MP 1.2.3 - minor issues/questions (1 Viewer)

[mm1352000]

Hello again

Yes, I tried disabling the plugin. Same result. I'll recheck the mapping in the built-in area, and try again.

I don't understand. TVserver has to use both tuners because it's recording two different tv shows at the same time. One tuner per show -- right? (this is why I have a two tuner card)

Not at all. This is true for WMC, but not for MP. With MP you should be able to timeshift/record all the channels from one frequency (AKA "physical/RF channel") with one tuner.

It should result in TV Server using the first tuner for both recordings.

Again, I don't see how this is possible. What am I missing? (I'll try this if you still want me to)

This is possible because physically speaking, the limitation is that each tuner chip can only tune to one frequency at a time. However there is no reason why all the channels that frequency/RF channel cannot be received at the same time. Please do try it.

signal -- all the channels are fine by themselves

So I wonder if that internal signal splitter is a factor.

tuning parameters -- sure, maybe, but how?

If the frequencies were wrong, but close enough to barely get lock then this could happen.
For example I'm checking the frequencies for Cox Cable, Tuscon in the Silicondust channel DB (link) and the frequencies seem quite different. That could be because I'm looking for the wrong postcode or because the frequencies have changed recently and not been updated in the SD Db. Since you know your own post code, perhaps you could check:
http://www.silicondust.com/support/channels/

hard drive activity - don't think so, no other user level programs run on this machine.

So to be clear: no real time scanners (virus protection, search indexing etc.)?

overheating - I really don't think so. Again, it's the same hardware as runs 1.1.0.0 successfully.

Okay.

I don't see why the 2nd partition would cause any extra disk activity (eg, file system administration) - but could it?

I guess it could (depending on how you made it) if the data is very fragmented.

I think I'll have to wait until you've tried the patch that I posted. If it works, it suggests the HDD is *not* the problem (since theoretically the HDD should experience roughly the same workload for recording two channels regardless of whether they are recorded from the same tuner or not); otherwise it suggests the HDD is the problem.
When you try the patch, could you please test recording the two digital channels twice; once for each tuner. This will help to confirm that the tuner hardware itself is okay and that each tuner performs identically.

mm

 

[mm1352000]

Here's a little results matrix for our amusement --

Did you try MCE and HID enabled at the same time (yes, this can be done - quite a few people have to do this I think)?

Also discovered that I am unable to add keys and mappings in Keys and Sounds(? or Default Mappings, one of those)), even though there is an Add key.

Just tried it, and I can. Depending on where the cursor is when you click "add" it may create a new window, and then you have to add actions within that. The "key" field is always disabled, but you should be able to select it and press the key on your remote to map.
To be honest though, I have always used EventGhost to handle remotes. I find it much easier. You can map the remote buttons per-application and do many more actions (including generic actions like maximising or minimising a window) than are available in MP.

 

[cewillis]

mm,

1) the simple things -
You caught me -- I had indexing enabled (like on the 1.1.0.0 partition). I turned it off, and it didn't make any difference. 2 qam tunings at the same time still mess up.

I tried MCE and HID both enabled, KIP disabled. Right works, but media controls don't. Enable KIP, Right doesn't work, media controls do. Is there a FINAL, FINAL mapping file somewhere, after all versions of mapping have been 'applied'?
Or is there some way to DISABLE changing the zoom setting in MP?

I turned off Dynamic Refresh (same as 1.1.0.0 partition). Still get the 2 second blank screen when exiting any playback.

Yes, I can add to Keys and Sounds. Don't know what happened yesterday. (and don't know if adding will affect anything)

2) a review and opinion -
I have 2 disk partitions using exactly the same hardware, except for the partioning. Specifically, the identical, same tuner in each. I use the 1.1.0.0 partition every day for routine activities - recording and playback, and it still works fine. I use the other partition to experiment with a 1.2.3 install, but no real 'work', yet.
So it seems to me that there is no point in trying to identify a tuner hardware problem.
But I can certainly believe that the exact same driver could (shouldn't, but could) behave differently in different installs depending on what else is installed, or even the order of installs. For instance, I have still have Hauppauge WinTV installed in the 1.1.0.0 partition from when we were trying to figure out why the tuner wasn't working (solved when my provider moved some qam channels out of the FM band).

