Sign In


User Group
Join date
Last activity

Post History

The GOUT Sync Thread

It’s not necessarily true that lossy audio tracks could not be conformed to a new standard that includes more frames, especially when those frames occur at reel transitions. One could simply split the ac3 track at the appropriate position and insert a silent ac3 frame before rejoining. Now, this would not be a perfect solution, as Dolby digital uses an audio frame rate of 31.25 fps, while the film runs at 23.976 FPS.

The result for a single missing frame is that the audio is off by roughly 0.232 frames, rather than a whole frame. This would arguably be a “good enough” compromise for many of the tracks. This would work especially well in the case of Star Wars as that bit of silent audio would not be noticeable for the affected reel transition. However, it is admittedly more problematic for the two missing frames in ROTJ, as those do not occur at a reel transition.

Analog Releases of Films That Contain Deleted, Extended, & Alternate Footage That've Never Been Released on DVD/BluRay

schorman13 said:

The US theatrical cut of “Yellow Submarine” has not been released in a digital format. It appears on VHS and LD only.

I think this says the Theatrical Cut was released on a German DVD.

No, that is just the same version released everywhere since 1999. It matches the UK version of the film as it premiered. Upon international release, the film was shortened, with some alternate footage during the final battle. That was the widely available version of the film before the 1999 restoration released on dvd at the time.

The GOUT Sync Thread

CatBus said:

oohteedee said:

CatBus said:

“The good thing about standards are there are so many to choose from!”

IMO GOUT (NTSC GOUT to be precise) meets the “good enough” standard, and I’m for sticking with it forever. Adding a frame that was never actually seen in theatres would break sync for all projects (not that it would matter for subtitles, but I’m thinking of others), for a pretty esoteric benefit.

There’s a better argument to be made for matching theatrical prints, since they have fewer frames than the GOUT, but it does not seem that they do so in a consistent manner, so there’s no standard that would apply to all prints. And losing those frames would also break sync for all projects for a pretty esoteric benefit.

Theatrical prints are missing frames because of projectionists cutting them off from time to time and wear and tear on the print. 4k77 is GOUT synced because the one missing frame is at the end of reel 5 for which we didn’t have a 35mm source. We would have had to use a BluRay frame to add it, so it was just left out.
Jedi GOUT is missing two frames from the middle of reel 3. We have “perfect” 35mm 4K scans for those two non GOUT frames. Those will be included in 4k83. It would be dumb IMHO to exclude 35mm theatrical frames to maintain a flawed video standard like the GOUT.

So essentially back to the PAL GOUT for Jedi again, then. We’ve already been there.

I’m fairly certain the intention is to use all known frames for 4K83, including the one frame unique to the NTSC GOUT.

Project Threepio (Star Wars OOT subtitles)

CatBus said:

These things change quickly, but this site says it’s on Amazon, but not Netflix:

EDIT: I don’t see anything but English subs though… iTunes also has it but I’m not clear on the subs there either.

Arabic subs are available on iTunes. I would be very surprised if they differ from the Blu Ray though. Let me know which films you want.

Phantom Menace '99 - HD Theatrical Version by Chewielewis

The Foobar decoder does outputs a true 5.1 stream, whereas the old winamp decoder would only allow you to decode a single channel at a time. In addition to that, the winamp filter did not decode the LFE and left it as part of the surround channels, as it was encoded. The Foobar decoder does use lowpass/highpass filtering to create the LFE channel and Surround channels, but but does not take into account that the hardware DTS decoders are set to attenuate the surround channels by 3dB. The result is that the surround channels are 3dB too hot which can lead to digital clipping in those channels during loud passages, and of course over emphasizes the surround channels in the mix. The solution is to simply declip and attenuate the surround channels after decoding.

When using the Winamp decoder, the Surround and LFE filtering must all be done in stages using software, while it’s mostly automated with Foobar, except for the 3dB gain eduction. The drawback is that the filtering used by the decoder is a bit of a black box. There’s no way to know the quality of the filtering being done, the cutoff being used or it’s steepness, or whether it follows the DTS white paper. It would probably be preferable to use higher quality software like iZotope Ozone to do the EQ filtering, but since the Winamp decoder has shown those audio problems, the method described seems to be the best possible solution.

Phantom Menace '99 - HD Theatrical Version by Chewielewis

I’ve synced all six of the Films’ theatrical DTS mixes. I’ve also shared that with Chewielewis. It syncs perfectly with the other two LD tracks posted.

Of note:

Hal9000 discovered, and I have verified that the old winamp APT-X100 decoder has a problem with decoding the front right and surround right channels. Basically, there is intermittent noise in those two channels throughout, which can be especially noticeable during quieter passages.

For this reason I’ve reconstructed each mix using the Foobar2000 decoder, which does have a different set of decoding issues. Basically, the surround channels are decoded 3dB too loud. This can cause clipping in the audio during louder portions, but I have been able to correct this using de-clip in Izotope RX.

