BD Rebuilder Beta - Bug Reports Only

[MrVideo]

Originally Posted by Lathe

(this almost reminds me of the old Monty Python sketch, 'Well, when I was young, we had to etch our own recordings in granite...')

Doesn't go back far enough.

[thekiller]

Hello,
How to have an HEVC + AC3 profile instead of AAC in the Alternate Output profile?
Thank you

[jdobbs]

Originally Posted by thekiller

Hello,
How to have an HEVC + AC3 profile instead of AAC in the Alternate Output profile?
Thank you

Just add a new profile to the ALTERNATE.TXT file in the MISC folder. At the top of the file you'll see the settings that can be used. Here is an example of an HEVC profile with AC3 that could be added to the end of the file (assuming it hasn't been altered):

[00039]
caption=MKV Container, HEVC 1920x1080, 448Kbs DD5.1
vEncoder=1
vBitrate=2000
vKeyint=Auto
aBitrate=448
aType=0
vFormat=5
cType=1

[CV91913]

I obviously have some setting screwed up somewhere as every DVD import to Alternate MP4 output I do has problems with "jumpy" video anytime there is action. I have never gotten a DVD conversion to work properly.

The psudo BD structure in the Import directory plays just fine. It is only the resulting MP4 that is jumpy.

Here are my logs and Alternate output settings.

[23:07:28] Importing DVD: SNOWDEN
 - Processing DVD title [1 of 1 ]
   - Collecting audio/video/subtitle streams...
   - Converting subtitles to BD format...
   - Multiplexing audio/video/subtitles...
 - Creating Quick-Play Menu...
[23:13:38] DVD import completed successfully.
----------------------
[02/03/17] BD Rebuilder v0.50.20
[23:17:53] Source:  SNOWDEN_00001
  - Input BD size: 5.46 GB
  - Approximate total content: [02:14:49.581]
  - Windows Version: 6.1 [7601]
  - MOVIE-ONLY/ALTERNATE OUTPUT mode enabled
  - Mode: DVD to MP4, No Resize, Stereo-AAC
  - Quality: Good (Very Fast)
  - Decoding/Frame serving: X264/LAVF
  - Audio Settings: AC3=0 DTS=1 HD=0 Kbs=640
[23:17:53] PHASE ONE, Encoding
 - [23:17:53] Processing: VID_00001 (1 of 1)
 - [23:17:53] Extracting A/V streams [VID_00001]
 - [23:21:14] Reencoding video [VID_00001]
   - Source Video: MPEG-2, 720x480
   - Rate/Length: 29.970fps, 242,445 frames
   - Bitrate: 3,288 Kbs
 - [23:21:14] Reencoding: VID_00001, Pass 1 of 2
 - [23:52:47] Reencoding: VID_00001, Pass 2 of 2
 - [00:23:54] Video Encode complete
 - [00:23:54] Processing audio tracks
   - Track 4352 (eng): Reencoding audio to AAC...
[00:30:26]PHASE ONE complete
[00:30:26]PHASE TWO - Rebuild Started
 - [00:30:26] Building ALTERNATE OUTPUT Structure
[00:35:41] - Encode and Rebuild complete
 - WORKFILES folder removed.
[00:35:41] JOB: SNOWDEN finished.

[Status]
LABEL=SNOWDEN
VERSION=v0.50.20
SOURCE_SIZE=5861591040
SOURCE_VIDEO_SIZE=5861591040
TARGET_SIZE=24641536000
REDUCTION=4.20389887862255
RESIZE_1080=0
RESIZE_1440=0
AUDIO_TO_KEEP=eng;
KEEP_HD_AUDIO=0
SUBS_TO_KEEP=eng;
BACKUP_MODE=1
MOVIEONLY_TYPE=42
USE_LAVF=-1
INSTANCES=1
DGDECNV=0
DGDECIM=0
FRIMSOURCE=0
FFMS2=0
SSIF_MODE=0
QUICK=0
ENCODE_STEP=0
COMPLETED=1
REBUILD_COMPLETE=1
[00001]
AUDIO=1
PGS=100
MSCAN=2
SCANCOUNT=193960
VIDEO2=0
V2MBRATE=0
M2TS_TARGET=24641536000
SPLITS=1
NSIZE=0
FLINK=0
MLINK=0

[00042]
caption=DVD to MP4, No Resize, Stereo-AAC
vBitrate=2500
aBitrate=256
aType=1
vFormat=7
cType=5

[02/04/17] Checking System Settings
  - BD-Rebuilder v0.50.20
  - Windows Version: 6.1 [7601]
  - Working Path Free Space: 347.16GB
  - AVISYNTH Version: 2.5.8.0, Ok
  - LAVFILTERS:  Ok
  - AnyDVD settings check: Ok.
  - X264:  Ok
  - AFTEN:  Ok
  - FAAC:  Ok
  - MP4BOX:  Ok
  - WAVI:  Ok
  - TSMUXER:  Ok
  - FRIMEncode:  Ok
  - FRIMDecode:  Ok
[02/04/17] Systems Settings Check complete

Hope someone has some ideas.