3) debugging anyway, maybe

1. Stop the TV service in the "manual control" section of TV Server configuration.
2. Close TV Server configuration.
3. Take a backup of the TvLibrary.Interfaces.dll in your TV Server install folder (typically C:\Program Files\Team MediaPortal\MediaPortal TV Server on XP).
4. Download and extract the attached DLL into that same folder.
5. Open TV Server configuration again (you'll get a prompt that the service is not running - click yes to start the service) then try the test again.

I think I did this, and it didn't make any difference.
But, now I'm not sure what was suppose to happen. The extract created a sub directory in \Program Files\Team MediaPortal\MediaPortal TV Server. Is that what you wanted?
I wasn't using TV service manual control to test. I was doing Record now in MP to start one recording, then another. Does the latter accomplish the test you wanted?

4) education
You are the expert, and I still don't understand something.
What I think I know is that channels, not frequencies, are mapped to tuners in TV service configuration. Right so far?
Sounds to me like you are saying that there is a test somewhere, such that if a new channel tuning request specifies the same frequency as is already being tuned, a new tuner is not allocated. Is that right?
If so, can't we test that? I should be able to tune 3 (or more) channels at the same time, even though I only have 2 tuners. That doesn't work for me: if I start recordings on the two channels you said were on the same frequency, all other channels are then shown as unavailable. (this on 1.1.0.0 since that can record 2 qam channels at once with no problem)
So, where am I going wrong?

Sorry to be dense. Not deliberately trying to waste your time ---. Much appreciate your efforts.
Cal

 

[mm1352000]

Hello again

Is there a FINAL, FINAL mapping file somewhere, after all versions of mapping have been 'applied'?

Lets be clear, because there are two mappings involved here. We have:
remote button --> key
key --> MP action

I know very little about the first mapping (I've always used EventGhost for this) except that it is handled in the "remotes" section of MP configuration.
The second mapping is handled in the keys and sounds section of MP configuration.

I note from the "keys and sounds" section in the wiki:

Note:

This section does not support the "media keys", i.e. the volume up/down, play, stop, next, previous, etc. buttons you can find on most modern keyboards. To support these keys, you need to enable the General HID Device in the Remote section.
Keys and Sounds are stored in Program Data keymap.xml Before you begin making major key changes, you may wish to back up this file first!

So part of the answer to your question is the keymap.xml file, for the mapping from key to MP action.

Or is there some way to DISABLE changing the zoom setting in MP?

It may be possible to remove or change the key that is linked to this action, but I don't think you can disable it entirely.

2) a review and opinion -
I have 2 disk partitions using exactly the same hardware, except for the partioning. Specifically, the identical, same tuner in each.

Well, yes it is the same tuner card but not necessarily the same tuner tuner... if that makes sense. In other words, there are two tuners on the card but we haven't ruled out whether performing both recordings with one tuner results in the same issues.

So it seems to me that there is no point in trying to identify a tuner hardware problem.

I think this is the crux of where we have the differing understanding/opinion. I'll try to explain what I'm saying in more detail once we've got through the other questions/points...

But I can certainly believe that the exact same driver could (shouldn't, but could) behave differently in different installs depending on what else is installed, or even the order of installs. For instance, I have still have Hauppauge WinTV installed in the 1.1.0.0 partition from when we were trying to figure out why the tuner wasn't working (solved when my provider moved some qam channels out of the FM band).

It could I guess, but as you say - it shouldn't.

3) debugging anyway, maybe
...
I think I did this, and it didn't make any difference.
But, now I'm not sure what was suppose to happen. The extract created a sub directory in \Program Files\Team MediaPortal\MediaPortal TV Server. Is that what you wanted?

No, it shouldn't have created a subdirectory. The zip file contains one file - the patched TVLibrary.dll. What I am asking you to do is replace the TVLibrary.dll that TV Server is currently using with the one from the zip file.

I wasn't using TV service manual control to test. I was doing Record now in MP to start one recording, then another. Does the latter accomplish the test you wanted?

Yes, kind of, but not entirely.
When you start recordings in MP you can't control or know which tuner will be used to perform the recording. Manual control allows you to see and/or specify which tuner should be used. I'll come back to this.

) education
You are the expert, and I still don't understand something.
What I think I know is that channels, not frequencies, are mapped to tuners in TV service configuration. Right so far?