Details on the process:

  1. Convert to 16bit/44.1kHz 5.1 wav with Foobar2000 apt-x100 DTS Decoder.

  2. Split to 6 mono wavs with ffmpeg.

  3. Pad each wav with 16 samples at the beginning using SoX to match decoder offset in the winamp decoder. (This is done to allow me to use my old Adobe Audition projects, originally created for use with the winamp decoder, in order to sync the tracks.)

  4. Edit Reels together in Adobe Audition and output full film as 6 mono wavs at 44.1kHz/16bit.

  5. Declip and attenuate the Ls and Rs channels by 3dB using Izotope Rx 6 Advanced (64Bit).

(For 24fps/44.1kHz/16bit)
6a. Dither Ls and Rs channels using Izotope Rx 6 Advanced (64bit) MBIT+ (noise shaping off, dither amount set to “Normal”, Auto-blanking on).

7a. Combine into a 5.1 channel wav using ffmpeg.

8a. Encode to 5.1 channel FLAC using eac3to.

9a. 5.1 Channel flac file @ 44.1kHz/16bit/24fps

(For 23.976fps/48kHz/24bit)
(Convert to 23.976 fps)

6b. Leave Ls and Rs channels at 32bit float, then, for all channels, interpret Sampling Rate as 44,056Hz using Izotope RX 6 Advanced (64Bit).

7b. Resample to 48kHz using Izotope RX 6 Advanced (64Bit) SRC (Steepness=64, Cutoff Shift=1.00, Pre-ringing=0.50).

8b. Dither to 24bit using Izotope RX 6 Advanced (64Bit) MBIT+ (noise shaping off, dither amount set to “Normal”, Auto-blanking on)

9b. Combine tracks into a 5.1 channel wav using ffmpeg.

10b. Encode to 5.1 channel FLAC using eac3to.

11b. Encode to DTS-HD MA using the official encoder suite.

12b. Strip superfluous 2 frames and DTS header by running through eac3to with -21ms delay. Ex: “eac3to -input.dtshd -output.dtshd -21ms”

13. 5.1 Channel flac file @ 48kHz/24bit/23.976fps

  1. 5.1 Channel DTS-HD MA file @ 48kHz/24bit/23.976fps
Preserving DTS LaserDisc tracks, specifically Jurassic Park

Buster D said:

Been trying to capture the DTS from the Japanese LD for “Hard Target”.

First recorded a minute of test audio from coaxial on a Pioneer X9 at 44.1 kHz, 16 bit wav using Sound Forge. BeSplit v0.9b6 had no trouble converting it with the following command:
BeSplit -core( -input c:\test16bit.wav -output H:\test16bit.dts -type dtswav -fix )

I then split the test16bit.dts file into 5 wav files using eac3to, and they all played fine without any distortion or anything. So far so good…

I went and captured the whole movie (settings unchanged), but when I tried to convert it to DTS, BeSplit stopped converting at about 45 minutes in (a few minutes before the side change). When trying to convert to 5 wav files, eac3to gave me this error:
This track is not clean.
libDcaDec reported the error “Invalid bitstream format”.
Aborted at file position 45350912.

I also tried converting the .wav to .dts using DTS Parser, but it stopped only about 10 minutes in (eac3to was able to convert this .dts file to 5 wavs, however).

Tried recording some tests at 24 bit and 32 bit float, and also tried recording using Audacity, but BeSplit wouldn’t convert any of these.

Anyone have any clues at to what could be wrong?

The issue your having is likely due to the differences in the frame headers for DTS on Laserdisc vs. DTS on DVD or BD.

Basically, you’ll need to run your wav through DTS parser, then correct the frame headers using a hex editor. After that you should have no problem decoding. This of course assumes you don’t have any ripping errors in the original file.

More detailed directions here:

Project Threepio (Star Wars OOT subtitles)

CatBus said:

Okay, I’m going to make the call. We have a new 4K preservation released out there and I want the current version of this project to support it better, even if not in a 4K-native form.

By this, I mean that the project won’t yet include 2160p subtitles, but it will include instructions for upscaling the 1080p subtitles to 2160p, which will give them their intended look on any 4k player.

So if anyone out there is working on translations, they will not make the cutoff for version 10.1 unless they’re done within about a week!

From what I’ve been able to determine from ripping my own UHD discs, they do not seem to use 2160p PGS streams. They’re 1080p like standard Blu Ray.

DTS Encoding - 16 bit?

Do rips of commercial discs read as 16bit? Just curious. You may also want to look for a program called LeeAudBi.exe for a more detailed look at what is in the frame headers. My guess is that the Surcode and eac3to encoders just don’t set the necessary bit in the frame header to describe it as a 16bit source. This source lists the contents of the frame header, and shows that there are three bits describing the source audio bit depth.

My guess is that your encoders just leave it set to 000 or 111, one of which probably corresponds to 24-bit. This is a total guess though. Remember that this is just a description of the source bit depth. Your other sources are correct in saying that bit depth is not really relevant to lossy compression codecs. The decoding of the audio is not necessarily effected, except that some decoders may try to match the bit depth of the decoded output to that of the source.

other info here: