PGCedit Seamless layer break discussion..

[bigotti5]

And before anyone asks a seamless joint cannot follow a slideshow because of the need for a sequence_end.

I made an example in maestro, 00 01 B7 is present, Vobu_se_e_ptm is non zero but the following vob joint seems to be seamless if I compare audio PTS values


http://rapidshare.de/files/19500930/...b_Mux.rar.html

[mpucoder]

That's not what I see. Both vob 2 and vob 3 (there are two angles) do not have vobu_se_e_ptm set in the last vobu, and the sequence_end was replaced with 00's. Besides, this is not what Scenarist calls a slideshow (I know, the term is not very specific). I was referring to a single I picture extended in time, as in my 7 color subpicture example.

edit: Maybe you were talking about vob 1? It doesn't look very spec, I'll see what the verifier says. For a vob prior to angles the last 3 or 4 vobu's should be marked as PREU, but they are not. The mux rate is OK, though.

[mpucoder]

Wow, lots of errors, including

>>> [DVD] ERROR 4621  (ref. DVD-3 4.5.2 (1)) :
    SML_PBI : VOB is allocated in a Contiguous Block and connected seamlessly
	with the next VOB in an Interleaved Block, and the former VOB is not defined as PREU.
for DSI unit 10 at byte 48 bit 0;
  PES stream-byte 20148 (byte 1072 of packet 19);
    PS stream byte 198704 (byte 2096 of pack 96).

and

>>> [DVD] ERROR 4644  (ref. DVD-3 5.1.1) :
    SML_PBI : The VOB_V_E_PTM value 1033200 (11.480 seconds) must be the
	same as the VOBU_E_PTM 475200 (5.280 seconds) of the last VOBU.
for DSI unit 9 at byte 48 bit 0;
  PES stream-byte 19131 (byte 55 of packet 19);
    PS stream byte 183351 (byte 1079 of pack 89).

The verifer does not check for a still before a seamless joint, but in my tests and research I discovered it does not work properly. However, we cannot discount the fact that many authoring programs (MuxMan included) will allow a still before a seamless joint.
This example brings up another subject - interleaving and the layer break. I believe the original intent of the rule about "seamlessly linked" was referring to interleaving. And it would be asking a lot of a player to not only refocus but start de-interleaving at the same time. It is my opinion that LB should not be allowed if either vobu is marked as interleaved (offset 0x427 non-zero)

[bigotti5]

Maybe you were talking about vob 1?

Yes, talking about vob1

Besides, this is not what Scenarist calls a slideshow

Why not?
Cant see the difference..

Interra tells

Error Name: N-DSI-2-040 Severity: Minor Error in: Data Search Information(DSI)
Error: Error in DSI playback information
Location:
Description: VOB termination time is not same as presentation termination time of last Video frame of the last GOP.It should be 475200 which is VOBU_E_PTM in last VOBU but it is 1033200
Specification Page Number: V14-139

just an example what is possible with an approved app as spruce

[mpucoder]

That was in reference to vobs 2 & 3, vob 1 is indeed a slideshow.
I'm surprised that is considered a minor error, maybe based on how badly it affects players. It definitely causes problems with several demuxers (PgcDemux, DVDSubEdit, the internal demuxing/positioning of MuxMan).

But back to seamless LB - A very strong warning should be given if the vobu either before or after a proposed LB is interleaved.

[jeanl]

mpucoder, this is a very interesting thread, especially the part about the seamless playback flag in the IFO.
Now, this brings another question: How is this flag used by the player???
From what you say, the player can play a seamless transition even if the flag isn't set, but setting the flag is the joint isn't muxed seamlessly can crash some players...
So how do players use the flag?

jeanl

[mpucoder]

Looks to me like the flag tells the player how to play. If it's set the player attempts seamless playback, if it's clear the player lets all buffers run out and checks for a cell command.