[jdobbs]

@CV91913

Have you tried turning on "IVTC sources with 3:2 pulldown" in the SETUP dialog. I think you almost always need that on when converting from an NTSC DVD.

[Matrix789]

Hi everybody,

I demuxed a movie with eac3to, put only the video track into an mkv with mkvtoolnix and then encoded with bd rebuilder (best video quality and custom size 10240MB).
The encoding went fine but, when I made my final mkv with the encoded video track and the untouched audio, I noticed that the movie is now out of sync. The video track is in fact 5 seconds longer than the audio/original video track.

Do you know why, how to fix it and how to prevent it to happen again? I encoded only the video track because I wanted the target output size to be only focused on the video track.

Thank you

EDIT: I actually noticed that the video track appears to be 5 seconds longer even before the encoding. How is it even possible?

[CV91913]

Originally Posted by jdobbs

@CV91913

Have you tried turning on "IVTC sources with 3:2 pulldown" in the SETUP dialog. I think you almost always need that on when converting from an NTSC DVD.

I just tried that and can see no difference. Still jumpy. Creating MP4 from Bluray works great. Just DVD's are a problem.

I have tried constant and variable bitrates and 1 pass and 2 pass. Nothing seems to fix it. Any other suggestions?

[jdobbs]

Originally Posted by CV91913

I just tried that and can see no difference. Still jumpy. Creating MP4 from Bluray works great. Just DVD's are a problem.

I have tried constant and variable bitrates and 1 pass and 2 pass. Nothing seems to fix it. Any other suggestions?

Is there any way you can create a snippet of the jumpy section so I can see what you mean?

[jdobbs]

Originally Posted by Matrix789

Hi everybody,

I demuxed a movie with eac3to, put only the video track into an mkv with mkvtoolnix and then encoded with bd rebuilder (best video quality and custom size 10240MB).
The encoding went fine but, when I made my final mkv with the encoded video track and the untouched audio, I noticed that the movie is now out of sync. The video track is in fact 5 seconds longer than the audio/original video track.

Do you know why, how to fix it and how to prevent it to happen again? I encoded only the video track because I wanted the target output size to be only focused on the video track.

Thank you

EDIT: I actually noticed that the video track appears to be 5 seconds longer even before the encoding. How is it even possible?

Sometimes the audio has a different start time than the video. You see that with intro studio symbols (not typically, but every now and again).

Why not just create a profile that leaves the audio untouched -- and let BD-RB do all the work? It automatically adjusts when the audio timestamp deviates from that of the video.

[meadrocks]

I'm having challenges with this title. I'm doing movie only, alternate output, mkv file, auto-aac, auto-gop. It is Playlist 00001. The audio is out of sync and there are no subtitles being displayed. I selected original swedish audio & english subtitles, not forced on.

Here is the setup info.
[Status]
LABEL=GIRL_WITH_THE_DRAGON_TATTOO
VERSION=v0.50.23
SOURCE_SIZE=19614609408
SOURCE_VIDEO_SIZE=19614609408
TARGET_SIZE=4686086144
REDUCTION=.238907951034127
RESIZE_1080=0
RESIZE_1440=0
AUDIO_TO_KEEP=eng;
KEEP_HD_AUDIO=0
SUBS_TO_KEEP=all
BACKUP_MODE=1
MOVIEONLY_TYPE=17
USE_LAVF=0
INSTANCES=1
DGDECNV=0
DGDECIM=0
FRIMSOURCE=0
FFMS2=0
SSIF_MODE=0
QUICK=0
ENCODE_STEP=0
COMPLETED=4
REBUILD_COMPLETE=1
[00001]
AUDIO=
PGS=1
APULLDOWN=0
S1440=0
VIDEO2=0
V2MBRATE=0
M2TS_TARGET=1397394
NSIZE=0
FLINK=0
MLINK=0
[00000]
APULLDOWN=0
S1440=0
VIDEO2=0
V2MBRATE=0
M2TS_TARGET=4684376098
SPLITS=1
AUDIO=10
NSIZE=0
FLINK=0
MLINK=0
[00017]
APULLDOWN=0
S1440=0
VIDEO2=0
V2MBRATE=0
M2TS_TARGET=54310
SPLITS=1
AUDIO=
NSIZE=0
FLINK=0
MLINK=0
[00018]
APULLDOWN=0
S1440=0
VIDEO2=0
V2MBRATE=0
M2TS_TARGET=258342

Here's the output from bd-rebuilder log file.

==========================
[19:58:45] BD Rebuilder v0.50.23
[19:58:45] BEGIN - BATCH SEQUENCE
==========================
-----------------------
[19:58:45] PROCESSING BATCH FILE [1]
----------------------
[02/02/17] BD Rebuilder v0.50.23
[04:21:40] Source: GIRL_WITH_THE_DRAGON_TATTOO_00001
- Input BD size: 18.27 GB
- Approximate total content: [02:33:02.214]
- Windows Version: 6.1 [7601]
- MOVIE-ONLY/ALTERNATE OUTPUT mode enabled
- Mode: MKV Container, 1920x1080, Auto-AAC, AutoGOP
- Quality: High Quality (Default)
- SHUTDOWN at completion enabled
- Decoding/Frame serving: DirectShow
- Audio Settings: AC3=0 DTS=0 HD=0 Kbs=640
[04:21:41] PHASE ONE, Encoding
- [04:21:41] Processing: VID_00001 (1 of 4)
- [04:21:41] Extracting A/V streams [VID_00001]
- [05:03:47] Reencoding video [VID_00001]
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 240 frames
- [05:03:47] Reencoding: VID_00001, Pass 1 of 1
- [05:03:58] Video Encode complete
- [05:03:58] Processing audio tracks
- Track 4352 (swe): Reencoding audio to AAC...
- [05:22:25] Processing: VID_00000 (2 of 4)
- [05:22:25] Extracting A/V streams [VID_00000]
- [05:31:33] Reencoding video [VID_00000]
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 219,720 frames
- [05:31:33] Reencoding: VID_00000, Pass 1 of 1
- [10:30:15] Video Encode complete
- [10:30:15] Processing: VID_00017 (3 of 4)
- [10:30:15] Extracting A/V streams [VID_00017]
- [10:30:21] Reencoding video [VID_00017]
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 96 frames
- [10:30:21] Reencoding: VID_00017, Pass 1 of 1
- [10:30:25] Video Encode complete
- [10:30:25] Processing: VID_00018 (4 of 4)
- [10:30:25] Extracting A/V streams [VID_00018]
- [10:30:29] Reencoding video [VID_00018]
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 97 frames
- [10:30:29] Reencoding: VID_00018, Pass 1 of 1
- [10:30:33] Video Encode complete
[10:30:34]PHASE ONE complete
[10:30:34]PHASE TWO - Rebuild Started
- [10:30:34] Building ALTERNATE OUTPUT Structure
[10:35:18] - Encode and Rebuild complete
[10:35:18] JOB: GIRL_WITH_THE_DRAGON_TATTOO finished.
==========================
[10:35:18] END - BATCH SEQUENCE
==========================
- Computer SHUTDOWN initiated.

Thanks.

[CV91913]

Originally Posted by jdobbs

Is there any way you can create a snippet of the jumpy section so I can see what you mean?

No problem. I have about 2 min that should demonstrate the problem. It is about 43mb. Where should I upload it?

[Matrix789]

Originally Posted by jdobbs

Sometimes the audio has a different start time than the video. You see that with intro studio symbols (not typically, but every now and again).

Why not just create a profile that leaves the audio untouched -- and let BD-RB do all the work? It automatically adjusts when the audio timestamp deviates from that of the video.

Thank you for your reply.
The reason why I don't encode all together is because I want the video track to be the size that I set and not the whole file.
Also, I tried to encode the whole thing (leaving the dts audio untouched) but at the end of the process I couldn't hear any audio and on MediaInfo it looked like the audio track had been "touched" anyway.

Will it work I delay the audio track by 5 seconds with mkvtoolnix? And if yes, do I have to delay also the subtitles and chapters tracks or they are fine?

Thanks

EDIT: ok this is weird. I demuxed everything from the blu ray and I put it into a mkv: the movie duration is correct, 1hr 30min 40sec.
If I import the same mkv into bd rebuilder it shows that the movie length is 5 sec longer, 1hr 30min 45 sec. That's why after encoding the video track I can't put it back with the untouched audio: bd rebuilder "thinks" that the video is 5 sec longer than it actually is.
How do I fix it? I need bd rebuilder to see the correct length. Delaying or stretching the audio track didn't work.

[mparade]

Hello,

I was away from this forum for a long time...
Sorry for asking such things here, but can someone send a link/tell how the donation is working now?

After pushing "donation" I was linked to "Amazon.com eGift Cards"
page. E.g. what to write into the "to" field?

Sorry for asking such things that may have been answered here several times already.