Yes, that is correct. In the mapping section of TV Server configuration you link channels to tuners. I'll try and clear up the confusion with the frequency stuff in a mo.

Sounds to me like you are saying that there is a test somewhere, such that if a new channel tuning request specifies the same frequency as is already being tuned, a new tuner is not allocated. Is that right?

Not quite. Let me try and explain more clearly.

Your cable provider divides the capacity of the connection to your house into 6 MHz slots. Each of these slots is called an RF channel (some people also call it a physical channel). Same principle applies with over-the-air TV. That "magic" 6 MHz number comes from ATSC, SCTE and/or FCC rules and standards (other countries may use 7 or 8 MHz slots). In the "bad old analog days" it took one 6 MHz slot to carry just one analog channel either over-the-air (NTSC) or via cable (analog cable).

In the past few years the US has been transitioning from analog TV transmission to digital TV transmission. So NTSC -> ATSC and analog cable -> digital cable. Digital transmission technologies have enabled TV broadcasters to use much less bandwidth per channel - more like 1 MHz rather than 6. However as mentioned above, they still divide the available capacity into the old 6 MHz slots.

Isn't that a waste, I hear you ask (?). Well, yes... if they only transmitted one channel in each slot. In practice what they do is transmit a stream that contains a bunch of channels, so it is not so wasteful after all. This idea of a bunch of channels being transmitted together is called multiplexing, the combined stream is called a transport stream or TS for short (note that a TS may contain only one channel, but for efficiency reasons the broadcasters usually multiplex as many channels as they can fit within the 6 MHz slot).

Changing tach here for a moment, lets talk about tuners...

Tuners are capable of tuning 1 RF channel (or slot) at any given time. In the past that meant that a person would need one tuner per [analog] channel. Now with digital, tuners can still only tune one RF channel at a time but they receive the combined transport stream. This means that you can receive more than one channel with just one tuner. The channels that you get depend on which channels the broadcaster decided to multiplex together.

Does that all make sense so far?
I hope so. If not just ask... but for now I'll assume the answer is yes.

Okay. So getting to TV Server now. TV Server is smart. It allows you to timeshift or record all the channels which are available in the multiplexed transport stream. You can almost think of it like a splitter that splits the channels that you want to watch out of the full multiplex into recording or timeshift files on your hard drive. Compare TV Server with WMC: WMC can only receive one channel per tuner like in the analog days, so in that sense TV Server has a huge advantage D).

When I said that the two channels you had tested were on the same frequency, I meant that they were transmitted in the same slot or multiplex. In other words, TV Server should be able to receive both of them using just one tuner. As it turns out, there is a bug in TV Server that is preventing that feature from working properly for ATSC and cable channels... and that is what the patch is for - to re-enable the feature so-to-speak.

So with the patch, you should be able to disable each of your tuners in turn but still be able to record both the channels that you previously tested with. In fact, if you disable all except one tuner in TV Server configuration and then start timeshifting/recording one channel in MP, MP will tell you which other channels you can still watch or record (see here - the icon key)

Hopefully you're still with me?

So what I want you to do is to test recording those two channels with each of the two tuners on your HVR2250 in isolation. This will take two tests; one with the first tuner enabled and the second disabled, and the second test with the first tuner disabled and the second tuner enabled. You can disable/enable tuners in the tuner list in TV Server configuration by unticking them:
http://wiki.team-mediaportal.com/1_.../TV-Server_Configuration/02_TV_Servers#Tuners

This way we can check whether the problem only occurs when both tuners are active.

4
If so, can't we test that? I should be able to tune 3 (or more) channels at the same time, even though I only have 2 tuners. That doesn't work for me: if I start recordings on the two channels you said were on the same frequency, all other channels are then shown as unavailable. (this on 1.1.0.0 since that can record 2 qam channels at once with no problem)
So, where am I going wrong?

Hopefully the explanation above covers off the questions here. To answer briefly: yes, that is exactly what I want you to test (). You should be able to record/tune 2 or more channels at once *as long as the patch is installed correctly*.

Sorry to be dense. Not deliberately trying to waste your time ---. Much appreciate your efforts.

