This is topic The DCP for American Beauty was dreadful in forum Digital Cinema Forum at Film-Tech Forum ARCHIVE.


To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=16;t=001449

---

Posted by Aaron Garman (Member # 1653) on 03-28-2013, 12:01 AM:
 
I just saw it as part of the Cinemark Classics series. The thing looked like an edge enhanced and sharpened Blu-ray that was blown up with a very weak audio mix. I saw this in 35mm a couple times in the same cinema (well one house over but roughly the same size) and it was glorious as I remember it, including the sound.

Anyone else catch this or have you run the DCP? It was great to see this again with an audience but wow, it was also very disappointing to see it with such a dreadful transfer.

It really did feel like video and not film. Too bad.

AJG
 

---

Posted by Brad Miller (Member # 2) on 03-28-2013, 02:31 AM:
 
Sadly, some of these repertory DCPs are sourced from the master used for the blu-ray. You can probably chalk up the poor audio to one of those spectacular near-field mixes too.
 

---

Posted by Bobby Henderson (Member # 840) on 03-28-2013, 08:26 AM:
 
American Beauty is one of those movies that had a noticeably flawed video transfer on Blu-ray -one bad enough to warrant a re-do like Dreamworks did with the Gladiator Blu-ray. Unfortunately American Beauty wasn't as big a hit as Gladiator so a re-do of the Blu-ray may never end up being done. It's a shame too. Any movie that wins an Oscar for cinematography ought to have a top tier transfer.

IIRC, Dreamworks also did a 2nd pressing of the Saving Private Ryan Blu-ray to solve an audio problem.

IMHO, movie studios should not be using video encodes for Blu-ray disc as a source for repertory DCPs. At least I hope that's not what they're doing. The DCI spec has some pretty important rules in how JPEG 2000 DCPs should be encoded. Video streams in AVC and VC1 break many of those rules in order to reduce file sizes. If you take a video stream from an already compressed video encode for Blu-ray and re-encode that again for JPEG 2000 DCP format the end result may actually end up looking worse than the Blu-ray. And the American Beauty Blu-ray is already compromised on its own.

DCPs should only be encoded from the highest quality master available. For Hollywood movies that's usually going to be a 2K or 4K uncompressed video master.

If the tech guys doing the video master go nuts with artificial sharpening and noise reduction it won't matter if the master is uncompressed and the DCP is running at a high bit rate. The end results are still going to look bad.
 

---

Posted by Manny Knowles (Member # 1171) on 03-28-2013, 10:16 AM:
 

quote: Bobby Henderson

---

Video streams in AVC and VC1 break many of those rules in order to reduce file sizes.

---

quote:

---

77GB total for both parts

---

quote: Aaron Garman

---

Let's hope Paramount had a great DCP for the Godfather in 2 weeks. I'd guess they will: the Blu-ray is stunning.

---

quote: Jock Blakley

---

THE BLUES BROTHERS - especially considering that was 1920x1080 and only had L/R sound... that drifted in and out of sync.

---

But F said the studios REALLY CARE!
 

---

Posted by Bobby Henderson (Member # 840) on 03-29-2013, 09:46 AM:
 
Y'know, this "DCP from a bad Blu-ray master" thing has me a little confused. What are the properties of the "master" source from which the bad DCP is derived?

Is the bad DCP source an uncompressed 2K video master? Modern movies have their DCPs and Blu-ray encodes sourced from the same master files produced from the digital intermediate. In the case of a repertory movie that pre-dated DI use, such as American Beauty, the HD video master could be something originally intended for DVD authoring. An older HD master with some telecine wobble, dust, specks and other issues was a good enough source for anamorphic DVD and low bit rate HD on cable. Such masters are not good enough for Blu-ray and big monitors with native HD resolution. And they're sure as hell not good enough for d-cinema. A modern film scan and digital intermediate treatment process is required.

Is the bad DCP source an already compressed AVC/VC1 1080p/24 video stream trans-coded over into JPEG 2000 format and slightly blown up to 2K? If this is the case someone just needs to hook up a Blu-ray player to the d-cinema system and play the damned Blu-ray disc. I really hope this scenario is not happening at all. Converting an already compressed video stream into another compressed video stream is only going to be stupid. Because you know what you get when you pile some shit on top of some other shit? You get even more shit. This scenario reminds me of the idiots who send me low resolution JPEG images for large format printing jobs and then artificially up-rez it in Photoshop (and save it as JPEG again) when I ask for something better. Just playing the Blu-ray disc would reduce the shit quotient down to a more modest level.
 

---

Posted by Bajsic Bojan (Member # 4928) on 03-29-2013, 11:57 AM:
 
Sadly, i would say that is the most common practice. Same source for Bluray, which they make worse (some claim it is better after the 1920->2048 conversion) by reencoding again (sometimes reencoding in 1920x1080).

