Device / Channel Assignment

[dixxfixx]

I always thought that if I put on a new device and gave it a single channel that was not in any other device's channel list, viewing or recording that channel would necessarily cause that device to be chosen. Was I wrong, or is something interesting happening?

I added a fourth HDPVR2, added a fake channel (666, Zulu) and selected it from the guide. NextPVR chose one of the other devices.

Logs attached, does anyone want xml or db3 files?

[johnsonx42]

what makes you think it chose one of the other devices? the log says it chooses the only one that had the channel, and it all looks consistent to me:

showing livetv on "zulu" starting on hdpvr #4:

Code:

2015-10-02 11:57:10.437    [DEBUG][10]    StartStream(7945:Zulu)
2015-10-02 11:57:10.437    [DEBUG][10]    looking for capture source
2015-10-02 11:57:10.437    [DEBUG][10]    checking: Hauppauge HD-PVR
2015-10-02 11:57:10.437    [DEBUG][10]    enabled and present
2015-10-02 11:57:10.437    [DEBUG][10]    does not have channel
2015-10-02 11:57:10.437    [DEBUG][10]    checking: Hauppauge HD-PVR     (#2)
2015-10-02 11:57:10.437    [DEBUG][10]    enabled and present
2015-10-02 11:57:10.437    [DEBUG][10]    does not have channel
2015-10-02 11:57:10.437    [DEBUG][10]    checking: Hauppauge HD-PVR     (#3)
2015-10-02 11:57:10.437    [DEBUG][10]    enabled and present
2015-10-02 11:57:10.437    [DEBUG][10]    does not have channel
2015-10-02 11:57:10.437    [DEBUG][10]    checking: IPTV Device
2015-10-02 11:57:10.437    [DEBUG][10]    enabled and present
2015-10-02 11:57:10.437    [DEBUG][10]    does not have channel
2015-10-02 11:57:10.437    [DEBUG][10]    checking: Hauppauge HD-PVR     (#4)
2015-10-02 11:57:10.437    [DEBUG][10]    enabled and present
2015-10-02 11:57:10.437    [DEBUG][10]    [color=#FF0000]has channel[/color]
2015-10-02 11:57:10.437    [DEBUG][10]    [color=#FF0000]live tv is possible[/color]

showing recording on channel 666 being recording from capture source 29:

Code:

2015-10-02 11:57:38.437    [DEBUG][11]    allocating recording target filename: D:\(no show details)\(no show details)_20151002_11301200.ts
2015-10-02 11:57:38.437    [DEBUG][11]    [color=#FF0000]About to start recording (1394 on 29)[/color]: D:\(no show details)\(no show details)_20151002_11301200.ts...
2015-10-02 11:57:38.437    [DEBUG][11]    Free space 287290277888  (D:\)
2015-10-02 11:57:38.437    [INFO][11]    HDPVRRecorder.StartStream(D:\(no show details)\(no show details)_20151002_11301200.ts)
2015-10-02 11:57:38.437    [DEBUG][11]    About to tune graph to:
<tuning>
  <type>HDPVR</type>
  <locator>
    <channel>666</channel>
    <input>3</input>
    <audio_input>RCA Audio Input</audio_input>
    <blaster_executable></blaster_executable>
    <blaster_args></blaster_args>
  </locator>
</tuning>

showing that capture source 29 is hd-pvr #4:

Code:

2015-10-02 11:57:10.359    [DEBUG][1]    GetServerStatus() returns:
<Status>
  <Device oid="20" identifier="Hauppauge Siena Crossbar:1">
  </Device>
  <Device oid="24" identifier="Hauppauge Siena Crossbar:2">
  </Device>
  <Device oid="26" identifier="Hauppauge Siena Crossbar:3">
  </Device>
  <Device oid="23" identifier="IPTV">
  </Device>
  [color=#FF0000]<Device oid="29" identifier="Hauppauge Siena Crossbar:4">[/color]
  </Device>
</Status>

[dixxfixx]

Hi Johnson! You notice I zipped my entire logs folder to keep you from spanking me!

"what makes you think it chose one of the other devices? the log says it chooses the only one that had the channel, and it all looks consistent to me:"

These are all HDPVR2's, so you can just sit there and look at them and see which one lights up.

