Programme Extraction, Transport Stream Analysis and transportx Package

Newcoppiceman

Active Member
I've e-mailed TPTV, Regency Productions and the "technical sponsor" re a bugbear of mine - jerky motion on The Heritage Chart Show:

I appreciate many of the low-budget videos featured are a bit “iffy” technically – but that clearly doesn’t apply to the beautifully-crafted Now and Then video. Again this week it suffered from the jerky-motion problem that, I believe, afflicts the entire programme, and has done for as long as I can remember.

To illustrate this problem, there’s a shot near the start of the video where the camera pans left from George to Paul, lasting 5.8 seconds. During the pan there are four instances of repeated frames which cause the pan to be jerky/hesitant.


I've never copied a programme to a USB stick, let alone tried to run it through a transport stream analyser or similar. Just wondered whether anyone had successfully done this?
 
Just wondered whether anyone had successfully done this?
What do you mean? We've done all sorts of things, it's not difficult. I'm not sure any broadcaster is going to welcome the idea that you've grabbed their output and put it under a microscope though!
 
That's a "yes" then. Nothing is difficult if you know how; I need to check what file type a recorded programme is; that it copies to a USB stick ok; and that it can be examined by a suitable software tool. Can you recommend the latter?

If the broadcaster did their job right, I wouldn't have to examine their output. The suggestion is that the programme has been made as a 24fps file and frames have been repeated to make 25fps for UK broadcast. Yuk.
 
The Humax seemingly took ages to copy the programme to a FAT32 USB stick, so I used the hidden menu to force an eject. So now I have:
1702509289854.png
Does that look right? I found this promising tool: https://www.promaxelectronics.com/i.../free-software-for-transport-stream-analysis/ ("a useful tool to understand how digital television systems work"). I've used Promax kit in the past and they're a reputable manufacturer. Needless to say, having downloaded the software it doesn't seem to recognise the TS file above. Could be very interesting if I can get this to work.
 
.TS files as recorded by the Humax are 192 byte packets not 188 byte packets, which I expect is what the software probably wants to see.
 
The suggestion is that the programme has been made as a 24fps file and frames have been repeated to make 25fps for UK broadcast. Yuk.
This happens a lot on the local channel Notts TV who seem to play out old, out of copyright, NTSC videotapes (yes I do mean videotapes - even if they have then been digitised) of films. The jerky 24, or maybe ~30, to 25 FPS conversion is dreadful.
It's not as though TPTV can't obtain conversions of video clips. Many of their old films seem to be the American release (with an unfamiliar title). They don't usually appear jerky.
 
It's not as though TPTV can't obtain conversions of video clips. Many of their old films seem to be the American release (with an unfamiliar title). They don't usually appear jerky.
The problem, I think, lies with the production company - Regency Productions. The Heritage Chart Show is presented by Mike Read whose outfit it is (and he's obviously well-connected with the Cronins who run TPTV). The show has worldwide distribution (and an American sponsor) so that might explain why it's (maybe) created 24fps. That said, back in the day, hardware made by companies like Snell & Wilcox could standards convert cleanly; now that everything is file-based I've no idea what the equivalent gold standard is. Dropping or adding frames to convert between different temporal standards is just the pits. In telecine, the practice was to run the 24fps film at 25fps, then pitch-correct the audio for critical material (film for TV was shot 25 fps).
 
The Humax seemingly took ages to copy the programme to a FAT32 USB stick
200MB/minute. See http://hummy.tv/forum/threads/black-holes-trail-guide-to-hdr-fox-t2-hacking.354/

.TS files as recorded by the Humax are 192 byte packets not 188 byte packets
M2TS rather than actually TS. You might do better with software if you rename the file .M2TS or .MTS. If the anaylyser software has to have proper TS, it's simple enough to convert.

You seem determined to beat your own path rather than stand of the shoulders of those who put in the leg work years ago.
 
For the purposes of any analysis you probably only need a short 5min segment rather than a full programme copy anyway.
I am not sure what you hope to achieve - companies are very reluctant to change procedure that, in their view, work perfectly adequately
 
I'm not sure any broadcaster is going to welcome the idea that you've grabbed their output and put it under a microscope though!

I am not sure what you hope to achieve - companies are very reluctant to change procedure that, in their view, work perfectly adequately
TPTV has been known to change procedures if someone points out a problem. (Ages ago, I complained about the loss of 0.5-1s of sound going into the break in films. That got sorted.) Jerky pictures can hardly be described as adequate. Although I have to agree that examining the .ts file isn't likely to further the evidence. The contents of post #1 should be enough. Whether they do anything about it :dunno:.
 
I have to agree that examining the .ts file isn't likely to further the evidence
Chewing this over with other members of the BBC Engineering Facebook group, it's likely we've identified the problem with the programme, namely production at 24fps and then poor conversion to 25fps. Idle curiosity led me to wonder what examining the transport stream might reveal, which is why I originally asked whether anyone had successfully done this and, if so, using what tool.

While it might not reveal anything about the issue with the programme, discovering that a tool with such excellent provenance (Promax) is available makes this especially worth pursuing and I will follow-up the helpful steers which have been provided to get the Promax software to work with the Humax file. As ever, the devil will be in the detail of this procedure.
 
Idle curiosity led me to wonder what examining the transport stream might reveal, which is why I originally asked whether anyone had successfully done this and, if so, using what tool.
Probably nothing that something like VLC or ffmeg won't already tell you. If you want to poke at things more closely, then dvbsnoop but you need to target what you are looking for and where, and you probably want to turn the .ts file into an .mpg file first. And it might not help if the problem has already been baked in upstream.
 
Well, that was easy. In light of:
and the fact that I couldn't copy the 959,520KB TS file due to a file-corrupted warning, I tried copying the programme to the USB stick again. PVR still said "Copying" after 30 minutes. Left for another 75 minutes (while I had a nap) and still saying "Copying". Force ejected as before, but this time double-clicking the new TS file - this time only 216,888KB - played the 11-minute file in Windows Media Player and it now opens ok in the Promax TS analyser software. Don't know what's going on with the copying, but I'll take that as a win.
 
You could have just downloaded it by WebIF; although it wouldn't have been quicker than a trouble-free USB transfer, it would have taken no more than about 5 mins!
 
You could have just downloaded it by WebIF; although it wouldn't have been quicker than a trouble-free USB transfer, it would have taken no more than about 5 mins!
Yep, that worked fine - complete TS file is 1.18GB. No NTS or THM file this time, though; is their purpose explained somewhere? Unfortunately Windows Media Player won't play, and the Promax TS analyser software doesn't like it.
 
Back
Top