Joining m2ts files

[cheerful]

Let me try to clarify

3 m4v files and 1 ac3 forms a timeline in Encore, none requiring transcoding. Encore builds the Blu-ray folder and creates 3 m2ts files, essentially mux the video and audio and split into 3 m2ts.

So I'd conclude

1. audio spec are identical.
2. all 3 m2ts files meet Blu-ray spec. I can verify them but create timeline for each of the m2ts file and burn another BD disc to play back

So if I have 3 m2ts file, each meet Blu-ray spec. I'd want to join them to create a single one that still meet Blu-ray spec.

From what I can understand, you are saying some of the spec must be different. So my question is then how can I tell the difference? Then I can re-encode the middle segment to correct its spec.

Appreciate your help

[cheerful]

I just noticed the 3rd m2ts file has "Format settings, GOP : M=3, N=18" while as the second one has "Format settings, GOP : M=3, N=29". I guess this means GOP mismatched.

I can't find GOP settings if I use Adobe Media Encoder under format H.264 Blu-ray. They are only available for MPEG2 Blu-ray.

How can I control GOP settings?

Thanks

[theewizard]

I'm not sure what you are going to use to do it
But I'm saying you have to create 3 mv4 files that match spec so they can be joined into one m2ts file in encore
Or create the three files and join them before putting them in encore
But the specs have to be matched BEFORE they can be joined
This means at least Two have to be transcoded not just the middle one since number one shows no Gop data
I would transcode one and two , to match number three
There is no magic quick easy change into setting way of doing this

Or put all three thru some encoder so they all get coded to the exact same spec

[poisondeathray]

I think Adobe Encore is quite picky.

Picky compare to what? If it were "picky" or more robust in it's checking you wouldn't have this problem in the first place. Didn't you say there was a "glitch" between 1 & 2? It should have flagged one or more of the original videos, or re-encoded at least one of them. It's known to be buggy - producing discs that fail professional verifiers. And it can't produce a true BD BDCMF image . Not picky in terms of professional BD software.

The GOP difference is a "red herring" . You can have variable length GOP's on BD. Perfectly legal. For 1920x1080i29.97 , all that is required is <30 , max 3 b-frames

m4v produced by Adobe is an elementary stream. The confusion lies in that "m4v" is also used by Apple to indicate a modified mp4 container which can hold AC3 audio


What command did you use in dos when trying to join the m4v segments ?

In tsmuxer, try it again with 1.10.6 , and in general track options, uncheckmark "add picture timing info", uncheckmark "continually insert sps/pps"

3 individual compliant files are not necessarily compliant when appended. I mentioned the most common reason why earlier. There is no way to ensure VBV compliance doing it individually. You should have produced it correctly as 1 video in the first place. So even if you get it working - it might not be "fully" compliant, and fail to play properly in some players, especially early generation models

[theewizard]

Can you have Gop differences inside the same file?
Inside the same movie?
On the disc with different files understand ,
But joining files with different Gop values?

[poisondeathray]

Can you have Gop differences inside the same file?
Inside the same movie?
On the disc with different files understand ,
But joining files with different Gop values?

Yes, the only restriction are the ones I listed above

Joining files can be done with different GOP characteristics (in general) , if you have closed GOPs . By that I mean different GOP lengths, I,B,P makeup. But for BD, the potential issue is a VBV violation, and you have to satisfy the 2 conditions I listed earlier

[theewizard]

thanks
what should he use to transcode the three segments into three matching segments, can that be done ?
what software to use ?
and what settings ? what would be the most common or the easiest to achieve ?

[cheerful]

The GOP difference is a "red herring" . You can have variable length GOP's on BD. Perfectly legal. For 1920x1080i29.97 , all that is required is <30 , max 3 b-frames

What command did you use in dos when trying to join the m4v segments ?

In tsmuxer, try it again with 1.10.6 , and in general track options, uncheckmark "add picture timing info", uncheckmark "continually insert sps/pps"

What should be < 30?

Just tried with the latest tsMuxer with the following meta file

MUXOPT --no-pcr-on-video-pid --new-audio-pes --vbr --vbv-len=500
V_MPEG4/ISO/AVC, "A.m4v"+"B.m4v"+"C.m4v"
A_AC3, "audio.ac3"

The output is playable by software (no mosaic or jittery around B segment). However, Encore still require transcoding.