[meadrocks]

Originally Posted by mparade

Hello,

I was away from this forum for a long time...
Sorry for asking such things here, but can someone send a link/tell how the donation is working now?

After pushing "donation" I was linked to "Amazon.com eGift Cards"
page. E.g. what to write into the "to" field?

Sorry for asking such things that may have been answered here several times already.

Buy an Amazon Gift card & PM jdobbs the code. Don't need anything in the To field.

[gonca]

Originally Posted by mparade

Hello,

I was away from this forum for a long time...
Sorry for asking such things here, but can someone send a link/tell how the donation is working now?

After pushing "donation" I was linked to "Amazon.com eGift Cards"
page. E.g. what to write into the "to" field?

Sorry for asking such things that may have been answered here several times already.

http://forum.doom9.org/showthread.ph...30#post1784130

[jdobbs]

Originally Posted by meadrocks

Buy an Amazon Gift card & PM jdobbs the code. Don't need anything in the To field.

Or send it to dvd-rb@jdobbs.net

[CV91913]

Originally Posted by jdobbs

Is there any way you can create a snippet of the jumpy section so I can see what you mean?

Yes, I uploaded a 8mb snippet to your dvd-rb@jdobbs.net email.

[jdobbs]

I've been doing some interesting testing recently. I've been thinking about using HEVC for my SERVIIO playback rather than AVC. I've been using X264 and X265 as the encoders with CRF encoding. I'm kinda' disappointed with the results I'm seeing in the tests. It seems that, while the same CRF value on the surface appears to save space, HEVC/X265 is actually using more space to get equivalent SSIM results. This definitely isn't what I was expecting. In fact the only time I'm seeing comparable or better results is at high CRF values that I wouldn't normally use.

Has anyone else seeing this kind of behavior? I saw a report that Netflix did some tests that disagree with what I'm seeing. In fact, they're saying X265 can compete with X264 at half the bitrate. Only at very-low bitrates, maybe?

Also -- at what value does SSIM reach a point that is generally considered "acceptable"? (I know this is really subjective and probably changes based on resolution, but I'm hoping that there's a consensus somewhere).

I think I'm going to run some tests using fixed bitrates rather than CRF and see what happens.

[Sharc]

Originally Posted by jdobbs

I've been doing some interesting testing recently. I've been thinking about using HEVC for my SERVIIO playback rather than AVC. I've been using X264 and X265 as the encoders with CRF encoding. I'm kinda' disappointed with the results I'm seeing in the tests. It seems that, while the same CRF value on the surface appears to save space, HEVC/X265 is actually using more space to get equivalent SSIM results. This definitely isn't what I was expecting. In fact the only time I'm seeing comparable or better results is at high CRF values that I wouldn't normally use.

Has anyone else seeing this kind of behavior? I saw a report that Netflix did some tests that disagree with what I'm seeing. In fact, they're saying X265 can compete with X264 at half the bitrate. Only at very-low bitrates, maybe?

Also -- at what value does SSIM reach a point that is generally considered "acceptable"? (I know this is really subjective and probably changes based on resolution, but I'm hoping that there's a consensus somewhere).

I think I'm going to run some tests using fixed bitrates rather than CRF and see what happens.

I think that we have to take CRF out of the equation for comparing x264 with x264 because the CRF are not really calibrated (aligned) between the two encoders. So SSIM (or similar) and file size are probably the only objective measures for comparison (apart from subjective visual impression, and encoding speed if this matters).
There exist some threads in doom9 comparing the two encoders, and if I remember this correctly a sort of general consensus was (or still is?) that x265 shines at low bitrates (where x264 collapses), and at high resolutions (4k), but the savings in bitrate of 1:2 as a rule of thumb seems still very ambitious. It's certainly worth to follow the discussion in the relevant threads.
I haven't been doing much testing with x265.

[jdobbs]

Originally Posted by Sharc

I think that we have to take CRF out of the equation for comparing x264 with x264 because the CRF are not really calibrated (aligned) between the two encoders. So SSIM (or similar) and file size are probably the only objective measures for comparison (apart from subjective visual impression, and encoding speed if this matters).
There exist some threads in doom9 comparing the two encoders, and if I remember this correctly a sort of general consensus was (or still is?) that x265 shines at low bitrates (where x264 collapses), and at high resolutions (4k), but the savings in bitrate of 1:2 as a rule of thumb seems still very ambitious. It's certainly worth to follow the discussion in the relevant threads.
I haven't been doing much testing with x265.

Yeah, I'm aware that the CRF values aren't aligned. But the surprising part is that to get the same SSIM result, the file size of the X265 encode is sometimes larger than that of X264 (especially when the SSIM value is larger than .99xxx). I never would have expected that.