No worries - I'm fully aware that this stuff is a little complex, and I breezed over it quite quickly assuming that you knew about this from the previous problem. My fault for assuming.

mm

 

[cewillis]

Hi mm,

First something I forgot about

Since you know your own post code, perhaps you could check:
http://www.silicondust.com/support/channels/

Silicon dust doesn't show my provider (local, small time group) -- only comcast and satellites.

Does that all make sense so far?
I hope so. If not just ask... but for now I'll assume the answer is yes.

Yes, indeed. Thanks. I did 'mostly' understand the situation with broadcast/cable frequencies and multiple channels, and it's even more clear now with your (repeat) explanation. I never thought about MP extracting multiple channels from one tuned frequency since as far as I knew/know, MP doesn't do that. (instead allocating a separate tuner to every channel, without regard to the channel's frequency)
In other words --

TV Server is smart. It allows you to timeshift or record all the channels which are available in the multiplexed transport stream.

As it turns out, there is a bug in TV Server that is preventing that feature from working properly

-- I wasn't understanding that.
But now I do.

So on to the tuner tests ---

tv-new-dll-1sttuner has test results on tuner 1: signal quality 100. tv-new-dll-2ndtuner, on tuner 2: signal quality << 100.
I seem to remember from somewhere that quality isn't really a useful measure for digital, since the signal is either there or it isn't. Either way, there is a difference in the tuners.

tv-new-dll-2tuner has results with both tuners enabled: It looks to me like TVserver took the free tuner (#3) for the second channel, even though the busy tuner (#2) was already tuned to the correct frequency. Again, tuner 2 signal quality << 100.

2012-08-06 11:42:03.703125 [(5)]: Controller: StartTimeShifting started on card:2 to D:\Documents and Settings\All Users\Application Data\Team MediaPortal\MediaPortal TV Server\timeshiftbuffer\live2-0.ts.tsbuffer
2012-08-06 11:42:15.281250 [(15)]: Controller: StartTimeShifting KGUNHD 9
2012-08-06 11:42:15.281250 [(15)]: Controller: find free card for channel KGUNHD
2012-08-06 11:42:15.281250 [(15)]: Controller: find card for channel KGUNHD
2012-08-06 11:42:15.281250 [(15)]: Controller: got 1 tuning details for KGUNHD
2012-08-06 11:42:15.281250 [(15)]: Controller: channel #1 ATSC:tv:unknown KGUNHD Freq:139250 ONID:0 TSID:0 SID:3 PMT:0x20 FTA:True LCN:10000 phys:18 maj:9 min:1 mod:Mod256Qam
2012-08-06 11:42:15.281250 [(15)]: Controller: card:1 typevbIP is disabled
2012-08-06 11:42:15.281250 [(15)]: Controller: card:2 type:Atsc can tune to channel
2012-08-06 11:42:15.281250 [(15)]: Controller: card:2 type:Atsc users: 1
2012-08-06 11:42:15.296875 [(15)]: Controller: card:3 type:Atsc can tune to channel
2012-08-06 11:42:15.296875 [(15)]: Controller: card:3 type:Atsc users: 0
2012-08-06 11:42:15.296875 [(15)]: Controller: card:4 type:Analog is disabled
2012-08-06 11:42:15.296875 [(15)]: Controller: card:5 type:Analog is disabled
2012-08-06 11:42:15.296875 [(15)]: Controller: card:6 type:RadioWebStream is disabled
2012-08-06 11:42:15.296875 [(15)]: Controller: found 2 card(s) for channel
2012-08-06 11:42:15.296875 [(15)]: AdvancedCardAllocation.GetAvailableCardsForChannel took 11 msec
2012-08-06 11:42:15.296875 [(15)]: Controller: card:3 type:Atsc is available
2012-08-06 11:42:15.296875 [(15)]: Controller: found 2 free card(s)
2012-08-06 11:42:15.296875 [(15)]: AdvancedCardAllocation.GetFreeCardsForChannel took 13 msec
2012-08-06 11:42:15.296875 [(15)]: Controller: try max 2 of 2 cards for timeshifting
2012-08-06 11:42:15.296875 [(15)]: Controller: Timeshifting failed, lets try next available card.
2012-08-06 11:42:15.296875 [(15)]: card: CardTune 3 KGUNHD setuptv-9-3:3:-1
2012-08-06 11:42:15.296875 [(15)]: card: Tune 3 to KGUNHD
2012-08-06 11:42:15.296875 [(15)]: card: user: setuptv-9-3:3:-1 tune ATSC:tv:unknown KGUNHD Freq:139250 ONID:0 TSID:0 SID:3 PMT:0x20 FTA:True LCN:10000 phys:18 maj:9 min:1 mod:Mod256Qam
2012-08-06 11:42:15.312500 [(15)]: TimeShifter.OnBeforeTune: resetting audio/video events
2012-08-06 11:42:15.312500 [(15)]: atsc:Tune:ATSC:tv:unknown KGUNHD Freq:139250 ONID:0 TSID:0 SID:3 PMT:0x20 FTA:True LCN:10000 phys:18 maj:9 min:1 mod:Mod256Qam

