(UK) Ch4 AC3 filter

Begonnen von PerryM, November 27, 2017, 12:57:33

« vorheriges - nächstes »

PerryM

Anybody else in UK notice that Ch4 seems to have changed its transmission protocols on films such that the AC3 filter no longer works. It seems to be on recent films - there are some where it still works fine. Is it possible that TSDoctor is not detecting a more recent version of AC3? I have used 'Media Info' on the recordings and it doesn't detect any difference in AC3 between a recording that works and one that doesn't.

Cypheros

Could you offer a short sample of such a not working recording?

AC3 is a standard for more than 20 years. No changes known.
Or are you talking about EAC3 = DD+?
TS-Doctor should show you if it's EAC3 or AC3.

PerryM

This one doesn't work: I have saved a few seconds that covers the movie start out of the break. It is 28MB so too large to post. I have attached the Media Info details on this clip.
cheers

Cypheros

My sample looks the same:
Audio #1
ID                                       : 2306 (0x902)
Menu ID                                  : 21200 (0x52D0)
Format                                   : AC-3
Format/Info                              : Audio Coding 3
Format profile                           : Layer 2
Format settings, Endianness              : Big
Codec ID                                 : 6
Duration                                 : 2 h 0 min
Bit rate mode                            : Constant
Bit rate                                 : 384 kb/s
Maximum bit rate                         : 31.0 Mb/s
Channel(s)                               : 2 channels
Channel positions                        : Front: L R
Sampling rate                            : 48.0 kHz
Frame rate                               : 31.250 FPS (1536 SPF)
Bit depth                                : 16 bits
Compression mode                         : Lossy
Delay relative to video                  : -766 ms
Stream size                              : 330 MiB (4%)
Language                                 : English
Service kind                             : Complete Main


Do you use Windows Mediaplayer for playback?
Did you try to use other mediaplayers for playback?

PerryM

Just to be clear - the only thing that is not working is the AC3 filter for finding the Ad breaks. The film edits perfectly and plays back perfectly. Another Ch4 film recorded the same weekend works perfectly for the AC3 filter but looks the same in Media Info.
Prior to recent updates, Ch4 movies worked fine for the AC3 filter about 95%. Usually it was old movies (presumably without AC3) that didn't work. Recently it is perhaps 80% that don't work with the filter, and they are recent movies supposedly with AC3. Clearly something has changed!
It is not a big deal - I am just interested in why this should be!

Mam

I guess, what you see is just "normal"  ;D

The AC3 filter works by assuming that film has 5.1 audio, whereas the ads are 2.0 only. Therefore the filter just looks for the change in the amount of audio channels.

If you start it on an old movie which does not have 5.1 audio, it cannot tell the movie from the ads, so it looks like it did nothing.

But its totally normal that the filter cannot work under these conditions. Nothing to worry about.
Normally the TSDoc already notices on the entry scan that there are no changes and so it skips the detailed AC3 scan later on.

But of course, if you have a recording containing several movies, the filter might be fooled by the changes and manages to isolate all (newer) movies whilst skipping over older ones.

Note that there are even some stations which ALWAYS send out 5.1 audio by upscaling the given channels artificially or just transmitting silence in the additional channels. Of course the AC3 filter cannot help you here too.

Thats why the doc contains serveral methods for ad scanning. Usually one of them will be "good" for your recording situation, it might not always be the same. So experiment a bit with them and try to find one which works better with your input material.

Cypheros

OK, a missunderstanding. You mean the AC3 commercial break detection, not the playback.

Mam is right. It works great for movies with AC3 5.1 but not good for older Movies with AC3 2.0.


PerryM

So we may be back where I started surmising that Ch4 has changed its protocols. What is certain is that I have had films with AC3 5.1 where the AC3 filter didn't work. I mean that it simply wasn't available in the GUI, and if you forced it on then the ribbon turned red and (of course) it found no breaks. I have erased all the raw recordings but I'll do a bunch this weekend and get back.
Does the AC3 filter need to detect BOTH 5.1 & 2.0 content for it to become available in the GUI?
It is sort of ironic that we got the improvement in the Video Analysis mode (which now works fine on Ch4) so at least I have a back-up. I always edit manually (for accuracy) so the auto detect is just a start point.
To repeat - I am a happy camper but curious as to unexplained changes.