When we get bluray (i dont like the format that much, the media is crap, the players also), i will if possible, rip it, demux, repack to mpeg mxf without reencoding. A server version is still much more reliable than a flimsy dust specked disk. (or just use a 100$ media player to do the job if in a hurry)
 

---

Posted by Martin McCaffery (Member # 37) on 03-29-2013, 12:44 PM:
 
So, the industry was lying to us when they said digital will be an improved viewing experience?
 

---

Posted by Bobby Henderson (Member # 840) on 03-29-2013, 02:58 PM:
 

quote: Bajsic Bojan

---

Sadly, i would say that is the most common practice. Same source for Bluray, which they make worse (some claim it is better after the 1920->2048 conversion) by reencoding again (sometimes reencoding in 1920x1080).

---

That's not really answering the question. If the DCP is only using the same uncompressed HD master source that was also used to encode video for Blu-ray then that doesn't make the DCP a "bad Blu-ray DCP." The Blu-ray format wouldn't have anything to do with it. The fault of the DCP being bad would either point to the uncompressed video master itself being faulty and/or the way the DCP was encoded being faulty. Blu-ray wouldn't have anything to do with it.

The only way Blu-ray would figure into the equation is if someone took the especially stupid leap of copying an already severely lossy compressed AVC/VC1 format video file made specifically for Blu-ray disc and then transcoded that compressed video file into JPEG 2000 for DCP use. Turning a Blu-ray video file into a DCP video file would gain absolutely nothing and actually make the resulting faked DCP look even worse than the Blu-ray. It's taking an already severely compressed video stream and compressing it again. Like I said, if this shit is happening it would be far better to play the retail Blu-ray disc than the even worse quality DCP that was derived directly from it. It doesn't matter how big you make the bit rate of that DCP. The extra bits won't restore any native imagery lost in the initial lossy encode step creating video files for Blu-ray. The combination of how the JPEG2000 format works and DCI rules on DCP authoring would greatly limit what could be done with encoding that Blu-ray derived DCP from hiding compression flaws. There are plenty of tools at the disposal of video authors to make 1080p/24 video in AVC and VC1 look good at low bit rates. There is not nearly as much latitude when making a JPEG2000 DCP.

I wouldn't put it past someone to commit such an act of digital stupidity. I see it all the damn time in graphics work. People will take a lower resolution JPEG image and try to fake it into being a high-res image. Oh you prefer lossless compressed TIFFs or Photoshop images? Well I'll just uprez this JPEG to a bigger size and save it as a TIFF or PSD!.

I would be appalled if this same kind of shit was happening on the movie distribution level. If it is some people need to be fired over it. It's bush league, bottom of the barrel quality standards of practice. Not professional at all.
 

---

Posted by Bajsic Bojan (Member # 4928) on 03-29-2013, 05:02 PM:
 
A) That is what i am saying. It happens that the DCP is a direct rip of the already encoded mpeg2 or similarly lossy file. Isn't that great And yes, the resulting 'DCP' then looks worse than the 'original' bluray file as it is aditionally resized and reencoded.

B) There are no "DCI Authoring rules" on video standards, so free for all. You can make a REAL DCP from a VHS source.
 

---

Posted by Jock Blakley (Member # 6552) on 03-29-2013, 07:19 PM:
 

quote: Bajsic Bojan

---

There are no "DCI Authoring rules" on video standards, so free for all. You can make a REAL DCP from a VHS source.

---

quote: Bajsic Bojan

---

It happens that the DCP is a direct rip of the already encoded mpeg2 or similarly lossy file. Isn't that great And yes, the resulting 'DCP' then looks worse than the 'original' bluray file as it is aditionally resized and reencoded.

---

quote: Bobby Henderson

---

No rules to authoring DCP's? I guess the spec sheets on dcimovies.com are just an academic waste of time?

---

quote: Bobby Henderson

---

If that's really the case I'd like to congratulate anyone doing that for being an absolute moron. There's NOTHING gained from ripping video from a Blu-ray and making a DCP out of it. It's a complete waste of time. Since there's an additional decode/re-encode factor thrown into it the Blu-ray disc will always look better with such a DCP hack. The theater might as well just play the retail Blu-ray disc regardless of what they think about the disc format.

---

Yes, welcome to capitalism. They can charge more for a '2k (!) DCP(!)' then for a BR. It is also 'BETTER QUALITY BECAUSE IT IS DCP'.

How many of todays distros have quality control that is more than a 24" full hd (!), maybe IPS or sometimes TN monitor (!!), doing QC from easyDCP (!!!). Or sometimes not even that (!!!!). Of course they won't see anything wrong with the product. It makes them money, its cheaper to get by than 35mm, they can do it at their home, they don't really care what happens afterwards.

If we have noticed 35mm prints being worse and worse over the past 5 years because of lack of QC, lack of interest in making an end product that looks good, even for a thing that cost many many $$$$, how much do you think they care for a bunch of 1's and 0's they get over a file sharing service for 10 times less $$$?

quote: Bobby Henderson

---

This sort of thing would do a great job of selling people on NEVER VISITING THE MOVIE THEATER and watching movies exclusively at home ON BLU-RAY if they like.

