3D Support Discussion

[mparade]

I am quite a new user to BD-rebuilder and find something for which I would need your help:

My config file:

[Options]
VERSION=0.44.0.17
MODE=3
ENCODE_QUALITY=3
ONEPASS_ENCODING=0
AUTO_QUALITY=0
SUPPRESS_FF_WARNING=1
AUTO_BURN=2
MOVIEONLY_TYPE=0
ALTCRF=23
ALT_TARGET=1024
ALTMETHOD=2
ALTAUTOCROP=0
TARGET_SIZE=14000
AUDIO_TO_KEEP=eng;hun;
SUBS_TO_KEEP=all
SD_CONVERT=0
OPEN_GOP=1
RESIZE_1080=0
RESIZE_1440=0
RESIZE_720=0
DEINTERLACE=1
SD_TO_1080=0
CONVERT_WIDE=0
DTS_REENCODE=0
AC3_REENCODE=0
AC3_640=1
AC3_192=0
KEEP_HD_AUDIO=1
AVCHD=1
REMOVE_WORKFILES=0
MOVIE_ONLY_LOOP=1
REMOVE_OUTPUT=0
USE_FILTERS=0
BDMV_CERT_ONLY=0
USE_LAVF=0
IVTC_PULLDOWN=0
ASSUME_DVD_PAL=0
UNMASK_CHAPTER=0
COMPLETION_BEEP=0
DGDECNV=0
OUTPUT_3D=1
NEROAAC=0
SUPTITLE=0
AUDIO_TRACK_LIMIT=1
SUBTITLE_TRACK_LIMIT=0
CUSTOM_TARGET_SIZE=17000
AUTO_BIAS=3
B_PYRAMID=1
EXTENDED_GOP=1
FORCED_SUB_LANG=hun
FULL_3D=1
MAXIMIZE_REFS=1
MULTIPROCESS=1
PASS_TWO_STATS_UPDATE=1
WEIGHTP=2
[Paths]
SOURCE_PATH=F:\

The avs file of the main video (for VID_00063.m2ts) from my "workfiles" directory includes:

#Created by BD Rebuilder - v0.44.17 (beta)
LoadPlugin("E:\Users\X\Desktop\MVCTOAVI.EXE\DirectShowMVCSource.dll")
VIEW1=DirectShowMVCSource("F:\BDMV\STREAM\SSIF\00063.SSIF", framecount=164209)
VIEW2=DirectShowMVCSource("F:\BDMV\STREAM\SSIF\00063.SSIF", framecount=164209,decodeleft=true)
StackHorizontal(VIEW2,VIEW1).BicubicResize(1920,1080)
ConvertToYV12().AssumeFPS(24000,1001)

My problem is that I would like to make a Full 3D SBS movie- only backup conversion with BD compliant structure without any resize of the frames that is why the hidden parameter "FULL_3D=1" used in my config file, but inspite of giving this parameter the resize of the frames happen.

As I read the earlier threads keeping the 3D movie in full 3D format could be possible to do just for alternate mkv format movie-only outputs with parameter of NO_RESIZE. But I do not really want to make an mkv format file for full 3D movies.

Please help what I have missed out.

I very much appreciate your help.

[jdobbs]

I am quite a new user to BD-rebuilder and find something for which I would need your help:

...
My problem is that I would like to make a Full 3D SBS movie- only backup conversion with BD compliant structure without any resize of the frames that is why the hidden parameter "FULL_3D=1" used in my config file, but inspite of giving this parameter the resize of the frames happen.

As I read the earlier threads keeping the 3D movie in full 3D format could be possible to do just for alternate mkv format movie-only outputs with parameter of NO_RESIZE. But I do not really want to make an mkv format file for full 3D movies.

Please help what I have missed out.

I very much appreciate your help.Either I'm not understanding what you're saying, or your post contradicts itself. You can't to a "3D SBS" movie without resizing. That's impossible. SBS means "side-by-side".

[HWK]

I am quite a new user to BD-rebuilder and find something for which I would need your help:

My config file:

[Options]
VERSION=0.44.0.17
MODE=3
ENCODE_QUALITY=3
ONEPASS_ENCODING=0
AUTO_QUALITY=0
SUPPRESS_FF_WARNING=1
AUTO_BURN=2
MOVIEONLY_TYPE=0
ALTCRF=23
ALT_TARGET=1024
ALTMETHOD=2
ALTAUTOCROP=0
TARGET_SIZE=14000
AUDIO_TO_KEEP=eng;hun;
SUBS_TO_KEEP=all
SD_CONVERT=0
OPEN_GOP=1
RESIZE_1080=0
RESIZE_1440=0
RESIZE_720=0
DEINTERLACE=1
SD_TO_1080=0
CONVERT_WIDE=0
DTS_REENCODE=0
AC3_REENCODE=0
AC3_640=1
AC3_192=0
KEEP_HD_AUDIO=1
AVCHD=1
REMOVE_WORKFILES=0
MOVIE_ONLY_LOOP=1
REMOVE_OUTPUT=0
USE_FILTERS=0
BDMV_CERT_ONLY=0
USE_LAVF=0
IVTC_PULLDOWN=0
ASSUME_DVD_PAL=0
UNMASK_CHAPTER=0
COMPLETION_BEEP=0
DGDECNV=0
OUTPUT_3D=1
NEROAAC=0
SUPTITLE=0
AUDIO_TRACK_LIMIT=1
SUBTITLE_TRACK_LIMIT=0
CUSTOM_TARGET_SIZE=17000
AUTO_BIAS=3
B_PYRAMID=1
EXTENDED_GOP=1
FORCED_SUB_LANG=hun
FULL_3D=1
MAXIMIZE_REFS=1
MULTIPROCESS=1
PASS_TWO_STATS_UPDATE=1
WEIGHTP=2
[Paths]
SOURCE_PATH=F:\

The avs file of the main video (for VID_00063.m2ts) from my "workfiles" directory includes:

#Created by BD Rebuilder - v0.44.17 (beta)
LoadPlugin("E:\Users\X\Desktop\MVCTOAVI.EXE\DirectShowMVCSource.dll")
VIEW1=DirectShowMVCSource("F:\BDMV\STREAM\SSIF\00063.SSIF", framecount=164209)
VIEW2=DirectShowMVCSource("F:\BDMV\STREAM\SSIF\00063.SSIF", framecount=164209,decodeleft=true)
StackHorizontal(VIEW2,VIEW1).BicubicResize(1920,1080)
ConvertToYV12().AssumeFPS(24000,1001)

My problem is that I would like to make a Full 3D SBS movie- only backup conversion with BD compliant structure without any resize of the frames that is why the hidden parameter "FULL_3D=1" used in my config file, but inspite of giving this parameter the resize of the frames happen.

As I read the earlier threads keeping the 3D movie in full 3D format could be possible to do just for alternate mkv format movie-only outputs with parameter of NO_RESIZE. But I do not really want to make an mkv format file for full 3D movies.

Please help what I have missed out.

I very much appreciate your help.

Keeping full 3D is possible, however it won't BD compliant structure anymore there is nothing you can do about it at least in bd-rebuilder to get Full 3D and at the same time do not resize output. You can only choose one option either have full BD3D or have BD compliant structure.