Mam

Zitat von: PerryM am November 29, 2017, 12:57:56
Does the AC3 filter need to detect BOTH 5.1 & 2.0 content for it to become available in the GUI?
Because it triggers on the edge of a change: OF COURSE!  ;D
In the settings you can select how many samples the quick scan will do (default 18 or so). Within these samples there needs to be at least one change, else the filter will be skipped during further analysis. If you think it misses too often (you KNOW that there are changes, but they are not discovered automatically), raise this number. But beware that more samples also will mean a longer time for the quick scan to run.

There is another option: "prefer 5.1 channels". If selected the Doc thinks that 5.1 audio parts are part of the movie, otherwise it treats them as ads and selects the parts with 2.0 audio instead. Usually you dont need to bother with this one.

PerryM

We are again rather missing my point.
When you first load a TS file in the GUI, it reads the metadata and immediately offers to detect the commercials with 'Video Analysis' or (sometimes) AC3 mode. What allows the AC3 mode to be presented? It is clearly NOT the presence of an AC3 5.1 audio track (at least, in the present version of TSDoctor).
I shall check all the settings tonight (I am at work) to see if any have changed.

Mam

Zitat von: PerryM am November 29, 2017, 19:04:53
What allows the AC3 mode to be presented?
As far as I know it MUST be the detection of a channel change in the AC3 audio track. The whole filter makes no sense without an AC3 track.
I cant get Ch4 here (or? hmm, maybe I can, have to check. But then, it would only be a valid check if I could get it from the same source as you do. Different Sats may transmit different tracks, even for the same station) therefor I cant take a look.

but maybe Cypheros knows about another reason why this filter may be selected? ("use the source Luke....!")

wolfman

Lately I had a similar effect on a recording from a german  TV Station. TSD read the file and immediately started EPG Recognition - obviously because it found no AC3 Changes respectively commercial breaks.
Since I was sure there must have been commercials I manually started AC3 recognition and tadah: 4 Changes.
Topfield SRP 2410M, SRP 2410, Zidoo X9s, Sony PS3, Amazon FireTV mit Kodi

Mam

#12
Zitat von: wolfman am November 30, 2017, 08:51:31
TSD read the file and immediately started EPG Recognition - obviously because it found no AC3 Changes respectively commercial breaks.

Wrong too ???
This behaviour is also "normal". It comes from the precedence of the filters during quick scan. Like in the old Video game "Joust" ("...the higher lance wins!") is the test sequence (which is hardcoded into the doc) important. Once a filter reports that it found the needed infos in the recordings, lower filter tests are completly skipped.
So it never got to the point where it might have discovered the AC3 stuff.

The EPG scan comes ahead of the AC3 scan, therefor your file has never been tested at all against audio changes.

The main problem here is that the sequence of tests is hardcoded. Its not transparent to the user why sometimes this filter is picked and sometimes a different one.

Users assume that it would be always "the best", but thats far from reality. If you wanted this, it would be needed that ALL filters would have to be run in full lenght and then the final results (who found more blocks?) would have been compared, finally chosing the one that produced the most detailed results. But that would take ages, so the doc does only quick scans and picks the one that first reports it might produce a result.

You will ever have situations where you need to choose manually and then run your desired scan by hand afterwards.

wolfman

Thank you, didn't know this
Or in German "wieder was gelernt"
Topfield SRP 2410M, SRP 2410, Zidoo X9s, Sony PS3, Amazon FireTV mit Kodi

Cypheros

In the log you can find a sequence of audio information:
Audio stream 1: MPEG1/Layer2 2.0 48000Hz (NAR)
AC3 2 channels: 18 times
Audio stream 2: AC3 2.0 48000Hz (ENG)
EIT data found


So TS-Doctor checked at 18 different positions of the recording for the number of channels. In this sample it's always "AC3 2 channels". No AC3 commercial scan is offered because it's unlikely that commercials can be detected with this methode.

Here is another sample but this time different numbers of channels are detected. This time TS-Doctor will offer and prever AC3 commercial detection because it's very likely, that commercials can be detected with this methode:
Audio stream 1: MPEG1/Layer2 2.0 48000Hz (DEU)
AC3 2 channels: 6 times
AC3 6 channels: 12 times
AC3 detection mode: Prefer 5.1
$0203 2 different AC3 types detected!
$0203 selecting AC3 type with mode 5.1
Audio stream 2: AC3 5.1 48000Hz (DEU)



www.cypheros.de