Do I need to use older version of tsMuxer?

These are encoded with Adobe Media Encoder, NTSC, 1080i, VBR 2-pass. I can re-encode segment B and accept the loss of quality in a short segment. Which encoder should I try and what setting (note I can't find GOP settings with Adobe Media Encoder with H.264 Blu-ray format)?

Thanks

[poisondeathray]

thanks
what should he use to transcode the three segments into three matching segments, can that be done ?
what software to use ?
and what settings ? what would be the most common or the easiest to achieve ?

Not "into" three matching segments - they supposedly already match. If you were re-encoding, you would encode the whole thing as 1 (what he should have done in the first place)

Encore might be forcing a re-encode because alone, they are compliant, but together they are not (the most common would be a VBV violation, but Encore should have told him there was a problem the 1st time he tried to make a disc. He said there was a pause or problem during playback at the spot between 1 & 2.

What should be < 30?

Max GOP length. The maximum keyframe to keyframe interval. You can have it lower than than, but not higher. It can be variable (ie. some might be 1, some might be 15, some might be 16)

Just tried with the latest tsMuxer with the following meta file

MUXOPT --no-pcr-on-video-pid --new-audio-pes --vbr --vbv-len=500
V_MPEG4/ISO/AVC, "A.m4v"+"B.m4v"+"C.m4v"
A_AC3, "audio.ac3"

The output is playable by software (no mosaic or jittery around B segment). However, Encore still require transcoding.

Do I need to use older version of tsMuxer?

Try it with the older version (it works better) , and without audio. Then demux the joined video for importing into Encore. Add the audio as a separate track in encore

These are encoded with Adobe Media Encoder, NTSC, 1080i, VBR 2-pass. I can re-encode segment B and accept the loss of quality in a short segment. Which encoder should I try and what setting (note I can't find GOP settings with Adobe Media Encoder with H.264 Blu-ray format)?

The GOP settings are not relevant - as long as you use a blu-ray compatible profile

Which video is which ? A,B,C - you have 12MiB, 3.45GiB, 8.04GiB listed

[cheerful]

Just tried 1.10.6 as instructed, same as the latest (2.6.12).

The project originally has 1 m4v and 1 ac3 files, produced by AME (a much older version a few years back). After most of the source videos are deleted, it's found out that one segment (B) has the source video offline at the time of encoding. The goal here is then to replace this particular segment in this project. tsMuxer was used to cut out segment A and C. AME is used to produce B.

[poisondeathray]

You just going to have to try a bunch of things out , trial & error. You've already tried AME on the middle section (is that the 12Mib section?), I doubt even if you used the older AME version that it would work. You're probably going to have to re-encode everything. If the tsmuxer output in a software player was jittery at the seam, then at least that's an indication of what to look for or where to look, or what to try to "fix" .

To be clear, you're starting a completely new project in encore, not trying to edit an old existing project with existing clip references, correct ? As soon as you import the elementary stream it wants to transcode under the blu-ray status ? (and make sure you're looking at blu-ray not the DVD column. Don't laugh - people have made that mistake before)

[cheerful]

Yes, brand new project with no reference to anything in the old project. The Blu-ray status will show "untranscoded" once the import is completed.

My goal is not transcode segment A and C (the quality degradation is obvious). So I'm thinking some kind of transcode of B. But I don't have much idea on what can be tried

[cheerful]

Just to report what I have already tried with the middle segment B

tsMuxer to extract 264 from each m4v and copy /b to combine them. Import to Encore: untranscoded

AME to produce MPEG2 Blu-ray, transcode in VideoRedo to H.264 and then join with the others, Encore: ultranscoded

[poisondeathray]

Just to report what I have already tried with the middle segment B

tsMuxer to extract 264 from each m4v and copy /b to combine them. Import to Encore: untranscoded

AME to produce MPEG2 Blu-ray, transcode in VideoRedo to H.264 and then join with the others, Encore: ultranscoded

you wouldn't be able to use mpeg2 blu-ray 100% for sure, because the other A,C are h.264 . videoredo v4 supports h.264, but you don't want it to re-encode even a few frames

the highest probability of success would be using the old AME version with the same settings, but even that has low chance of working

the other things you can try are using x264 and setting --sps-id to a different value other than 0 (like 2) , and using BD compatible settings - but you're not really having a problem with appending (the file plays ok when joined by tsmuxer), so I doubt this will increase your chances. If the file was glitched in tsmuxer, this method would make it playable without glitches, but not necessarily pass thru in encore

[cheerful]

X264 is from VideoLan, right? What would be a good GUI for Windows?