[mparade]

Either I'm not understanding what you're saying, or your post contradicts itself. You can't to a "3D SBS" movie without resizing. That's impossible. SBS means "side-by-side".

I think maybe both of them, sorry for that.

My aim is to reencode the main movie part of 3D Blue Ray Discs, besides not reducing the number of pixels of the original material by using any of the resizing methods. So, just to keep the resolution of the original disc + reencoding with BD rebuilder with the proper setup. I do not really want to halve the resolution for each eye to receive by using either O/U or SBS converting methods. If I am right, I can do this with your software for an mkv alternate output.

Thank you very much for your efforts in advance.

[HWK]

I think maybe both of them, sorry for that.

My aim is to reencode the main movie part of 3D Blue Ray Discs, besides not reducing the number of pixels of the original material by using any of the resizing methods. So, just to keep the resolution of the original disc + reencoding with BD rebuilder with the proper setup. I do not really want to halve the resolution for each eye to receive by using either O/U or SBS converting methods. If I am right, I can do this with your software for an mkv alternate output.

Thank you very much for your efforts in advance.

oh, I see what you are saying you want full resolution of each frame with no resize, you can do that in side by side as well but it would be something like 3840*1080 pixels this include full view for both eyes and then joined together to create one large view. Be advised this method might not work on all players.

[jdobbs]

oh, I see what you are saying you want full resolution of each frame with no resize, you can do that in side by side as well but it would be something like 3840*1080 pixels this include full view for both eyes and then joined together to create one large view. Be advised this method might not work on all players.It won't work on any player that is built to BD specs.

[HWK]

It won't work on any player that is built to BD specs.

Quite true, this method is suitable for storage purpose for time being. However videofan3d has released mvc encoder which run on Intel media SDK and it can use full SBS for input and create view of 1920*1080 for each respective view.

[jdobbs]

I think maybe both of them, sorry for that.

My aim is to reencode the main movie part of 3D Blue Ray Discs, besides not reducing the number of pixels of the original material by using any of the resizing methods. So, just to keep the resolution of the original disc + reencoding with BD rebuilder with the proper setup. I do not really want to halve the resolution for each eye to receive by using either O/U or SBS converting methods. If I am right, I can do this with your software for an mkv alternate output.

Thank you very much for your efforts in advance.The problem would be in finding a player to support it. It is, though, a good way to keep a full resolution backup to be used later when full BD compliant 3D is supported via the new TSMUXER. I'll have to think about that one and give you an answer later. Right now I'm trying to debug another issue...

[Edit] Just read HWKs response, which kinda' says something similar.

[mparade]

The problem would be in finding a player to support it. It is, though, a good way to keep a full resolution backup to be used later when full BD compliant 3D is supported via the new TSMUXER. I'll have to think about that one and give you an answer later. Right now I'm trying to debug another issue...

[Edit] Just read HWKs response, which kinda' says something similar.

Ok. Until then I have just sent a donation of 50$ to solve my problem within your program asap. :) Anyway, the newest network media tanks are compatible with Full 3D contents. (The ones with using HDMI 1.4 standard for example Popcorn hour A-400). But I think all the ones are compatible with such medias which are fulfill standard HDMI 1.4 standard for the much higher data rate.) But please advise, I am not on the top of the theme. :)

[HWK]

Ok. Until then I have just sent a donation of 50$ to solve my problem within your program asap. :) Anyway, the newest network media tanks are compatible with Full 3D contents. (The ones with using HDMI 1.4 standard for example Popcorn hour A-400). But I think all the ones are compatible with such medias which are fulfill standard HDMI 1.4 standard for the much higher data rate.) But please advise, I am not on the top of the theme. :)

I just look at specs and from my finding it can not do full sbs. Max level it can handle is high profile 4.1. However full sbs is profile 5 or up.

MPEG-4.10 (H.264) : BP@L3, MP@L4.0, HP@L4.0, HP@L4.1

[mparade]

I just look at specs and from my finding it can not do full sbs. Max level it can handle is high profile 4.1. However full sbs is profile 5 or up.

Yes, you are right. I have found the same on the specification sheet of the player
about compatible x264 profiles. Anyway, just for your information the catalogue sheet of the product tells something about the main capability namely as follows: "this is the first Popcorn Hour media player which can even play 3D Blu-ray ISO files and folders. While on the earlier generation of media players only watching half-resolution side-by-side and top-bottom 3D contents was possible to enjoy as spatial images with model A-400 it also possible for 3D Blue-ray movies from USB, eSATA, shared network folders or from built-in hard disk. Here the 3D Blue-ray movies means Full 3D with intact resolution as I wrote previously, 1920x1080 pixel for each eye without any resize process made that is why the HDMI 1.4 compatibility with the amplifier or directly with the active 3D television. For passive televisions for sure it would be well enough for me to convert 3D Blu-rays to half-resolution top-and-bottom format due to detail loss by dividing the horizontal pixel rows between the eyes. I think I will wait for Mr. Dobbs to improve the software with such a function. I would like to see maximum details with the power x264 encoder...


az A-400 már a 3D Blu-ray filmekket is képes lejátszani USB-ről, eSATA-ról, hálózatról, vagy éppen a beszerelt merevlemezről.

[jdobbs]

Ok. Until then I have just sent a donation of 50$ to solve my problem within your program asap. :) Anyway, the newest network media tanks are compatible with Full 3D contents. (The ones with using HDMI 1.4 standard for example Popcorn hour A-400). But I think all the ones are compatible with such medias which are fulfill standard HDMI 1.4 standard for the much higher data rate.) But please advise, I am not on the top of the theme. :) Ok... I need clarification. What exactly are you asking for? Are you asking for output to a BD format -- but that has a (illegal by BD specs) 3840x1080 full resolution SBS picture in it? The hard part is in the video format definitions -- there is no code for that resolution.

Does that player not playback MKV files? X264 directly supports several stereoscopic frame-packing arrangements.

[HWK]

Ok... I need clarification. What exactly are you asking for? Are you asking for output to a BD format -- but that has a (illegal by BD specs) 3840x1080 full resolution SBS picture in it? The hard part is in the video format definitions -- there is no code for that resolution.

Does that player not playback MKV files? X264 directly supports several stereoscopic frame-packing arrangements.

My guess would be OP want to preserve 1920*1080 for both eye and still compatible with BD standard. Which as you mention 3840*1080 is not supported by players out there.

Although x264 can add frame packing signal inside source, not all decoders obey it.

[mparade]

Ok... I need clarification. What exactly are you asking for? Are you asking for output to a BD format -- but that has a (illegal by BD specs) 3840x1080 full resolution SBS picture in it? The hard part is in the video format definitions -- there is no code for that resolution.

Does that player not playback MKV files? X264 directly supports several stereoscopic frame-packing arrangements.

It is compatible with mkv files. Additionally, the player is compatible with the following video codecs according to manual:

XVID SD/HD
MPEG-1
MPEG-2 MP@HL
MPEG-4.2 ASP@L5, 720p, 1-pontos GMC
MPEG-4.10 (H.264): BP@L3, MP@L4.0, HP@L4.0, HP@L4.1
WMV9: MP@HL
SMPTE 421M (VC-1): MP@HL, AP@L3

I know your problem about BD specs and this is my problem, too: Blu-ray Discs only support level 4.1.