Seeing the persistent lower quality on tuner 2, but knowing that 2 qam channels record fine on 1.1.0.0, I simulated the test on 1.1.0.0 by starting two recordings from MP, since that TV server does not permit two simultaneous manual tunes.
tv-1100 does indeed show lower quality on tuner 2, but that doesn't show up in the actual recording ?!?!?!?! I watched the tuner 2 program for a numbing 20 minutes (a cooking show), and at most saw two minor flickers. Nothing at all like the abominations on tuner 2 with 1.2.3.

Are there 2 problems here? Does 1.1.0.0 handle errors better?

Again, many thanks for you help. I won't be surprised if you want to leave this for a while ---
Cal

 

[mm1352000]

Hello again Cal

Silicon dust doesn't show my provider (local, small time group) -- only comcast and satellites.

Ahhh, that would explain why the frequencies are so different then.

...

-- I wasn't understanding that.
But now I do.

Excellent.

So on to the tuner tests ---

tv-new-dll-1sttuner has test results on tuner 1: signal quality 100. tv-new-dll-2ndtuner, on tuner 2: signal quality << 100.
I seem to remember from somewhere that quality isn't really a useful measure for digital, since the signal is either there or it isn't. Either way, there is a difference in the tuners.

Well this isn't entirely true. The quality measure is usually derived from the BER (bit error rate) of the stream, which is in turn is determined by error checks that the tuner driver performs. Each packet in the transport stream is transmitted with extra bytes (16 for DVB, 20 for ATSC) that help the tuner to determine whether the stream is corrupted or okay - in some cases it is possible for the tuner to correct errors.

So in summary, the measure is definitely useful even for digital, and as you say there appears to be a significant difference in the quality of the signal received by the two tuners. Interestingly, we can tell that only the second tuner is affected by the result of your test with only tuner 2 - the signal quality is still significantly lower than the signal quality for the first tuner, even though the first tuner is not operating at the time.

tv-new-dll-2tuner has results with both tuners enabled: It looks to me like TVserver took the free tuner (#3) for the second channel, even though the busy tuner (#2) was already tuned to the correct frequency. Again, tuner 2 signal quality << 100.

This is strange. How did you perform this test - in the manual control section of TV Server configuration? If so, did you have a particular tuner selected when you clicked start timeshifting?

Seeing the persistent lower quality on tuner 2, but knowing that 2 qam channels record fine on 1.1.0.0, I simulated the test on 1.1.0.0 by starting two recordings from MP, since that TV server does not permit two simultaneous manual tunes.
tv-1100 does indeed show lower quality on tuner 2, but that doesn't show up in the actual recording ?!?!?!?! I watched the tuner 2 program for a numbing 20 minutes (a cooking show), and at most saw two minor flickers. Nothing at all like the abominations on tuner 2 with 1.2.3.

Are there 2 problems here? Does 1.1.0.0 handle errors better?

This is indeed a surprising result. I have a suspicion that the logging of continuity/discontinuity errors may have been disabled in MP 1.1.x, and that the actual action of logging the errors [in MP 1.2.3] may be causing more disruption than the stream corruptions themselves. This should be relatively easy to test - I'll provide a TsWriter.ax with the logging disabled as soon as I can find some time (at work at the mo, but may not be this evening as I'm quite busy this week).

Again, many thanks for you help. I won't be surprised if you want to leave this for a while ---

No, I'm game to keep looking - just a little time-constrained with other things at the moment. Lets see how we go...

mm

 

[cewillis]

This is strange. How did you perform this test - in the manual control section of TV Server configuration? If so, did you have a particular tuner selected when you clicked start timeshifting?