[jeanl]

Ok, that makes sense. So if I understand you correctly, if the flag isn't set, even though the joint is muxed seamlessly, you will still have a pause in the playback.
In other words, to get a seamless playback you need both a seamless mux and a set flag.
Jeanl

[jamos]

Originally Posted by jeanl

Ok, that makes sense. So if I understand you correctly, if the flag isn't set, even though the joint is muxed seamlessly, you will still have a pause in the playback.
In other words, to get a seamless playback you need both a seamless mux and a set flag.
Jeanl

I have normal mux and i set the seamless flag option in PGCedit, no pause on break. Some players (older ones) may have issues with small buffers running out of memory and you would get a pause (as with non seamless break), but most (if not all) newer ones will not have issues. In PGCedit we are setting the layer break flag and we are just leaving the joint flag seamless (either 8 or 10). Note: I am only talking about seamless cells here not interleave etc.

and yes even if you have seamless mux and the layer break flag is set normally it is a non seamless break and you will get a pause.

[r0lZ]

Originally Posted by mpucoder

A very strong warning should be given if the vobu either before or after a proposed LB is interleaved.

OK, I have added a warning, and the pink background in this case. Currently, this warning is issued if the First ILVU End value in the cell playback info table is non-zero for the current cell or the previous one. Is this test sufficient?

Note that, currently, it is not possible to select an angle cell for the LB. It's mainly because I don't know if the seamless flag must be cleared on all cells of the angle, or only on the selected cell. I suppose it's the first case, but I don't have the logic to do that easily. Therfore, it's forbidden.

However, it is possible to set the LB on the cell that follow an ILV or angle cell. Your advice is therefore useful, mpucoder. Thanks again!

[blutach]

Originally Posted by mpucoder

Looks to me like the flag tells the player how to play. If it's set the player attempts seamless playback, if it's clear the player lets all buffers run out and checks for a cell command.

And, I guess, to ensure the buffers are all clear, every now and again, we see the NOP cell command at the LB.

Regards

[mpucoder]

Originally Posted by r0lZ

OK, I have added a warning, and the pink background in this case. Currently, this warning is issued if the First ILVU End value in the cell playback info table is non-zero for the current cell or the previous one. Is this test sufficient?

As long as you check all PGCs I'd say it is, as even single angle multi-story, which uses cell types 12 and 14 (interleaved but not an angle block), sets this value. Possibly the "interleaved" flag in the cell category would be easier to check?

[r0lZ]

Easier, no, but it's also easy. I can even check both the First ILVU End and the interleaved flag, but, since the user can change it in PgcEdit, I prefer to trust the ILVU End!

[jamos]

Hey Rolz, not sure or not if this is intentional but if I have the layer break set on the title as non seamless IE 0 and the non seamless option turned on, it still burns non seamless...the seamless option does not override, hence to get a seamless break I would have to manually take out the layer break flag then burn. Is that intentional? If so can you disable the seamless option on burn if I have a non seamless break checked?

thanks

[r0lZ]

Yes, it's intentional. It's what is described in the text beside the option (at least in the current version.)
If I set the seamless falg, I have to test if the current VOB is really seamless, and can really be played seamlessly. (For exemple, I must also test if there is a cell command.)
A cell with the seamless cell clear appear in blue in the list. You can set the layer break manually on such cell with the "Seamless Cell" button, below the new option... but at your own risk.

I agree that it's not very intuitive, and maybe I'll change that in a future release (or beta?)

[Jenny Davis]

Hi guys

A lot of the previous discussion was over my head, but I think I understood that there are two mux rates and it is necessary to have the low mux rate going into a layer break. I would like to incorporate this on the DVD I am authoring but I also want to stay with DVD-lab pro and Gear Pro ME as my main authoring and mastering tools.

Is there a program that will let me set the low mux rate and also work with these two programs?

Will muxman do this?

Will making it multi angle just before the layer break force the lower mux rate?

Thanks for any replies.

Jenny

[mpucoder]

It is not absolutely necessary to lower the mux rate, but this is what Superbit does.
MuxMan is an authoring program, as is DVD Lab, one cannot do half the work for the other. The mux rate is determined by the multiplexer, which is at the core of every authoring program. To stay with DVD Lab it would need to allow lowering the mux rate without interleaving (the usual reason to lower it).
The current versions of MuxMan do not allow lowering the mux rate, that will be available in a future release.

[bigotti5]

...but this is what Superbit does.

Not always - I have a Superbit here with seamless layerchange at high mux rate

[mpucoder]

How about that - so it looks like the mux rate is really not very important at all.