Anyway, into BDinfo I loaded my Jack The Giant Slayer Blue-ray 3D and it was said that the profile used by the main video title inside were according to L4.1 (see the results in the
attachement).

Anyway, my aim is to get the best I can from a 3D Blue ray movie: without any resize of the frames, keeping full resolution images to each eye (1920x1080) with the amazing reencoding capability of x264 encoder in BD-RB. So, I do not want intact frames inside, I want to make reencoding on frames due to the powerful capabilitites of x264.

If the format could be only mkv it is not a problem until I can play it without any issue (if I can reach the "same" video quality and compressing efficiency level as I would do with a Blue-ray compliant iso structure for sure).

Just a question because as I told you I am not on top of the theme: If 3840x1080 full resolution SBS is illegal by BD specs (and it is) how come that the 3D Blue-Ray disk I have is using L4.1 profile (which is to BD specs) besides it includes unresized Full 3D pictures (1920x1080) for each eye and recommended to use with HDMI 1.4 standard and active televisions. Sorry, but a little bit I have been messed up finally.

[HWK]

It is compatible with mkv files. Additionally, the player is compatible with the following video codecs according to manual:

XVID SD/HD
MPEG-1
MPEG-2 MP@HL
MPEG-4.2 ASP@L5, 720p, 1-pontos GMC
MPEG-4.10 (H.264): BP@L3, MP@L4.0, HP@L4.0, HP@L4.1
WMV9: MP@HL
SMPTE 421M (VC-1): MP@HL, AP@L3
This simply means what kind of footage decoder can handle, one of them is HP@ L4.1 which means max resolution is 1920* 1080 for video frame. If you want to feed 3840*1080 to decoder you will need level 5.0 at least.

I know your problem about BD specs and this is my problem, too: Blu-ray Discs only support level 4.1.

Anyway, into BDinfo I loaded my Jack The Giant Slayer Blue-ray 3D and it was said that the profile used by the main video title inside were according to L4.1 (see the results in the
attachement).

Anyway, my aim is to get the best I can from a 3D Blue ray movie: without any resize of the frames, keeping full resolution images to each eye (1920x1080) with the amazing reencoding capability of x264 encoder in BD-RB. So, I do not want intact frames inside, I want to make reencoding on frames due to the powerful capabilitites of x264.
X264 cannot do 3D like blu-ray like spec demand. You need mvc encoder for this job and for simplicity purpose I have omitted why this is the case.

If the format could be only mkv it is not a problem until I can play it without any issue (if I can reach the "same" video quality and compressing efficiency level as I would do with a Blue-ray compliant iso structure for sure).

Just a question because as I told you I am not on top of the theme: If 3840x1080 full resolution SBS is illegal by BD specs (and it is) how come that the 3D Blue-Ray disk I have is using L4.1 profile (which is to BD specs) besides it includes unresized Full 3D pictures (1920x1080) for each eye and recommended to use with HDMI 1.4 standard and active televisions. Sorry, but a little bit I have been messed up finally.

Blu-ray disc never store movies files in SBS format to begin with, instead they use MVC encoder which take base and dependent views and create two different files for final output to be used by authoring program. Only difference is for second view it only need max of 50% additional bitrate to achieve same quality. However on player side, player simply read from these two files for playback and constructs two frames of 1920*1080 with 45 pixel of spacing between them.

Example how you player pack footage and send to display top one is base image and bottom one is dependent image, however it could be another way as well depending on movie for example if it was life of pi then it would be inverted since views are store differently in that movie.
http://s22.postimg.org/jiwu3s7e9/Packing.png

With 3D on Blu-ray, each eye gets its own 1,920x1,080-pixel image. These are combined into one extra-tall frame, of which there are 24 per second. 45 pixels of blanking are included (the black space in the middle), resulting in a 1,920x2,205/24 signal.
(Credit: Geoffrey Morrison: Image, Screenshot: Dreamworks)

Because of extra data included with the two frames, the total amount of bits is slightly more than if it were a true 1080p/48 (1080p/24 x 2) signal. It is, as you see above, a 1,920x2,205p/24 signal.

[mparade]

This simply means what kind of footage decoder can handle, one of them is HP@ L4.1 which means max resolution is 1920* 1080 for video frame. If you want to feed 3840*1080 to decoder you will need level 5.0 at least.


X264 cannot do 3D like blu-ray like spec demand. You need mvc encoder for this job and for simplicity purpose I have omitted why this is the case.



Blu-ray disc never store movies files in SBS format to begin with, instead they use MVC encoder which take base and dependent views and create two different files for final output to be used by authoring program. Only difference is for second view it only need max of 50% additional bitrate to achieve same quality. However on player side, player simply read from these two files for playback and constructs two frames of 1920*1080 with 45 pixel of spacing between them.

Example how you player pack footage and send to display top one is base image and bottom one is dependent image, however it could be another way as well depending on movie for example if it was life of pi then it would be inverted since views are store differently in that movie.
http://s22.postimg.org/jiwu3s7e9/Packing.png

With 3D on Blu-ray, each eye gets its own 1,920x1,080-pixel image. These are combined into one extra-tall frame, of which there are 24 per second. 45 pixels of blanking are included (the black space in the middle), resulting in a 1,920x2,205/24 signal.
(Credit: Geoffrey Morrison: Image, Screenshot: Dreamworks)

Because of extra data included with the two frames, the total amount of bits is slightly more than if it were a true 1080p/48 (1080p/24 x 2) signal. It is, as you see above, a 1,920x2,205p/24 signal.

Thank you very much for the efforts working on my problem. In this case I will stay with my averagely 50 GB 3D iso packages for the time someone finds out a "non image degrading" solution to reencode them.....

[jdobbs]

@HWK

I think this is what you mean, but I just want to clarify to those reading... the two 1080p images aren't actually combined into a large O/U picture on the BD itself -- as you said, they are combined -- but it happens at playback and represents the way it is presented via HDMI to the playback device (monitor). It can do that because the required bandwidth for 1080p/48fps is already supported by HDMI.

I'm sure you know this -- but again for the reading public. The base/dependent views aren't actually two independent streams. They are one stream. The MVC portion of the stream can't exist or be played back without the base view (thus the term "dependent view"). But.. by design, so it works on a 2D player, the base portion of the stream can be presented on its own. On the physical disc there isn't truly a separate SSIF and M2TS stream -- the M2TS file is just a group of pointers that points to the sectors that hold the base view within the SSIF stream. That's the reason the required space grows by such a huge amount when you rip a 3D disc to your hard drive -- and its also why the new TSMUXER is forced to create an ISO image rather than a disc representation.

I still don't understand why they decided on an odd sized O/U picture (2205 pixels). It makes it difficult if anyone ever actually wanted to encode that way with H.264. I assume it has something to do with making the necessary electronics simpler.

