• The forum software that supports hummy.tv has been upgraded to XenForo 2.3!

    Please bear with us as we continue to tweak things, and feel free to post any questions, issues or suggestions in the upgrade thread.

Beta My little projects

M

mumble_myopia

Member
I'm a long time user of the customised SW and there are a few improvements that would make my life easier (not a criticism of the current CF, just potential improvements). I'm happy to get my hands dirty if anyone can provide pointers in the right direction.

1) investigate wake on LAN - having a box not connected to the TV and in another room means I have to go upstairs and turn it on to access it. I realise it depends on the NW hardware in the box but it would be worth my looking at it
2) recognise overlaps causing conflict and adjust especially when recordings on same channel - because I set the beginning and end 'extra times', I often get a dialog on the screen asking me to select from a combo of recordings even if there isn't really a conflict.
3) sub-title synch issue - investigate. As I get older I find this more and more frustrating as I rely more on subtitles with the modern directorial trend of amplifying the background noise to 'improve' the character.
4) email notifications eg those that appear in the WebIf banner, 0 length recordings, alerts of longed for program appearance (retune independant) and alerts of recording clashes

I have worked on monolithic non-authored code bases in the past and in my current incarnation (17 years now) have been an embedded engineer working with ST hardware so think I have a reasonable chance of getting somewhere.

Any tips (or pointers to these issues being solved via another route) gratefuly appreciated
 
mumble_myopia said:
investigate wake on LAN - having a box not connected to the TV and in another room means I have to go upstairs and turn it on to access it. I realise it depends on the NW hardware in the box but it would be worth my looking at it
Click to expand...
The major part of the system (including anything which could receive and act on a WoL) is powered down in standby, so there's no chance without hardware mods. I have considered and dismissed it; it would require an auxiliary system to monitor network traffic and then inject a wake-up on the signal line the front panel module uses to assert a wake-up. That would require a separate network connection, and yet somehow share the same IP address...

mumble_myopia said:
recognise overlaps causing conflict and adjust especially when recordings on same channel - because I set the beginning and end 'extra times', I often get a dialog on the screen asking me to select from a combo of recordings even if there isn't really a conflict.
Click to expand...
Take a look at the existing work, schedchk comes to mind. Contribute to it by all means.

mumble_myopia said:
sub-title synch issue - investigate. As I get older I find this more and more frustrating as I rely more on subtitles with the modern directorial trend of amplifying the background noise to 'improve' the character.
Click to expand...
The cause of this will be embedded within the system-on-chip stuff. If you look up the hardware references on the wiki, you will see (links to) outline specifications for what makes the PVR functions work, and Humax have simply (like other manufacturers) taken the hardware systems provided to implement all the basic functionality and added their own front-end. If you're going to crack that one, you'll have to figure out how the system-on-chip hardware is configured, and hope it isn't because of a fundamental hardware inadequacy. All that information is kept private to industrial users: agree to buy x-thousand units of the SoC chip and you'll be provided with the data needed to use it, otherwise nada.

What "we" (not me!) have done is examine the raw machine code contained in the .hdf update files, with knowledge of the MIPS machine code structure and the elements of the code base which were open source and had to be published under the open source license requirements. That then enabled "us" to insert additional code to provide access to the operating system, and then load that as a firmware update... but only by first having cracked the security hash which marks any particular .hdf as a valid update. What we have not done is determine what the API is to the hardware (and if you can do that, you can figure out how to turn off encryption).

mumble_myopia said:
email notifications eg those that appear in the WebIf banner, 0 length recordings, alerts of longed for program appearance (retune independant) and alerts of recording clashes
Click to expand...
RS already does some of that - alerts to failed recordings, persistent text and rule-based searches...

Anything you want over the external Internet is handled by the RS server, and the WebIF "phones home" to upload status and receive commands.

mumble_myopia said:
have been an embedded engineer working with ST hardware so think I have a reasonable chance of getting somewhere.
Click to expand...
You do? Best of luck. If you do manage to crack the hardware stuff I'm sure we'll be delighted. Everything else is pretty much already covered.
 
Last edited:
mumble_myopia said:
1) investigate wake on LAN - having a box not connected to the TV and in another room means I have to go upstairs and turn it on to access it. I realise it depends on the NW hardware in the box but it would be worth my looking at it
Click to expand...
I think you will find this to be impossible, I sure it has been investigated in the past.
Networking is normally disabled when the Humax is recording scheduled programmes in the background
mumble_myopia said:
2) recognise overlaps causing conflict and adjust especially when recordings on same channel - because I set the beginning and end 'extra times', I often get a dialog on the screen asking me to select from a combo of recordings even if there isn't really a conflict.
Click to expand...
Look at my schedchk package which is available in the Advanced section of the packages catalogue. It already attempts to detect and reschedule conflicting recordings but can't spot those where there is fake conflict when the AR codes for programmes overlap, I rarely use padding - on most channels AR is pretty good.
If schedchk does not meet your needs we can discuss and it does show how to adjust the recording schedule
mumble_myopia said:
3) sub-title synch issue - investigate. As I get older I find this more and more frustrating as I rely more on subtitles with the modern directorial trend of amplifying the background noise to 'improve' the character.
Click to expand...
I am not sure whether there is anything that can be done - but I haven't looked since fortunately I don't need subtitles (yet).
mumble_myopia said:
4) email notifications eg those that appear in the WebIf banner, 0 length recordings, alerts of longed for program appearance (retune independant) and alerts of recording clashes
Click to expand...
I don't know of any fundamental reason why you couldn't email from the humax, as BH says the RS server has been used for emails until now
Rs already does
alerts of longed for program appearance (retune independant) and alerts of recording clashes
Click to expand...
 