Thanks

[poisondeathray]

Can you join A+C (omit B for now) , import that into Encore, what happens ? How do you know "B" is the problem for certain ? If that test fails, anything else you do short of re-encoding everything is going to fail and you're wasting your time

[cheerful]

Can you join A+C (omit B for now) , import that into Encore, what happens ? How do you know "B" is the problem for certain ? If that test fails, anything else you do short of re-encoding everything is going to fail and you're wasting your time

Very good point. As it turns out, tsMuxer A+C still fails Encore test.

[poisondeathray]

What about dos A+C, on elementary stream ?

[cheerful]

Update.

It appears that I have messed up the individual m2ts files. So I just recreated them.

Encore to import 1 m4v + 1 ac3, build Blu-ray folder, get the whole m2ts file

VideoRedo to cut out segment C, export directly to m2ts, Encore accepted it without transcoding

VideoRedo to cut out segment A, can't save to m2ts (VideoRedo re-encodes), had to save to ts first and remux in tsMuxer, Encore requires transcoding. So I used tsMuxer to cut out A. Encore accepted it without transcoding.

Re-test join of A/B/C by VideoRedo, tsMuxer, SolveigMM. All require transcoding in Encore.

Then I tried to join these m2ts with tsMuxer

A + C, accepted by Encore without transcoding.
A + B, Encore requires transcoding
B + C, Encore requires transcoding

So the issue appears with the recoded segment B

[poisondeathray]

And was this "new" replacement B is from the newer AME version ?

Do you have the original single m4v that has the "bad" middle offline segment ?

What happens if you join the A + (old B) + C ? The old B being the "bad" offline video segment . ie just cut up your original single m4v and rejoin then import into encore (I know, stupid, but just making some observations if it really is the new B compression)

When you cut in videoredo, you have to cut on keyframes, otherwise it will re-encode a few frames. It's safer for what you are doing to cut with a GOP level cutter (for example tsmuxer)

[cheerful]

The new B is from new AME.

Originally, A/B is NOT cut at i-frame where B/C is cut at i-frame.

I just took the original m4v and ac3, used Encore to build, saved the m2ts in the output folder. I then used VideoRedo to cut at keyframe (I-frame) to get the new A and B segments.

tsMuxer to join original A, B and C (without insertSEI and contSPS) back into a single m2ts. Encore accepted it without transcoding

Adobe Media Encoder to encode B with fixed clip, produced B1.m4v and B1.ac3, mux with tsMuxer. Encore accepted it without transcoding

tsMuxer to join A, B1, C (without insertSEI and contSPS). Encore requires transcoding.

In tsMuxer, I can see the following output when I did A + B/B1 + C. Does it say anything useful?

Decoding H264 stream (track 1): Profile: High@4.1 Resolution: 1920:1080i Frame rate: 29.97
H.264 muxing fps is not set. Get fps from stream. Value: 29.97
B-pyramid level 2 detected. Shift DTS to 3 frames
Decoding AC3 stream (track 2): Bitrate: 192Kbps Sample Rate: 48KHz Channels: 2

[poisondeathray]

That log is normal for tsmuxer, nothing additionally useful

Post mediainfo (view=>text) of B0 (original), B1 (new AME encode)

[ron spencer]

export a new file of all 3 joined as DNxHD, then reencode

or read this:

http://avchdvideos.blogspot.ca/p/joining-my-avchd-videos.html

[cheerful]

export a new file of all 3 joined as DNxHD, then reencode

or read this:

http://avchdvideos.blogspot.ca/p/joining-my-avchd-videos.html

I don't want to re-encode. Otherwise I would have used AME.

[poisondeathray]

And if you want a more detailed analysis of the differences between the "two b's" beyond mediainfo, the only free tool is h264_parse . Otherwise there are various stream analyzers, some very expensive

[cheerful]

That log is normal for tsmuxer, nothing additionally useful

Post mediainfo (view=>text) of B0 (original), B1 (new AME encode)

See below.