The question is whether the player he mentioned can take an interleaved, SBS, or O/U picture and output it to an HDMI 1.4 compatible device. If so, and it supports MKV, there's no reason why he can't use the ALTERNATE output to make a full resolution backup.

[mparade]

@HWK

I think this is what you mean, but I just want to clarify to those reading... the two 1080p images aren't actually combined into a large O/U picture on the BD itself -- as you said, they are combined -- but it happens at playback and represents the way it is presented via HDMI to the playback device (monitor). It can do that because the required bandwidth for 1080p/48fps is already supported by HDMI.

I'm sure you know this -- but again for the reading public. The base/dependent views aren't actually two independent streams. They are one stream. The MVC portion of the stream can't exist or be played back without the base view (thus the term "dependent view"). But.. by design, so it works on a 2D player, the base portion of the stream can be presented on its own. On the physical disc there isn't truly a separate SSIF and M2TS stream -- the M2TS file is just a group of pointers that points to the sectors that hold the base view within the SSIF stream. That's the reason the required space grows by such a huge amount when you rip a 3D disc to your hard drive -- and its also why the new TSMUXER is forced to create an ISO image rather than a disc representation.

I still don't understand why they decided on an odd sized O/U picture (2205 pixels). It makes it difficult if anyone ever actually wanted to encode that way with H.264. I assume it has something to do with making the necessary electronics simpler.

The question is whether the player he mentioned can take an interleaved, SBS, or O/U picture and output it to an HDMI 1.4 compatible device. If so, and it supports MKV, there's no reason why he can't use the ALTERNATE output to make a full resolution backup.


It can play the following 3D formats:

side-by-side, top-bottom, MVC, frame packed, frame sequential, MKV 3D, Blu-ray 3D iso and folders.

The method the player is working with half-resolution mkv content is that the one has to complement the file name with a word 3DSBS (in case of using SBS method during encoding) or 3DTAB (in case of using O/U method during encoding) to automatically inform the display device how to cut and then intelligently upscale each of the halved images to 1920x1080 to each eye. This way the player automatically switches the display into 3D mode. The manual tells also that MKVToolnix from version 4.5 can provide the infos with the mkv files according to Matroska 3D specification and the player will use these information to automatically switch the display to 3D mode.

The player is prepared for HDMI 1.4 surely.

[jdobbs]

Try selecting "MKV Container, NO_RESIZE, AutoGOP, Auto-AAC" from the ALTERNATE output dialog and see how that looks/plays back.

[mparade]

Try selecting "MKV Container, NO_RESIZE, AutoGOP, Auto-AAC" from the ALTERNATE output dialog and see how that looks/plays back.

OK, thank you, I will let you know the results asap.

[HWK]

@HWK
I'm sure you know this -- but again for the reading public. The base/dependent views aren't actually two independent streams. They are one stream. The MVC portion of the stream can't exist or be played back without the base view (thus the term "dependent view"). But.. by design, so it works on a 2D player, the base portion of the stream can be presented on its own. On the physical disc there isn't truly a separate SSIF and M2TS stream -- the M2TS file is just a group of pointers that points to the sectors that hold the base view within the SSIF stream.

Sorry, but this is not correct info. I am gone post more through information later today and hopefully clear any confusion. I would try to cover basics of how it is created on disk, checking playback environment and decoding and finally sending signal to sink (Display)

I still don't understand why they decided on an odd sized O/U picture (2205 pixels). It makes it difficult if anyone ever actually wanted to encode that way with H.264. I assume it has something to do with making the necessary electronics simpler.

Once again, you are correct here.

The question is whether the player he mentioned can take an interleaved, SBS, or O/U picture and output it to an HDMI 1.4 compatible device. If so, and it supports MKV, there's no reason why he can't use the ALTERNATE output to make a full resolution backup.

[jdobbs]

Sorry, but this is not correct info That's possible, it wouldn't be the first time I've been wrong, but you'll have to show me where it's incorrect. I have the BD 3D spec and can confirm as needed.

[HWK]

That's possible, it wouldn't be the first time I've been wrong, but you'll have to show me where it's incorrect... I have the complete set of specifications for 3D on Blu-ray.

First point where you are wrong is statement "The base/dependent views aren't actually two independent streams. They are one stream."

They are indeed two seperate stream link with virtual ssif file. There are two m2ts files one for base and one for dependent view. (not including seamless branching) However for example purpose let look at avatar if you look inside ssif folder and look at 00001.ssif file it would be combination of 00001 + 00002 m2ts size and linking is part of UDF feature.

[jdobbs]

I'm not an expert on MVC, but I don't think that's correct. As I understand it, the dependent stream can have frames that backward of forward reference a frame in the base stream. That's why it only increases the bandwidth requirement by 50% rather than 100% -- because it doesn't have to repeat the common information. You can certainly pull out only the base or dependent portions into separate files (M2TS) in the STREAM folder -- and the base view would be nothing more than AVC, making it available for 2D reference and playback -- but I don't believe the dependent view can stand alone except when it has been decoded (e.g. to YUV) using the base stream for reference.

Maybe one of the AVC/MVC experts could weigh in... but I think I'm right.

[jdobbs]

OK, thank you, I will let you know the results asap.BD-RB doesn't set the stereo-mode flag in MKV output. So if the output doesn't automatically playback in 3D, try remuxing manually using MKVMerge using the "--stereo-mode" command line option. It should only take a few minutes.

If that works correctly, I can add the command line option within BD-RB.

[HWK]

I'm not an expert on MVC, but I don't think that's correct. As I understand it, the dependent stream can have frames that backward of forward reference a frame in the base stream. That's why it only increases the bandwidth requirement by 50% rather than 100% -- because it doesn't have to repeat the common information. You can certainly pull out only the base or dependent portions into separate files (M2TS) in the STREAM folder -- and the base view would be nothing more than AVC, making it available for 2D reference and playback -- but I don't believe the dependent view can stand alone except when it has been decoded (e.g. to YUV) using the base stream for reference.

Maybe one of the AVC/MVC experts could weigh in... but I think I'm right.

I am gone do few encode with mvc encoder and hopefully with that info I can prove ssif file is only used to link two files.

[jdobbs]

I am gone do few encode with mvc encoder and hopefully with that info I can prove ssif file is only used to link two files. They won't be "linked" exactly -- but they will be interleaved into a single MVC "stream" -- which, when multiplexed with audio/pgs streams in a BD, is the SSIF file. I guess that's a form of linking. Frankly I'm not even sure we're in that much of a disagreement. My point is that if the dependent stream is useless on its own -- it can't be called an "independent stream".

Try this: Look in a STREAMS folder. Play one of the two MVC based M2TS files with Media Player Classic -- the base view will play fine. But if you try to play back the dependent stream it won't work. I also doubt that you will find a reference to it in any of the MPLS files -- which means it is only referenced through its interleaving in the SSIF. Which, of course, matches what I meant by saying they are "one stream".

Each of the base and dependent parts does have its own PID, though, so I guess that could be used as an argument that they are at least separate in one way.

I wish someone who is an MVC programmer would speak up, though... I'm only familiar with how MVC works at a theoretical level and I'm certainly not an authoritative source.

