many card sellers...

• 
  [*] Nova-t PCI/HVR-1100/HVR-1300 - Range of values = -41000 to +5000
  [*] HVR-1110 - Range of values = Unknown
  [*] Nova-t USB2 - Range of values = -20 to unknown
  [*] Nova-t Stick - Range of values = -100, 0 to 100

 

[tourettes]

Few other HW (tuner makers actually):

DVB-t - BDA Signal Stats

For example, Conexant use values from -41000 to +5000, DiBcom use -20 to +70 (ish), and eMPIA use values such as 25000.

 

[framug]

Hi tourettes,

How are you, today ?
The first link you posted talk about a specific .exe.
Then, are you sure it's really low-level reported or, an interpretation of the program ?
(I mean already modified values).

This guy, is he a Hauppauge salariee ?
How can he affirm this ?
If it's really what he wrote then, no problem, I agree that, it could exist many signal calculation.
And, your proposals to sort this out are also really excellent.

Therefore, I would like you take a look in the bdamedia file, please.
It's not me who write rules, it's MicroSoft for a proper signal strength return on windows OS.

//  OPTIONAL
//  Carrier strength in mDb (1/1000 of a DB).
//
//  A strength of 0 is nominal strength as expected for the given
//  type of broadcast network.
//
//  Sub-nominal strengths are reported as positive mDb
//
//  Super-nominal strengths are reported as negative mDb
//

Now, if vendors doesn't respect it, it's vendors problem and, users will be aware of that and finally, they will not buy such cards wich return bad values.
Finally, you will see that, this will force vendors to change their driver, to make more money.

One example, by chance : Nova-t PCI/HVR-1100/HVR-1300 - Range of values = -41000 to +5000
Since, it appears it's mDb, the +5000 is +5 and, I wonder if this case really can exist.
I'm very curious to really see +5000 during more than 1-3 cycles.
I suppose it's because of a 5% hardware tuner error marge.

But finally, you pointed it to me that unfortunately, there are many signal level wich exist then it appears that a boolean is probably not enought.
Now, if you have time to dev your proposals, no problem, please, reject this proposal.
If you think that, my solution with an INT instead of a boolean could be interesting please, let me know.
This solution could evolve in the time, depending on feedback users.

The last thing I would like to say :
I proposed my help nicely therefore, it appears to me that, it's more a "Payment of accounts in OK Corral" than an interesting exchanges of ideas.
If it's case, I really wonder it is worth to continue this way.

Regards.