Ask questions to Sony Pictures Entertainment insider "paidgeek"

[Penton-Man]

Originally Posted by Objectivity

Also, since things seem to be getting a bit nitpicky and defensive, I just wanted to say thank you for all you do here.

Perhaps I missed some posts that were deleted but, for some perspective, I’m always amused when a forum member *debates* a person who chairs one of the ASC
(American Society of Cinematographers) Technology Sub Committees and questions the judgment of someone else that chairs another ASC Sub Committee directly involving film transfers. (and no, I’m not referring to you Patrick)

Everyone makes mistakes in their profession but, most often I’ve found those of top professionals to be the result of just too much stuff on one’s plate and the need to prioritize which could result in an oversight. In the case of one of the most valuable assets (LoA) in a film studio’s library, rest assured that this title received the most careful consideration and work.

[paidgeek]

Originally Posted by cjamescook

Sir Geek-

Not to beat this film title to death, but concerning Across the Universe:

- I'm sure picture quality is wonderful

- I would like to (politely) note that I would have loved to see earlier cuts of this film. Penton (and probably Hollywood) indicated it went through a number of edits. I'm not sure how many of the edits were for reduced running time as they were for content, but as this is I would classify as an "art film", I would have loved to see and understand the earlier edits.

In this case, while seamless edits might have been put to good use, if the edits were many and small, it might have been a good idea to offer the viewer the option of choosing seamless vs. pauses (and hit continue) at each edit point.

So, to end this with a question, is there a reason you can share why seamless edits of discarded material (or even bonus tracks) were not offered? Reviews have noted only one or two minor deleted bonus scenes, but I thought there would be much more.

Thanks in advance,
-Jim

The best way I can think to answer this is that as a studio, we distribute a film, and sometimes we have input on it creatively, but it really belongs to the film maker. As a rule, we can't re-cut and distribute a film without the blessing of the director.

The only way I can see us creating a title with some edit decisions left to the viewer is if the title is created with that goal in mind from the beginning and the director is involved.

[Penton-Man]

Originally Posted by paidgeek

On reflection, my earlier response is unnecessarily sarcastic........................

Baloney, you’re too nice.
If I had my way and some people on some other forums are already positioning to turn LoA into the next Bram Stoker’s Dracula, I wouldn’t release the damn thing on Blu-ray for years.

[blitz6speed]

Originally Posted by Penton-Man

In the case of one of the most valuable assets (LoA) in a film studio’s library, rest assured that this title received the most careful consideration and work.

I understand what you are saying, but if you've seen the 1080p25 high bitrate AVC cap of LoA from europe, and then compare it to the HDNet one that aired which is lowbitrate MPEG2 (that is using the Blu-ray master), theres HARDLY any difference, and in fact, the euro cap looks better (due to the low bitrate mpeg2 awesomeness). Im sure a BD will look quite good using that master, but i dont think its going to be the improvement we were expecting, to be honest.

[TheRealBob]

Originally Posted by Penton-Man

Baloney, you’re too nice.
If I had my way and some people on some other forums are already positioning to turn LoA into the next Bram Stoker’s Dracula, I wouldn’t release the damn thing on Blu-ray for years.

Originally Posted by blitz6speed

I understand what you are saying, but if you've seen the 1080p25 high bitrate AVC cap of LoA from europe, and then compare it to the HDNet one that aired which is lowbitrate MPEG2 (that is using the Blu-ray master), theres HARDLY any difference, and in fact, the euro cap looks better (due to the low bitrate mpeg2 awesomeness). Im sure a BD will look quite good using that master, but i dont think its going to be the improvement we were expecting, to be honest.

Well, there goes Lawrence of Arabia for a few years...

[dialog_gvf]

Originally Posted by paidgeek

Increases in sample rate allow for proportional increases in frequency range and a reduction in filter sharpness, but for distribution, 48k is all that I need as a consumer.

And by some theories, 48K is all you need as a human being (Nyquist-Shannon: 24,000 Hz). So, why do some insist on 96K and even 192K? What's in there that we can perceive?

I would have thought the bit budget is better served boosting the bit depth, rather than increasing the sample rate.

Gary

[HeavyHitter]

Originally Posted by Penton-Man

Baloney, you’re too nice.
If I had my way and some people on some other forums are already positioning to turn LoA into the next Bram Stoker’s Dracula, I wouldn’t release the damn thing on Blu-ray for years.

Not to sidetrack, but would you believe me if I told you Dave Mack and Xylon are still discussing Bram Stroker's Dracula over at AVS?

[mhafner]

Originally Posted by paidgeek

The film was scanned as opposed to telecined from a 65mm IP. Do you know of a post facility that can handle that format and get a better result?

That depends on when it was scanned and on what equipment. If it was done in the 90s then yes, today's houses with new scanners built specifically for 65mm will do a significantly better job with far more high frequency content and far less noise from the scanner itself (and no EE of course). So, are we talking about a new scan from the last 2-3 years or the one made in the 90s?
I heard about a new 6K scan in the works since 2001. Can you confirm that?
All HD I have seen so far of LOA is definitely supbar and it's not the compression that is the primary cause.
This link suggests LOA has detail upto 6K:
http://www.jts2004.org/english/proceedings/Galt.html
So the 1080p should at least sometimes have full 1080p resolution. The HD I have seen so far has barely 720p detail, has EE and strange noise patterns. Why is that?