[HWK]

Each of the base and dependent parts does have its own PID, though, so I guess that could be used as an argument that they are at least separate in one way.

I wish someone who is an MVC programmer would speak up, though... I'm only familiar with how MVC works at a theoretical level and I'm certainly not an authoritative source.

Ok, let try again this time with some example. I will start with avatar in this case and use this as an example.

1.In this step I look inside STREAM folder and right away I notice 00001.m2ts is main movie, because of its size and there is 0002.m2ts which is smaller than 0001.m2ts however it is bigger than other files. At this stage it is not possible to conclude if second file belong mvc file and during 3d playback it reference 0001.m2ts file. As a result I run media info on both files and stack them right next to each other.

After examining media info report it can be concluded that file has link because of time frame or duration is same for both files and off course framerate etc. In following step I am gone do frame comparison between two m2ts files and see what happens.

http://s24.postimg.org/5qyoofth1/0002.jpg

2.In this step I am gone perform comparison of frames; however before we start lets have quick look at track id. Also I am gone use SSIF file for comparison track ID, but use m2ts for extraction and comparison of frames(AVC=4113, MVC=4114 see screenshot of tsmuxer below with blue highlight in metafile section)

http://s24.postimg.org/bug98cjqt/0001.jpg

2B I added 0001.m2ts file in right and 0002.m2ts in left, in following screenshot I will show result. You can also check track id by looking at tsmuxer meta file section. (I extract both files at same time as a result time taken is different)

http://s23.postimg.org/zcqs89ukr/0003.jpg
http://s17.postimg.org/leydgihnj/0004.jpg

In this screenshot under output, one can see both m2ts have equal number of frames (232608) along with other useful info. I know what you are thinking why we have equal of frames even though base view has (I P B) and dependent view (P B)? Simply put encoder add additional frames in dependent view in order to make sure it stays synchronous during playback. Otherwise GOP will not have same number of frames for both video files and thus it won’t work.

I have included screenshot for illustrative purpose how encoder would encode MVC file, although in right view you don't see frame reference each other but it is allowed and encoder would do that. However there some limitation B frames can reference P frames and naturally I frame from base, but P cannot reference B frames.

http://s2.postimg.org/nuq4hpfp5/Capture.jpg

Here is another example with multiple cameras, however it is true for Blu-ray as well.

http://s10.postimg.org/rmym02n7d/00012.jpg

TIP: One of the doom9 user asked if we just encode H.264 and rebuild with existing MVC, file would it work and off course answer is no to this question.

Which probably raise question why have SSIF file in first place? To get better understanding let’s go back in time before 3D specs was finalized and at that time Blu-ray spec said video bitrate max is 40Mbps, however with 3D max bitrate for video is 60Mbps combined. Right away one would notice if we use this for Blu-ray authoring it will violate 2D Specs and will not allow backwards capability since bit rate is higher than 40Mbps and base view or 2D can’t have more than 40Mbps (Screenshot below)

http://s21.postimg.org/gdmkhhzlj/0005.jpg

Based on this SSIF (Stereoscopic Interleaved File) file extension was invented and used in 3D, in this case one simply takes 00001.m2ts + 00002.m2ts and join them in such way that ther are referred by reading 00001.m2ts + 00002.m2ts, 00001.m2ts + 00002.m2ts, 00001.m2ts + 00002.m2ts etc. Other way to look at this to visualize 1,3,5,7,9 sectors belong to 00001.m2ts and 2,4,6,8,10 sectors belong 00002.m2ts when interleave inside ssif file which is virtual it would be 1,2,3,4,5,6,7,8,9,10 etc

When player plays in 3d it does not skip over sectors, however when in 2d it skips over 3d part. Screenshot is shown for illustrative purposes below. This method used to make sure during playback laser only go in one direction rather than back and forth

http://s8.postimg.org/6cy6grcgl/0006.jpg

If examined carefully one would notice block for 2D is bigger than 3D, this is because audio and subs are usually embedded with 2D or base video, as a result in visualization we make it bigger for block which represent 2D video.

http://s12.postimg.org/ywkystcsd/0007.jpg

After reading this I am hoping that I can prove my point of having separate stream of MVC, but not decodable until combined with base view. I could go on how player determine when to play in 3d or 2d but it is outside scope of this discussion.

To conclude I would give analogy of a course which has theory and practical part to it. In order to fully benefit from this kind of course one need understand theory and then reinforce concept by practical example and for our case I would say map 3DBD specs map to theory and practical part to encoding experience. I am not sure jdobbs you have mvc encoder or not but I do so I provide experience to you which is next best thing after trying yourself.

Sources:
Mainconcept MVC encoder documentation
BD3D specs documentation
BD3D white paper overview
TsMuxer 2.1.6 use to create screenshots (Film:Avatar)
http://www.merl.com/reports/docs/TR2011-022.pdf
http://mpeg.chiariglione.org/standards/mpeg-4/advanced-video-coding
http://blu-ray-software.org/

[HWK]

My point is that if the dependent stream is useless on its own -- it can't be called an "independent stream".


My interpretation for multimedia file is with different packet ID (PID) which is 13 bit in size, not on bases of file contents.

Each of the base and dependent parts does have its own PID, though, so I guess that could be used as an argument that they are at least separate in one way.

This is exactly how I interpret multimedia file in first place by reading PID from stream.

[jdobbs]

This is all good material. But I don't see how any of this contradicts anything I've said? In fact it seems to confirm it???one can see both m2ts have equal number of frames (232608) along with other useful infoOf course they'd have the same number of frames, base (AVC) and dependent (MVC) views representing left and right (or right and left) -- just as I said. You can certainly pull out only the base or dependent portions into separate files (M2TS) in the STREAM folder -- and the base view would be nothing more than AVC, making it available for 2D reference and playback -- but I don't believe the dependent view can stand aloneAnd your figure 6.6 along with paragraph 2.6.4 confirms exactly how I explained the interleaved nature of the SSIF. On the physical disc there isn't truly a separate SSIF and M2TS stream -- the M2TS file is just a group of pointers that points to the sectors that hold the base view within the SSIF stream The graphic shows that the P & B frames in the dependent stream can only exist with the base view (making the dependent stream unplayable without the base stream -- just as I said.

You even say After reading this I am hoping that I can prove my point of having separate stream of MVC, but not decodable until combined with base view which is exactly what I said here My point is that if the dependent stream is useless on its own -- it can't be called an "independent stream"
So back to the original point of contention. Where am I "incorrect"? This is the statement you say is wrong: The base/dependent views aren't actually two independent streams. But then your own words show how the dependent view portion isn't independent. If anything I only see a disagreement over semantics as to what constitutes the term "stream". I say that a complete MVC stream that can be played in 3D must consist of the combination of the two (base and dependent) views.

So why are we having this painful discussion if we are saying the same thing?

[HWK]

Jdobbs, I think it all boils down to one thing and that is how we interpret stream in first place I use PID which is different for avc and mvc and as such I will interpret as two different stream, however for you interpretation is "dependent stream is useless on its own - it can't be called an "independent stream" as a result you don't see as a separate stream.