And yes, the logs look sexy. I went through them and did some experimentation before I admitted defeat and came begging. I think the problem may be that even though they are individually identifiable (the serial numbers are part of the enumeration, as seen in NPVR.log.2, when it scanned for devices), that ID is not carried through to the device selection process. HDPVR2 #4, being the last one discovered on the devices screen, happens to have the lowest serial number. The one with the highest serial number is being chosen to watch Zulu.

So it knows that "#4" is the only one with Zulu, but it thinks they go in order of serial number, as Windoze reports them, which does not necessarily match up.

I can, of course, just attach them in order of serial number and make NextPVR discover them and number them in that order so that it all works out, or just use ArcSoft to record single-source occasional captures of some other computer (which is what this is all about), but I think there is some underlying flaw that sub might want to look at in case it ever comes up again.

Or I could be crazy.

[johnsonx42]

Hi Johnson! You notice I zipped my entire logs folder to keep you from spanking me!

and it's good that you did, as I believe those excerpts I posted were from 3 different logs. I know I'm getting a bit of a reputation on this point, sub even had to remove one of my posts recently (and quite rightly) when I exceeded the allowed snark-threshold!

if indeed what you say is happening, that the software isn't being consistent about which hd-pvr is #4 from one part of the program to another, then your issue has risen above the level of getting help from a minion (even a self-proclaimed log-analysis-ninja minion) and salvation can only be found at the hands of the master himself.

[dixxfixx]

Ninja indeed! I quite enjoy your posts, and heartily agree with them. Long have I waited for such a champion, being too non-confrontational (read: wimpy) to do it myself. You give good help.

And yes, it seems the logical device selection does not necessarily track the physical.

[johnsonx42]

Ninja indeed! I quite enjoy your posts, and heartily agree with them. Long have I waited for such a champion, being too non-confrontational (read: wimpy) to do it myself. You give good help.

yeah, well, when you have over 6,000 posts and 7 years on the forum, you can toss your weight around a little more. sub just has to yank my chain a little every couple of years ("bad dog, bad!").

[dixxfixx]

I decided to minimize the setup to see if I could exemplify the effect in logs or databases.

Two devices, one channel each, no epg, no channel changer. Pretty much as small as I can make it.

Started up with one device, the higher of the two serial numbers. Add Channel 1. Watch the channel, record the channel, stop NPVR and save the files away.

Attach the second device, start NPVR, detect second device, add Channel 2. Watch and record both channels alternately. It always chose the second device for Channel 1, the first device for Channel 2. Stop NPVR and save the files away.

I see nothing in the logs or database files that would indicate the physical reality.

The only way I can think to show this effect is to video record the process. If I wind up doing that, I will try to get a cute model to play me so you won't have to see my awfulness.

[sub]

NextPVR effectively only knows the devices as "the first HDPVR2", "the second HDPVR2", "the third HDPVR2" and "the fourth HDPVR2". Windows generally keeps this order the same, as long as you plugin them in, configure them, then leave them plugged in. If you start unplugging them, or reconnecting them to different USB ports etc, then that order is likely to change. For users with a symmetrical setup, where all device receive the same channels, this is no big deal. With an asymmetric setup, you have to be more careful to avoid this, or ideally use a different type of device for the one that receives different channels.

[dixxfixx]

Sure, and I can deal with that since I'm a horrible old retired engineer/programmer. My only concern is that some incapable end-user (housewife, secretary, politician, etc.) will give a black eye to the best piece of software ever written because they're unable to understand the intricacies.

And since NextPVR already sees the serial numbers, it could conceivably track devices that way. Just another added level of complexity for you. What fun!

Of course, it may never come up again. Just thought it deserved mentioning.

Oh and, by the way, as usual, <donate>.

[dixxfixx]

I thought I could get away with attaching my fourth device and using ArcSoft to record from it as long as I didn't tell NextPVR about it. Not so. It was still the go-to device.

Apparently NextPVR doesn't care if they haven't been introduced, so long as it's the right type. I guess it asks window$ for a device and accepts whatever is provided regardless of whether it has seen it before.

In the attached (pointless) logs, I attached number 4 at about 17:07 and tried to watch something. NextPVR immediately jumped on the new, unknown device, because it has the lowest serial number and windows reports it first. I guess for some reason I thought NextPVR would only abuse devices it was familiar with.

So I will either not attach it unless it is about to be used or use another computer entirely. Ooh, that gives me an excuse to build yet another PVR computer!