[mhafner]

Originally Posted by paidgeek

What I take issue with in your post is the theory that a 1080p image that is displayed at 1080p where no filter was present in the signal processing chain will produce a less accurate result than using oversampling/filtering. When you have a digital system with a single pixel of information being carried from capture to display, then filter theory does not apply because... wait for it... there are no filters.

We agree that if you have a 1080p source and a 1080p display we want 1:1 pixel mapping and no resampling of any kind (as unnecessary as it is it's still done sometimes for the dumbest of reasons (overscan...)).
But that was not my point. The question was if it's better to sample with 1080p sensor to get 1080p output or with a sensor that is >> 1080p followed by digital downsampling to 1080p. The latter is far better as it allows optimal tradeoff between sharpness/detail and aliasing while the former must optically low pass filter before the sensor losing fine detail that need not be lost (or have severe aliasing when the optical filter is too weak).

[TheLion]

Originally Posted by mhafner

We agree that if you have a 1080p source and a 1080p display we want 1:1 pixel mapping and no resampling of any kind (as unnecessary as it is it's still done sometimes for the dumbest of reasons (overscan...)).
But that was not my point. The question was if it's better to sample with 1080p sensor to get 1080p output or with a sensor that is >> 1080p followed by digital downsampling to 1080p. The latter is far better as it allows optimal tradeoff between sharpness/detail and aliasing while the former must optically low pass filter before the sensor losing fine detail that need not be lost (or have severe aliasing when the optical filter is too weak).

This is exactly my point as well. Michel, we are on the same page as usual...

I sure acknowledge that paidgeek has more experience and knowledge about these matters than I could ever achieve in my lifetime.

That being said it is quite obvious from my point of view that any given 2K sensor/scanning captures a NOMINAL resolution of 1080p, but depending on many factors a significantly lower REAL resolution.

The sensor itself acts as a filter ("optical low pass").

Following this logic to achieve "real" 1:1 pixel mapping (original source sample of information = pixel of output device) you "have to" scan at much higher resolution (eg. 4k) in order to "make up for lost resolution in this process" and capture REAL picture information that's even close to 2K.

Pixels are NOT resolution!

To quote Michel's link:

"So what is the answer to the question, how many pixels in “Lawrence Of Arabia”?

From the MTF characteristic of the negative it would appear that digitizing the negative at 6K pixels horizontally would capture all useful information. Considering the negative aspect ratio this would result in approximately 16 megapixels per color or 48 megapixels per frame. Running time is approximately 227 minutes. At 1440 frames per minute this would suggest about 16 million megapixels would be sufficient!"

I haven't seen a LoA 70mm film copy in years now. All in all I have seen pretty bad copies of it as well as really good ones - sadly I never had the opportunity to project one of the latter myself...

But I think I am quite familiar with the potential "level of detail and definition" of this particular movie. All I am saying is that my capture of the recent LoA airing is nowhere near this potential - and this purely is an aspect of the used master as I am already considering the influence of the sub-par low-bitrate encoding in this assessment.

The other aspect is that we are talking about compressed video on a 4:2:0 8bit delivery format. So real 1080p pixel mapping is out of the question anyway...

I fear that releasing a BD based on this transfer could be a disappointment for many enthusiasts.

Paidgeek asked me to "compare" it to other recent 70mm transfers and mentioned "Grand Prix". After watching the HD-DVD release on a state of the art front projection setup a couple of times now I think this release is "quite good". For my tastes it just has this distinct "Warner VC-1 look" - too clean, too smooth. I would wish for way more high frequency detail (including grain) to be preserved - this is the one area this release leaves something to be desired IMHO. But it sure is much better than any LoA HD presentation I have seen yet.

All in all I must say that SPHE presently does an outstanding job title after title (the recent Gattaca release being the latest example) and for the time being I remain confident that they treat one of their most valuable assets the way it deserves. If it takes another 2-3 years I am not just fine - I am happy with that. I just hope that the first ( ) LoA BD release will not merely be a high-bitrate encoding of the recently used transfer put on disc.

[patrick99]

Originally Posted by paidgeek

There is always a possibility, but it will depend on many factors.

I am glad you are a Blu-ray fan, because with HD-DVD, the ceiling is 30.

The bandwidth and disc capacity advantages were of course the main reasons for my vigorous support of Blu-ray over HD DVD. Anyone who is familiar with my posting history at AVS can testify that I have been focusing on the benefits of higher bitrates for a long time now, and that my crusade against low bitrate dual format encodes was probably tiresome to many.

I know that obtaining customer/consumer feedback is one of the reasons for your participation here, so many thanks for giving me the chance to provide this feedback on the perspective of one customer/consumer on the benefits to be gained from using bitrates in the thirties.