Even though with documentation we both forum same opinion with different angle of approaching it. Which means neither of us are wrong and also I am not in disagreement with you at all only thing is our interpretation is different. In light of this I would say it would good if we leave at this stage. Since going further make no sense and also it waste mine and your resources which could be put to better use somewhere else.

[jdobbs]

Jdobbs, I think it all boils down to one thing and that is how we interpret stream in first place I use PID which is different for avc and mvc and as such I will interpret as two different stream, however for you interpretation is "dependent stream is useless on its own - it can't be called an "independent stream" as a result you don't see as a separate stream.

Even though with documentation we both forum same opinion with different angle of approaching it. Which means neither of us are wrong and also I am not in disagreement with you at all only thing is our interpretation is different. In light of this I would say it would good if we leave at this stage. Since going further make no sense and also it waste mine and your resources which could be put to better use somewhere else.On this we can agree.

[Sharc]

Amen.
And thanks to both of you guys for a nice tutorial on the subject. Much appreciated. :)

[HWK]

Amen.
And thanks to both of you guys for a nice tutorial on the subject. Much appreciated. :)

Thank you, glad I can contribute :D

[HWK]

Jdobbs, I think it all boils down to one thing and that is how we interpret stream in first place I use PID which is different for avc and mvc and as such I will interpret as two different stream, however for you interpretation is "dependent stream is useless on its own - it can't be called an "independent stream" as a result you don't see as a separate stream.

Even though with documentation we both forum same opinion with different angle of approaching it. Which means neither of us are wrong and also I am not in disagreement with you at all only thing is our interpretation is different. In light of this I would say it would good if we leave at this stage. Since going further make no sense and also it waste mine and your resources which could be put to better use somewhere else.

On this we can agree.

One of things I have learned from this lengthy discussion is if not sure about other person interpretation then declare how you view it. In that case we would have possibly avoided whole discussion in first place. Oh well you always learn something from others :cool:

[jdobbs]

OK, thank you, I will let you know the results asap.BD-RB doesn't set the stereo-mode flag in MKV output. So if the output doesn't automatically playback in 3D, try remuxing manually using MKVMerge using the "--stereo-mode" command line option. It should only take a few minutes.

If that works correctly, I can add the command line option within BD-RB. @mparade

Any luck?

[HWK]

BD-RB doesn't set the stereo-mode flag in MKV output. So if the output doesn't automatically playback in 3D, try remuxing manually using MKVMerge using the "--stereo-mode" command line option. It should only take a few minutes.

If that works correctly, I can add the command line option within BD-RB.

If I remember correctly once you add --stereo-mode parameter file extension is change from MKV to MK3D, I build 3D while ago with bd-rebuilder and then process with mkvmerge end result player refuse when extension was changed, however plays when extension was changed again.

[mparade]

@mparade

Any luck?

I have just done the first full 3D (2160x1920) Top and Bottom 3D mkv container with your program with intact audio setup.
For that I simply modified one of your profiles in file alternate.txt file. The reencoding process finished without any problem 1 hour ago. Just for comparison purposes I am just making now an other 3D movie-only BD-complaint half-resolution reencoding for the same video. Anyway the first reencoding process even with the fastest 2-pass method took 2.5 hours for me to finish. (With Intel XEON 8-core processor 8Gbyte RAM, 64 bit system). Even this evening I will try if popcorn can handle them and send a signal to my HD TV.
Only in the next few days I can test that the correct signal is sent to a big enough 3D display.I let you know asap the results.

[jdobbs]

I have just done the first full 3D (2160x1920) Top and Bottom 3D mkv container with your program with intact audio setup.
For that I simply modified one of your profiles in file alternate.txt file. The reencoding process finished without any problem 1 hour ago. Just for comparison purposes I am just making now an other 3D movie-only BD-complaint half-resolution reencoding for the same video. Anyway the first reencoding process even with the fastest 2-pass method took 2.5 hours for me to finish. (With Intel XEON 8-core processor 8Gbyte RAM, 64 bit system). Even this evening I will try if popcorn can handle them and send a signal to my HD TV.
Only in the next few days I can test that the correct signal is sent to a big enough 3D display.I let you know asap the results.Thanks. Just remember that the O/U format isn't the same thing as the HDMI presentation format. There isn't any blank space between the two pictures. H-264 video can't do that because of the odd number of lines. It is the player's responsibility to convert O/U and SBS video into HDMI formats during playback. The other option is to let the monitor do it -- but normally they expect half resolution SBS or O/U.

[HWK]

I have just done the first full 3D (2160x1920) Top and Bottom 3D mkv container with your program with intact audio setup.
For that I simply modified one of your profiles in file alternate.txt file. The reencoding process finished without any problem 1 hour ago. Just for comparison purposes I am just making now an other 3D movie-only BD-complaint half-resolution reencoding for the same video. Anyway the first reencoding process even with the fastest 2-pass method took 2.5 hours for me to finish. (With Intel XEON 8-core processor 8Gbyte RAM, 64 bit system). Even this evening I will try if popcorn can handle them and send a signal to my HD TV.
Only in the next few days I can test that the correct signal is sent to a big enough 3D display.I let you know asap the results.

I am interested as well when I tired before I was met with failure. If not much of hassle can you post media info report for file you just encoded.

[mparade]

I am interested as well when I tired before I was met with failure. If not much of hassle can you post media info report for file you just encoded.

Just until then here you can see the process status of each version. As you can see the second one is just being reencoded...:)

[11.09.13] BD Rebuilder v0.44.17 (beta)
[12:34:55] Source: JACK_THE_GIANT_SLAYER_00098
- Input BD size: 27,43 GB
- Approximate total content: [01:54:08.883]
- Windows Version: 6.1 [7601]
- MOVIE-ONLY/ALTERNATE OUTPUT mode enabled
- Mode: MKV Container, NO_RESIZE, AutoGOP, Intact Audio
- 3D conversion mode enabled
- Decoding/Frame serving: DirectShowMVC
- Audio Settings: AC3=0 DTS=0 HD=0 Kbs=640
[12:34:55] PHASE ONE, Encoding
- [12:34:55] Processing: VID_00063 (1 of 1)
- [12:34:55] Extracting A/V streams [VID_00063]
- [12:45:10] Reencoding video [VID_00063]
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23,976fps, 164*209 frames
- Converting output to Over/Under 3D Format
- Bitrate: 13*000 Kbs
- [12:45:10] Reencoding: VID_00063, Pass 1 of 2
- [13:54:23] Reencoding: VID_00063, Pass 2 of 2
- [15:03:03] Video Encode complete
- [15:03:03] Processing audio tracks
- Track 4352 (eng): Keeping original audio
- Track 4354 (hun): Keeping original audio
[15:03:03]PHASE ONE complete
[15:03:03]PHASE TWO - Rebuild Started
- [15:03:04] Building ALTERNATE OUTPUT Structure
[15:07:47] - Encode and Rebuild complete
[15:07:47] JOB: JACK_THE_GIANT_SLAYER finished.
----------------------
[11.09.13] BD Rebuilder v0.44.17 (beta)
[21:35:13] Source: JACK_THE_GIANT_SLAYER_00098
- Input BD size: 27,43 GB
- Approximate total content: [01:54:08.883]
- Target BD size: 7,81 GB
- Windows Version: 6.1 [7601]
- MOVIE-ONLY mode enabled
- Quality: Good (Very Fast), Two Pass
- 3D conversion mode enabled
- Decoding/Frame serving: DirectShowMVC
- Audio Settings: AC3=0 DTS=0 HD=0 Kbs=640
[21:35:13] PHASE ONE, Encoding
- [21:35:13] Processing: VID_00063 (1 of 1)
- [21:35:13] Extracting A/V streams [VID_00063]
- [21:45:28] Reencoding video [VID_00063]
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23,976fps, 164*209 frames
- Converting output to Over/Under 3D Format
- Bitrate: 6*125 Kbs
- [21:45:28] Reencoding: VID_00063, Pass 1 of 2
- [22:44:13] Reencoding: VID_00063, Pass 2 of 2

[mparade]

My first impression is that my confirmation within a few days on the quality and proper signal sending/decoding of Full 3D TAB mkv files will be quite positive. My Popcorn A-400 can playback both format flawlessly. What remained is the 3D display test....