Last edited:
thanks for that all, will digest.

wrt subtitles. I'm unconvinced it is purely down to bought in components. With some channels it is rock solid (and seems to be better on HD channels). I wonder if there are various modes to the actual data and if so whether it can be cured/improved by preprocessing whatever file(s) the subtitles are in - that was why I said investigate (there is more than one way to crack a chestnut) ;-)
 
Black Hole said:
...and (no doubt) restrictions on what you're allowed to do with it!
Click to expand...
Well, my understanding is a typical NDA says you are not allowed to distribute the information in any way shape or form. Typically any documentation supplied will be watermarked with the name of the company it was supplied to as an aid to tracking any unauthorised distribution.
 
Yes, but it might also say you agree (in blood) not to reverse-engineer anything, regardless of divulging it.
 
mumble_myopia said:
With some channels it is rock solid (and seems to be better on HD channels).
Click to expand...
Doesn't that just tell you the broadcast sources are the problem? Do you really think the Humax firmware has any part to play in stripping the subtitle stream and merging it with the video stream in real time?

The subtitle stream is video (not ASCII/Unicode), and I expect there is a hardware TS demuxer for it, a decoder, and a video output mixer. Problems will be due to "standards disagreements", and hardware units which can handle subtitles better than others will have been built to accommodate newer standards or be less picky of non-conforming streams.

I don't want to stop you having a look - you never know, you might be onto something - but I don't want you coming back in a year complaining I didn't warn you it might be a waste of time.
 
mumble_myopia said:
wrt subtitles. I'm unconvinced it is purely down to bought in components.
Click to expand...
It happens with some Humax satellite boxes too (our 1010S at least). We've dumped that as well as the FOX for the latest Freesat box (Arris) and an Aura in large part because of the subtitle issues.
That said the broadcasters still muck them up at times but at least the text and audio aren't 20 seconds apart half the time.
 
Black Hole said:
Doesn't that just tell you the broadcast sources are the problem? Do you really think the Humax firmware has any part to play in stripping the subtitle stream and merging it with the video stream in real time?

The subtitle stream is video (not ASCII/Unicode), and I expect there is a hardware TS demuxer for it, a decoder, and a video output mixer. Problems will be due to "standards disagreements", and hardware units which can handle subtitles better than others will have been built to accommodate newer standards or be less picky of non-conforming streams.

I don't want to stop you having a look - you never know, you might be onto something - but I don't want you coming back in a year complaining I didn't warn you it might be a waste of time.
Click to expand...
as I said I want to investigate it. It is not cut and dried by any means where the problem lies.
If it were simply embedded in the data stream at the relevant points then when the STs rush ahead of themselves, stopping playback and restarting wouldn't restore the positional accuracy (sometimes for a good while). I haven't yet even looked at the encoding but it has been an interesting exercise deducing what I can specifically without looking at how it is stored. The portrayal of the STs often seems unduly unsettled by the imminent arrival of adverts (a common trick for restoring them is to leap forward and then back the correct number of jumps but that often doesn't work near ads).
This is exactly the sort of challenge I love (sorry, sad isn't it?) ;-)
 
The main problem with both our boxes ('view and 'sat) was that with any prolonged lack of speech/subtitle the thing seemed to become impatient and start displaying them early. They rarely resynched without some skipping back and forth. I never noticed any correlation with adverts - that's a new one.
Incidentally, 99% of our viewing is recordings.
 
MikeSh said:
The main problem with both our boxes ('view and 'sat) was that with any prolonged lack of speech/subtitle the thing seemed to become impatient and start displaying them early. They rarely resynched without some skipping back and forth. I never noticed any correlation with adverts - that's a new one.
Click to expand...
agreed
MikeSh said:
Incidentally, 99% of our viewing is recordings.
Click to expand...
ditto


This discussion has reminded me to keep a recording that is particularly bad for testing
 
a bit of research indicates that it is in the data stream (sorry Black Hole!) but it is done as a sort of timed overlay using the old teletext system. There is some sort of time base and the overlays are shown relative to that. So it is the timebase that must be varying - which was my suspicion but assumed there was ACII text involved. Next I need to understand what causes the wandering and whether the base is start of recording or relative to a later point (virtual bookmark or adverts or something)
 
I decrypt my recordings on the HDR. If I wish to watch with synced subtitles I use VLC on another device to playback.
 
I suppose if the
bottletop said:
I decrypt my recordings on the HDR. If I wish to watch with synced subtitles I use VLC on another device to playback.
Click to expand...
one of the things that occured to me was to take a known Humax-unfriedly recording and see if it could be played back satisfactorily on VLC with STs in the right place
 
You must log in or register to reply here.
Back
Top