[patrick99]

Originally Posted by Penton-Man

Perhaps I missed some posts that were deleted but, for some perspective, I’m always amused when a forum member *debates* a person who chairs one of the ASC
(American Society of Cinematographers) Technology Sub Committees and questions the judgment of someone else that chairs another ASC Sub Committee directly involving film transfers. (and no, I’m not referring to you Patrick)

Everyone makes mistakes in their profession but, most often I’ve found those of top professionals to be the result of just too much stuff on one’s plate and the need to prioritize which could result in an oversight. In the case of one of the most valuable assets (LoA) in a film studio’s library, rest assured that this title received the most careful consideration and work.

Thanks, Penton!

[Esox50]

Originally Posted by Penton-Man

Baloney, you’re too nice.
If I had my way and some people on some other forums are already positioning to turn LoA into the next Bram Stoker’s Dracula, I wouldn’t release the damn thing on Blu-ray for years.

You'd really do that to poor Rob?

[WriteSimply]

I'll try not to hijack the thread too much. But hijacking is still hijacking.

Originally Posted by TheLion

That being said it is quite obvious from my point of view that any given 2K sensor/scanning captures a NOMINAL resolution of 1080p, but depending on many factors a significantly lower REAL resolution.

The sensor itself acts as a filter ("optical low pass").

Following this logic to achieve "real" 1:1 pixel mapping (original source sample of information = pixel of output device) you "have to" scan at much higher resolution (eg. 4k) in order to "make up for lost resolution in this process" and capture REAL picture information that's even close to 2K.

Pixels are NOT resolution!

I understand this. I also understand that a 1080p scan is an optical low scan, which in itself is a filter.

However, for a 4K/6K scan to be downsampled to 2K/1080p would require the use of an algorithm, which in itself is a filter. Not only that, the algorithm can be changed to tweak the picture to suit the needs for the studio, which makes it the use of filters. Direct 2K/1080p scan, 1 filter - optical. 4K/6K scan, multiple filters - algorithms.

The big question is could a significant number of people tell the difference between the two? And since paidgeek already gave this answer, is this going to be a topic conducted among professionals or professional vs amateur?

Originally Posted by mhafner

But that was not my point. The question was if it's better to sample with 1080p sensor to get 1080p output or with a sensor that is >> 1080p followed by digital downsampling to 1080p. The latter is far better as it allows optimal tradeoff between sharpness/detail and aliasing while the former must optically low pass filter before the sensor losing fine detail that need not be lost (or have severe aliasing when the optical filter is too weak).

Like I said, the algorithms become the filters. Would you apply different algorithms for different shots or one for the entire movie?

Since this started because of comparisons between different LOA broadcast versions, I suggest we put this to rest for now since none of us have seen the 1080p master or the BD version.

paidgeek: I watched Coppola's Dracula the other day (with commentary on) and I noticed that halfway through the film in the shot of the castle, there were scratches on the print - same shot at two different spots. Just thought you'd like to know before you do final archiving on it.


fuad

[fparais]

Hi Paid

Is there any plan to release "The Tango Lesson" in blu-ray. I think it was release by Sony Classic Film.

Thank you

[Penton-Man]

Originally Posted by HeavyHitter

Not to sidetrack, but would you believe me if I told you Dave Mack and Xylon are still discussing Bram Stroker's Dracula over at AVS?

Unbelievable, just unbelievable.
I think that the only thing that will ever convince some of those people that no screw-up occurred with the production of that disc is to invite them to the SP lot for a special screening of the answer print with a monitor set up in the aisle running the HD master tape with a synched side-by-side display running the Blu-ray edition.

And in the end what will that prove?...... that there was no technological misstep regarding that production…… but the hounds will then go after F.F.C. for his creative choice in using that answer print as the original source for the Blu-ray disc. My goodness, they (well, one member in particular) nearly drove Robert Harris bonkers on the Home Theater Forum with his *debate* of that Blu-ray title-- until the administrator and other moderation stepped in, essentially saying *enough already*.

And he (Bob) had no vested interest in endorsing that Blu-ray title whatsoever.

[dialog_gvf]

Originally Posted by fparais

Hi Paid

Is there any plan to release "The Tango Lesson" in blu-ray. I think it was release by Sony Classic Film.

Thank you

We request members to refrain from asking the insiders about the release plans of specific titles.

If information could be released, it would be.

Now, what you can ask is whether paid has seen the master and what he thought of it. But, the actual release plans are another matter entirely.

Gary

[darkedgex]

paid,

Have you ever considered starting a blog? It might make for a more organized way of communicating with Blu-ray supporters/users at large, while still allowing people to provide feedback (via comments).

Same thing for you Penton.

[Chris Beveridge]

Not blu-ray related but Sony related;

why oh why was the Blood+ season 1 box set release dubtitled? After so many good anime releases getting a proper treatment when they're theatrical movies, why not love for the TV series properties as well?

Seriously disappointed in Sony over this.

[fparais]

Hi paid

Did you see the "Tango Lesson" master. Do you think it it gonna look great in blu-ray ?

Thank you