[mparade]

Ok, let try again this time with some example. I will start with avatar in this case and use this as an example.

1.In this step I look inside STREAM folder and right away I notice 00001.m2ts is main movie, because of its size and there is 0002.m2ts which is smaller than 0001.m2ts however it is bigger than other files. At this stage it is not possible to conclude if second file belong mvc file and during 3d playback it reference 0001.m2ts file. As a result I run media info on both files and stack them right next to each other.

After examining media info report it can be concluded that file has link because of time frame or duration is same for both files and off course framerate etc. In following step I am gone do frame comparison between two m2ts files and see what happens.

http://s24.postimg.org/5qyoofth1/0002.jpg

2.In this step I am gone perform comparison of frames; however before we start lets have quick look at track id. Also I am gone use SSIF file for comparison track ID, but use m2ts for extraction and comparison of frames(AVC=4113, MVC=4114 see screenshot of tsmuxer below with blue highlight in metafile section)

http://s24.postimg.org/bug98cjqt/0001.jpg

2B I added 0001.m2ts file in right and 0002.m2ts in left, in following screenshot I will show result. You can also check track id by looking at tsmuxer meta file section. (I extract both files at same time as a result time taken is different)

http://s23.postimg.org/zcqs89ukr/0003.jpg
http://s17.postimg.org/leydgihnj/0004.jpg

In this screenshot under output, one can see both m2ts have equal number of frames (232608) along with other useful info. I know what you are thinking why we have equal of frames even though base view has (I P B) and dependent view (P B)? Simply put encoder add additional frames in dependent view in order to make sure it stays synchronous during playback. Otherwise GOP will not have same number of frames for both video files and thus it won’t work.

I have included screenshot for illustrative purpose how encoder would encode MVC file, although in right view you don't see frame reference each other but it is allowed and encoder would do that. However there some limitation B frames can reference P frames and naturally I frame from base, but P cannot reference B frames.

http://s2.postimg.org/nuq4hpfp5/Capture.jpg

Here is another example with multiple cameras, however it is true for Blu-ray as well.

http://s10.postimg.org/rmym02n7d/00012.jpg

TIP: One of the doom9 user asked if we just encode H.264 and rebuild with existing MVC, file would it work and off course answer is no to this question.

Which probably raise question why have SSIF file in first place? To get better understanding let’s go back in time before 3D specs was finalized and at that time Blu-ray spec said video bitrate max is 40Mbps, however with 3D max bitrate for video is 60Mbps combined. Right away one would notice if we use this for Blu-ray authoring it will violate 2D Specs and will not allow backwards capability since bit rate is higher than 40Mbps and base view or 2D can’t have more than 40Mbps (Screenshot below)

http://s21.postimg.org/gdmkhhzlj/0005.jpg

Based on this SSIF (Stereoscopic Interleaved File) file extension was invented and used in 3D, in this case one simply takes 00001.m2ts + 00002.m2ts and join them in such way that ther are referred by reading 00001.m2ts + 00002.m2ts, 00001.m2ts + 00002.m2ts, 00001.m2ts + 00002.m2ts etc. Other way to look at this to visualize 1,3,5,7,9 sectors belong to 00001.m2ts and 2,4,6,8,10 sectors belong 00002.m2ts when interleave inside ssif file which is virtual it would be 1,2,3,4,5,6,7,8,9,10 etc

When player plays in 3d it does not skip over sectors, however when in 2d it skips over 3d part. Screenshot is shown for illustrative purposes below. This method used to make sure during playback laser only go in one direction rather than back and forth

http://s8.postimg.org/6cy6grcgl/0006.jpg

If examined carefully one would notice block for 2D is bigger than 3D, this is because audio and subs are usually embedded with 2D or base video, as a result in visualization we make it bigger for block which represent 2D video.

http://s12.postimg.org/ywkystcsd/0007.jpg

After reading this I am hoping that I can prove my point of having separate stream of MVC, but not decodable until combined with base view. I could go on how player determine when to play in 3d or 2d but it is outside scope of this discussion.

To conclude I would give analogy of a course which has theory and practical part to it. In order to fully benefit from this kind of course one need understand theory and then reinforce concept by practical example and for our case I would say map 3DBD specs map to theory and practical part to encoding experience. I am not sure jdobbs you have mvc encoder or not but I do so I provide experience to you which is next best thing after trying yourself.

Sources:
Mainconcept MVC encoder documentation
BD3D specs documentation
BD3D white paper overview
TsMuxer 2.1.6 use to create screenshots (Film:Avatar)
http://www.merl.com/reports/docs/TR2011-022.pdf
http://mpeg.chiariglione.org/standards/mpeg-4/advanced-video-coding
http://blu-ray-software.org/


I have just finished reading your course. It was very good "material" for me to see "behind the curtain". Thank You!

[HWK]

I have just finished reading your course. It was very good "material" for me to see "behind the curtain". Thank You!

Thank you, certainly appreciated :)

[mparade]

Thanks. Just remember that the O/U format isn't the same thing as the HDMI presentation format. There isn't any blank space between the two pictures. H-264 video can't do that because of the odd number of lines. It is the player's responsibility to convert O/U and SBS video into HDMI formats during playback. The other option is to let the monitor do it -- but normally they expect half resolution SBS or O/U.

I have just finished with the test of a "Full 3D O/U" material made by using one of the BD-RB alternate mkv output profiles namely MKV Container, NO_RESIZE, AutoGOP, Intact Audio.

Test devices and contents: Popcorn hour A-400, Loewe Connect ID 46, Loewe 3D active glasses, video: Jack the Giant Slayer 3D.mk3d including Full Resolution 3D content made by BD-RB.

Test steps were as follows:

1. making an O/U mkv file with BD-RB using the correct mkv profile in the resolution of 1920x2160. Fast two pass encoding
was used with such options like open GOPS etc.

2. In mkv merge I loaded this file and done the following setup of the video file to let the player know what type of content is located inside: framerate to 24/1001; resolution for display kept as 1920x2160; stereoscopy mode set to O/U, format to 16:9, saved as with the extension of mk3d.