Yes, manual control, and I would have sworn there was no tuner selected. But I repeated it to be sure (1st attachment), and this time the 2nd channel was added to the first tuner (if I read this right). I'm going to have to try this a few more times ----

To throw in some more mystery, using 1.1.0.0, I tried disabling the first tuner, and using manual control t0 start timeshift on the 2nd tuner. (2nd attachment). In this case, the initial quality report is 100. (unlike using only the 2nd tuner in 1.2.3)

No, I'm game to keep looking - just a little time-constrained with other things at the moment. Lets see how we go...

Great - no hurry. Sure appreciate it.

Cal

 

[mm1352000]

This is strange. How did you perform this test - in the manual control section of TV Server configuration? If so, did you have a particular tuner selected when you clicked start timeshifting?

Yes, manual control, and I would have sworn there was no tuner selected. But I repeated it to be sure (1st attachment), and this time the 2nd channel was added to the first tuner (if I read this right). I'm going to have to try this a few more times ----

Yes, the repeat test looks like it used the first tuner twice - strange. Let me know if you think there is something wrong that you can reproduce.

To throw in some more mystery, using 1.1.0.0, I tried disabling the first tuner, and using manual control t0 start timeshift on the 2nd tuner. (2nd attachment). In this case, the initial quality report is 100. (unlike using only the 2nd tuner in 1.2.3)

<sigh>
The worst problems are the ones that can't be explained!
In my last post I said it looked like only the second tuner was affected by the problems. I wonder if the onboard signal splitter is causing these issues - maybe it is disabled by default (one tuner or the other gets all signal), and only when enabled (ie. both tuners need signal) does it cause problems. An interesting test that you could do to confirm this is force the second tuner to tune the first channel, then force the first tuner to tune the second channel - we'd be looking to see whether the signal quality switches (ie. tuner 2 gets 100% and tuner 1 gets ~30%).

Great - no hurry. Sure appreciate it.

The TsWriter that I promised is attached. Installation is the same routine as with the TVLibrary.dll that you already have. Remember to back up the original one just in case. The only change here is that the continuity/discontinuity logging is partially disabled - I've disabled the line "Continuity error..." that appeared most in your log files but left the others enabled for now.

mm

 

[cewillis]

Yes, the repeat test looks like it used the first tuner twice - strange. Let me know if you think there is something wrong that you can reproduce.

Tried it a few more time, just looking at the tuner statui in Manual control, using 1st and 2nd tuners separately. Put both channels on the same tuner each time. Looks like my mistake -- sorry.

To throw in some more mystery, using 1.1.0.0, I tried disabling the first tuner, and using manual control to start timeshift on the 2nd tuner.

Also did this a few more times, but watched the quality reading on manual control. The reading would 'occasionally' did to 90, 80, 70, 60, but was usually at 100. Dipped 7 times in one 2 minute period. I think this correlates pretty well with subjective viewing impression that the channel is fine.

Then, I installed your newer tswriter, and repeated the previous test on 1.2.3, using tuner 2. 33 observed quality dips in 2 minutes -- but in limited looking, tv looks fine on tuner 2 (with tuner 1 disabled). Unfortunately, trying to record/watch 2 channels still has major problems, on tuner 2, or 1, or both (using the old tvLibrary.Interfaces.

I'm really thinking there much be something wrong with my 2nd partition, or the whole installation. The 1.2.3 results just aren't making any sense.

Cal

 

[cewillis]

mm,

A large part of my problem might be the 2nd partition on my disk. First, the disk is a WD 'green' disk, and I've found a few reports of poor performance compared to 'non-green?' disks. (although the 1.1.0.0 install on the first partition seems fine) Second, it's a 1TB disk, Advanced Format (4096 byte/sector) disk, that is apparently suppose to have partitions aligned on 4096 byte boundaries. Mine isn't, as of now. I tried to do another complete install starting with re-formatting the partition. During install of .net updates, it was obvious the the disk was running much slower (less that 1/2 as fast) than the primary disk in my office computer.
Not sure how this would affect quality reporting, but it certainly seems reasonable that it would affect simultaneous recordings.
I'll try to get the partition aligned correctly, and / or try a different disk, and report back.

Cal