Trying to understand GOP

[cjv]

Hi,

I am trying to figure out why Scenaraist rejects my file and states the GOP of 38 is too big, but Maestro doesn't complain at all. I do know anything over 36 is invalid.

I have been demuxing the m2v file from the VOB, and loading it up into bitrate viewer. Most NTSC hollywood movies I've tried, have a GOP of 12...thus I set 3/4 in CCE.

Recently, I got my hands on the "24" first season. Demuxed the m2v files, and they show a max GOP of 15 (in bitrate viewer). Thus, I set my GOP in CCE to 3/5. Maestro accepted my files no problem, and I encoded the first 12 or so episodes. Just for fun, I decided to learn how Scenarist works, but it flat-out rejects the 12 m2v files that Maestro accepts...stating the GOP of 38 is too big. When running it through DVD2AVI, it states 95% FILM, so I turn Force Film it, and run pulldown later.

So, I sat down with a pen and paper:
29.970 / 23.976 = 1.25 (to account for the pulldown I am doing)
1.25 * (GOP of 15) * (2 fields) = 37.5 == 38 rounded (invalid :()
1.25 * (GOP of 12) * (2 fields) = 30.0 == 30 (valid)

Thus, if my calculations are correct, then Scenarist is right. That leaves me with a few questions:
1) Are my calculations right? :)
2) Why can the original m2v from hollywood have a GOP of 15 be accepted into Scenarist, but my re-encode cannot?
3) How come Maestro accepts it?
4) I don't have a standalone player yet...is having a GOP of 38 really going to make a big difference. I hear that part of the standard is loosely enforced.

Thanks,

cjv

[auenf]

ok, i think your getting a little confused here;

first, i dont know where you are getting 1.25 from, 3/2 is 1.333333 in my books ;)

when you demux the m2v from the VOB's, it has 3:2 pulldown applied on it, and the 15 you are seeing is probably the 3:2 pulled down GOP size (which prior to pulldown was 10 frames)

10 * 3 / 2 == 15
12 * 3 / 2 == 18

now, they are 30 and 36 frames respectively, and both of which are under the 36 field GOP limit. i think when scen is checking the files, it gets to 19 frames (38 fields) and thinks, this GOP is too big, i better error now, and displays 38 fields as the reason.

just out of curiosity, you are encoding to a pre-pulldown GOP of 15;

15 * 3 / 2 == 22.5

now that is a long way above the required GOP size.

now, regarding DVDMaestro and its handling; it may accept the file into the asset bin, but will it compile with that file?

Enf...

[slk001]

You are both confused. 2:3 pulldown doesn't mean either 1.25 or 1.333. It means "three frames from two", or to "add" a frame to the video stream. The GOPs both before and after pulldown are the same - pulldown does nothing to the GOPs (however, the player adds the extra frames using the pulldown flags when it sends the video to the TV). The TV doesn't even know what a GOP is or what it does, because it cannot handle a MPEG input stream.

Most likely, your copy of Scenarist is corrupt - contact Sonic for exchange/replacement of your original disks. Both Maestro and Scenarist are strict adherents to the GOP length, and if Maestro will accept your stream and Scenarist won't, the finger is pointing at Scenarist as the "offender".

[cjv]

Yea, looks like I was pretty confused.

I was thinking to myself that since the episodes were originally made for TV, they might be encoded as interlaced and that would explain some of my problems. In any event, encoding with a CCE GOP of 3/4 (12 * 3 / 2 == 18) doesn't cause any problems with either program.

About the Sonic thing, you could be quite right. Unfortunately, I'm just a student who enjoys learning new stuff and playing around with video technology for pure entertainment. Enough said. :)
I figure if you're going to learn something, you might as well learn how to do it right. Actually, I don't even own a DVD burner or even a player!

Now I think about, I guess I should know better, than to post about a problem that could have been caused by my "verion" of the software.

Thanks for the responses,

cjv

[Atlantis]

Cjv, your calculations are correct. His 1.25 is not coming from 3/2! It’s the 3:2 pulldown factor when applied to a video to achieve the frame rate after the pulldown.

I had the same problem about 2 years ago. It’s almost a classic for people going from Maestro to Scenarist. 2 Years ago I made the definite jump to Scenarist and it’s almost 2 years that I use only Scenarist. Here I will explain some basic things acquired by experience and long days of getting errors. So don’t ask me why because I don’t know about them. I have discovered them by experience.

1) The calculation you have done is right. You should always multiple by 1.25 an NTSC film and then 2 and make sure that it is under 36.
2) So for an NTSC film with a 3:2 flag, always use 12. As you have seen 15 is too much.
3) As for your NTSC TV series and why it is in 15, my guess is that it has no 3:2 flaq and is encoded as real 29.97
4) You might ask next that why if it’s a 29.97 material, they didn’t use 18 GOP. My experience is that they never use 18 GOP for NTSC because of encoders. At least I use TMPGEnc and know about it. I have inspected the GOPs closely. When you encode and choose a GOP number it’s never strict. The encoder might produce 1 or rarely 2 GOPs more. It means when you choose 15, you might get sometimes 16! That’s why they don’t use 18! Because 19 is a no no!

I hope this answers some of your questions.

[Antonio S.]

slk001 said:You are both confused. 2:3 pulldown doesn't mean either 1.25 or 1.333. It means "three frames from two", or to "add" a frame to the video stream. The GOPs both before and after pulldown are the same - pulldown does nothing to the GOPs (however, the player adds the extra frames using the pulldown flags when it sends the video to the TV). The TV doesn't even know what a GOP is or what it does, because it cannot handle a MPEG input stream.

This is true...

You can add the pulldown "BEFORE" encoding, then you are creating additional frames (23.976 > 29.97), but if you are adding the "PULLDOWN FLAGS" after encoding you are not creating or adding any frames, just the flags or instructions so your decoder can recreate them...

Antonio S.