I've noticed that segment B has very low bitrate b/c offline clip becomes a static image in the output. So I lower the bitrate to nearly match it. Yet the joined m2ts file still requires transcoding in Encore.

B

Video
ID : 4113 (0x1011)
Menu ID : 1 (0x1)
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L4.1
Format settings, CABAC : Yes
Format settings, ReFrames : 4 frames
Format settings, GOP : M=3, N=18
Codec ID : 27
Duration : 25s 158ms
Bit rate mode : Variable
Bit rate : 595 Kbps
Maximum bit rate : 35.0 Mbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate : 29.970 (30000/1001) fps
Standard : NTSC
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan type, store method : Separated fields
Scan order : Top Field First
Bits/(Pixel*Frame) : 0.010
Stream size : 1.79 MiB (72%)
Color range : Limited
Color primaries : BT.709
Transfer characteristics : BT.709
Matrix coefficients : BT.709

B1
Video
ID : 4113 (0x1011)
Menu ID : 1 (0x1)
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L4.1
Format settings, CABAC : Yes
Format settings, ReFrames : 4 frames
Format settings, GOP : M=3, N=29
Codec ID : 27
Duration : 25s 158ms
Bit rate mode : Variable
Bit rate : 580 Kbps
Maximum bit rate : 17.0 Mbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate : 29.970 (30000/1001) fps
Standard : NTSC
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan type, store method : Separated fields
Scan order : Top Field First
Bits/(Pixel*Frame) : 0.009
Stream size : 1.74 MiB (78%)
Color range : Limited
Color primaries : BT.709
Transfer characteristics : BT.709
Matrix coefficients : BT.709

[ron spencer]

export a new file of all 3 joined as DNxHD, then reencode

or read this:

http://avchdvideos.blogspot.ca/p/joining-my-avchd-videos.html

I don't want to re-encode. Otherwise I would have used AME.

DNxHD is virtually lossless...what are you trying to accomplish. Just joining?? If your clips are different it will not work without an eventual re-encode. Simple as that. Try the link I sent.

This is not like DV where you can join any output of DV. You could be doing this for years.

[cheerful]

export a new file of all 3 joined as DNxHD, then reencode

or read this:

http://avchdvideos.blogspot.ca/p/joining-my-avchd-videos.html

I don't want to re-encode. Otherwise I would have used AME.

DNxHD is virtually lossless...what are you trying to accomplish. Just joining?? If your clips are different it will not work without an eventual re-encode. Simple as that. Try the link I sent.

This is not like DV where you can join any output of DV. You could be doing this for years.

The description says "Avid Codecs contains Avid DV, DV100, DNxHD codecs. Avid HD intermediate codec (DNxHD) is useful if you want convert HDV/AVCHD to a format/codec that is supported in more editors(a mov with Avid DnxHD). Use the codecs in Quicktime Pro, Avid, Vegas Movie Studio, Vegas for example."

I already have Premiere which support AVCHD natively. I need m2ts in the end for authoring with Encore. What would this tool do to help?

The tutorial is about joining which I have already tried. Is there anything special I'm missing?

Thanks

[ron spencer]

My point is that your files are not DV. DV was great in that no matter what you shot, as long as it was DV you could join no matter what. With H264, this is not the case because of the flexibility provided in encoding parameters. In your case, you have a clip that is different than the others, which is very problematic, as you are finding out. For the time and effort spent, re-encoding is not a bad option, or moving to an archival format like DNxHD and then encoding with MeGUI or some other encoder. This has happened to me when a Canon camcorder I was using had its settings changed. I had no choice but to just archive to DNxHD and encode later...it is not so bad as the bitrate for DNxHD is huge, and color can be 10 bit as well.

[cheerful]

My point is that your files are not DV. DV was great in that no matter what you shot, as long as it was DV you could join no matter what. With H264, this is not the case because of the flexibility provided in encoding parameters. In your case, you have a clip that is different than the others, which is very problematic, as you are finding out. For the time and effort spent, re-encoding is not a bad option, or moving to an archival format like DNxHD and then encoding with MeGUI or some other encoder. This has happened to me when a Canon camcorder I was using had its settings changed. I had no choice but to just archive to DNxHD and encode later...it is not so bad as the bitrate for DNxHD is huge, and color can be 10 bit as well.

Appreciate your input but it does not help me to resolve my problem.