With this setup of the mkv container and such a HDMI1.4 compatible player like mentioned above which can interpret such a picture format I just had to manually switch my 3D active TV to topandbottom 3D mode. (unfortunately in automatic mode it did not recognized correctly the picture format the player sent, and was a bit uncomfortable, though.

I tested the player behaviour for about 2 hours while playing such a "processor eater" content with active 3D glasses worn.
(the player is prepared for HDMI 1.4 & 3D MVC Blue-ray content, so the higher data rates should not be a problem)

The results was so good that I didn't want to stop watching the video. :). Inspite of the small size of my 3D TV there is definitely a difference in quality (it is subjective for sure, so if you do not really believe, please test it by yourself) comparing to the "half-resolution" 3D content which is generally used of discs authored according to BD specs. Every function of the player worked flawlessly (chapter skipping, switching over to other languages/subtitles, forwarding/backwarding upto 8X) however if I would have to compare the speed of these functions to the ones while watching 2D Blue-ray contents a very bit of hanging should be mentioned.

I think Mr. Jdobbs should also consider these results. Maybe he can make use of them in the improvement procedure of BD-RB. Personally I was really surprised at the results.

[mparade]

BD-RB doesn't set the stereo-mode flag in MKV output. So if the output doesn't automatically playback in 3D, try remuxing manually using MKVMerge using the "--stereo-mode" command line option. It should only take a few minutes.

If that works correctly, I can add the command line option within BD-RB.

It was working correctly! I think it is not enough to add a command line option for such an mkv file to work correctly in my player. What I had to do manually in mkv merge after getting the mkv output from BD-RB to get the file to function correctly: framerate to 24/1001; resolution for display kept as 1920x2160; stereoscopy mode set to O/U, format to 16:9, saved as with the extension of mk3d.

Please setup your program so that I do not have to do any manual adjustments such these in mkv merge with each mkv file I got from BD-RB. Can you do that? It would be very appreciated.

Another thing to mention: as you know the output file now contains a very large resolution image (in case of O/U arrangement it is a 1920x2160). Could you please check also if the values of the general (which cannot be modified through the setup menu or hiddenopts) video encoding parameters of the actual "MKV Container, NO_RESIZE, AutoGOP, Intact Audio" profile I would like to use for my encodings is still optimal for encoding such a large big resolution image?

Anyway, when using an alternate output mkv profile all the parameters I set up in the setup dialog and through the hiddenopts are still valid? So, for example if I use open GOPs
in the setup dialogue is it valid for an alternate output profile also?

You know my aim is the same as yours: to make an virtually indistinguishable copy to the original. But this time the encoded picture is a double-sized one and the results should be compared to a 3D image according to 3D Blue-ray specs. (1920x1080p image to each eye).

Please help me setting up a profile for encoding Full 3D contents, I would really appreciate your efforts again.

Personally what is the minimum average bitrate you would accept for a the main movie part in the resolution of 1920x1080p/1920x2160p in case of 2-pass encoding? (It is subjective upto some degree for sure but not so much that cannot estimate) :)

Thank you for your answer.

[jdobbs]

But the framerate and resolution settings are always there anyway? The output aspect ratio isn't 16:9, it is 1:1 (it just happens that a 1:1 ratio at 1920x1080 is overall 16:9). Why would I need to do anything differently other than "--stereo-mode" (which I have already added)?

The presets for ALTERNATE output can be edited and new presets can be added. You just use notepad (or another editor) on the ALTERNATES.TXT file in the MISC folder. There is text at the top describing the editable options.

Can you give more details as to exactly what settings you used in MKVMERGE? I'm ready to release a new version, and if I don't get them soon it's going to have to wait for the next release (which, with the holidays, may be a while).

[mparade]

But the framerate and resolution settings are always there anyway? The output aspect ratio isn't 16:9, it is 1:1 (it just happens that a 1:1 ratio at 1920x1080 is overall 16:9). Why would I need to do anything differently other than "--stereo-mode" (which I have already added)?

The presets for ALTERNATE output can be edited and new presets can be added. You just use notepad (or another editor) on the ALTERNATES.TXT file in the MISC folder. There is text at the top describing the editable options.

Can you give more details as to exactly what settings you used in MKVMERGE? I'm ready to release a new version, and if I don't get them soon it's going to have to wait for the next release (which, with the holidays, may be a while).

This is the command line of the mk3d file (which was made from the mkv file I got directly from BD-RB using an mkv profile with the description of "MKV Container, NO_RESIZE, AutoGOP, Intact Audio" + Type_of_3D=1 (O/U)) in mkvmerge that could be properly played on my popcorn A-400.

"mkvmerge" -o "E:\\Users\\X\\Desktop\\JACK_THE_GIANT_SLAYER OU_left first (1).mk3d" "--default-track" "0:yes" "--forced-track" "0:no" "--display-dimensions" "0:3840x2160" "--stereo-mode" "0:3" "--language" "1:eng" "--default-track" "1:yes" "--forced-track" "1:no" "--language" "2:hun" "--default-track" "2:no" "--forced-track" "2:no" "--language" "3:eng" "--default-track" "3:no" "--forced-track" "3:no" "--language" "4:hun" "--default-track" "4:no" "--forced-track" "4:no" "-a" "1,2" "-d" "0" "-s" "3,4" "-T" "--no-global-tags" "(" "E:\\Users\\X\\Desktop\\JACK_THE_GIANT_SLAYER OU_left first.mk3d" ")" "--track-order" "0:0,0:1,0:2,0:3,0:4"

Please extract all the data what is required by you, and let me know if it is lack of something to make the adjustments in mkvmerge needed. Generally, I am going to use BD-RB to produce the highest quality Full 3D content with a specific mkv profile, namely the one: "MKV Container, NO_RESIZE, AutoGOP, Intact Audio" with modified/complemented with some hidden opt parameter.

[jdobbs]

I don't think that could be correct. For example: "--display-dimensions 0:3840x2160". I would think it would be 3840x1080 for "--stereo-mode 0:1" or 1920x2160 for "--stereo-mode 0:3" I also have to be careful about calling out the dimensions directly -- because there is no reason why a source stereo image couldn't be 1280x720 (2560x720,1280x1440) in SBS or O/U format when you have NO_RESIZE set.

[HWK]

I don't think that could be correct. For example: "--display-dimensions 0:3840x2160". I would think it would be 3840x1080 for "--stereo-mode 0:1" or 1920x2160 for "--stereo-mode 0:3"

I'll have to wait until the next release.

Jdobbs, On mkvmerge website it list all available options

--stereo-mode TID:n|keyword

Sets the stereo mode for the video track with the track ID TID. The mode can either be a number n between 0 and 14 or one of these keywords:

'mono', 'side_by_side_left_first', 'top_bottom_right_first', 'top_bottom_left_first', 'checkerboard_right_first', 'checkerboard_left_first', 'row_interleaved_right_first', 'row_interleaved_left_first', 'column_interleaved_right_first', 'column_interleaved_left_first', 'anaglyph_cyan_red', 'side_by_side_right_first', 'anaglyph_green_magenta', 'both_eyes_laced_left_first', 'both_eyes_laced_right_first'.