---

quote: Bobby Henderson

---

I think it should be mandatory for the picture and sound to be better than what can be delivered at home. DCPs have the capability of being a hell of a lot better than the audio and video on a Blu-ray disc. Movie studios need to be using that capability all the time.

---

But that costs money. Execs will just say 'WHY?' and be done with it. Digital is here to make things cheaper for them in all the stages. Production, post, audio, distribution. Quality has been deemed acceptable. By media, by the audiences and <GASP> by industry professionals. Not as good as 35mm you say? Pffft, its marvelous, no scratches and flicker.

And 35mm is only serving the cinema industry, while a file on a hard disk can easily be pushed to cinema, ht, broadcast, streaming, archives, mobile phones, ipods, ipads, rental companies,... who cares what happens to it. Who cares how this will get stored and if we dont have access to a movie from 2008 anymore. Who cares if it is compressed video and sound. Who cares if any company who gets the file can do with it as they like, put pauses inside, compresses it even more, does some fps changes, adds a banner or commercial or two. Who cares if it is projected over a 3D lens at 15% of the intended light. Who cares if it is projected in a 2.00:1 ratio, regardless if it is a scope or a flat feature. Who cares if the sound is not 7.1 PCM but a stereo AC3 decode.

It's all good.
 

---

Posted by Martin McCaffery (Member # 37) on 03-30-2013, 08:35 AM:
 

quote: Bajsic Bojan

---

And 35mm is only serving the cinema industry, while a file on a hard disk can easily be pushed to cinema, ht, broadcast, streaming, archives, mobile phones, ipods, ipads, rental companies,... who cares what happens to it.

---

quote: Brad Miller

---

They are using the MASTER that the bluray was created from, not an actual bluray to create the DCP. The DCPs are higher quality than the bluray, but in a case like this (and others, such as Ferris Bueller) the original TRANSFER that was used for both the bluray and the DCP sucked.

---

quote: Bajsic Bojan

---

Yes. You can scour the DCI specs for any rules about DCP source material and you will find nothing.

---

quote: DCI DCSS v12 with errata 2012-1010

---

4.4. Codestream Specification
All codestreams shall fully conform with [ISO 15444-1:2004/PDAM 1 (soon to be Amendment 1)], as
more fully constrained as follows:

• The capability parameter for a 2K distribution shall be Rsiz = 3, for a 4K distribution it shall be
Rsiz = 4.

• All image frames shall be untiled. More precisely, the entire image shall be encoded as a single
tile.

• The image and tile origins shall both be at (0, 0).

• There shall be no more than 5 wavelet transform levels for 2K content and no more than 6
wavelet transform levels for 4K content. There shall be no less than one wavelet transform level
for 4K content. Additionally, every color component of every frame of a distribution shall have
the same number of wavelet transform levels.

• Codeblocks shall be of size 32x32.

• The codeblock coding style shall be SPcod, SPcoc = 0b00000000.

• All precinct sizes at all resolutions shall be 256x256, except the lowest frequency subband, which
shall have a precinct size of 128x128.

• There shall be no region of interest, i.e., Region of interest (RGN) marker segments are
disallowed.

• Coding style Default (COD), Coding style Component (COC), Quantization Default (QCD), and
Quantization Component (QCC) marker segments shall appear only in the main header.

• Packed Packet headers, Main header (PPM) and Packed Packet headers, Tile-part header (PPT)
marker segments are forbidden.

---

quote: Bajsic Bojan

---

But that costs money. Execs will just say 'WHY?' and be done with it. Digital is here to make things cheaper for them in all the stages.

---

quote: Bobby Henderson

---

Those are some pretty specific rules. Rules that any AVC/VC1 video encode on Blu-ray would break. That PDF at dcimovies.com is full of rules. The rules are not only there to establish some kind of quality standard. They're also there more importantly to ensure (or at least improve) file compatibility with a variety of d-cinema hardware.

---

quote: Bobby Henderson

---

That's what I figured. But the way Bajsic is putting it anything goes with how Hollywood studios make their DCPs. Where is he getting his information?

---

quote: Andrew Thomas

---

Those are all specifications for the DCP itself. That doesn't specify the input source file going into the DCP.

---

quote:

---

As said, those have nothing to do if my source is a 320x240 flash video i converted to tiff images of 2048x858. I am DCI compliant!

---

quote: Bajsic Bojan

---

The reality is, DCPs are getting created at a much much more larger scale than only the five or six HW studios. The reality is that only a smaller part of those movies receive ANY quality control.

---

quote: Bobby Henderson

---

Those are specifications for how the video code stream in the DCP should be encoded from the uncompressed DCM (Digital Cinema Master).

Every consumer grade video format violates those rules. They have many levels of wavelet transforms, they heavily use inter-frame compression, every frame has dozens of tiles rather than the frame being just one tile. It goes on and on. Those consumer grade formats can't conform to the DCP JPEG 2000 encoding rules. They do so in return for much smaller file sizes and far more severe compression ratios.

---

quote: Bobby Henderson

---

...if this shit is happening...

---