Originally Posted by jsaliga
Forgive me if this has already been asked. This is directed mainly at Amir but any insider who can answer this is welcome to do so.
Indeed I think others may be able to add as much to the answer as I can.
There is currently a discussion taking place in one of the digital projector forums:http://www.avsforum.com/avs-vb/showt...&&#post9258509
The above post quotes a four year old thread that in part discusses the actual resolvable detail in D5 master tapes, and the number that was kicked around was about 800~1300 lines of horizontal resolution. I think that number came from a FAQ by Joe Kane, who never participated directly in that discussion. I don't think the question was ever settled definitively so I thought I would bring it here to the insiders.
Indeed, I also believe the original assertion came from Joe as Gary Reber asked me about the same thing a couple of years ago, in an interview for WSR regarding WMV-HD discs, attributing this claim to Joe.
Making obvious allowances for a wide variety of filming techniques that would impact how much apparent detail we see in the film sources we have, just how much horizontal resolution is really in the HD we are getting in HD DVD and Blu-Ray? Is it really 800~1300 lines or is it higher? If D5 master tapes only go out to 1300 lines then what is the limiting factor? Is it the D5 system, optics, or something else? If it is not a limitation of the master then under ideal conditions how much horizontal resolution are getting in these new HD formats?
First, let me start by saying that the link points to a high quality discussion
. I expected a bunch of guess by laymen but the quoted sections do touch on issues that could really be there. I don't want to pester Stacey over the holidays but can do so once the new year comes, to see what he had found out. So here, I can only speak to my own experience, not scientific tests which they may have run. Also note that it has been a decade since I have been involved in Telecine equipment so others may have more current experience.
So back to your question. Let's work backward through this chain to see if we can spot where the problem spots may be.
Both HD DVD and BD encode content at true 1920 horizontal resolution. The codecs are given that resolution and barring any filtering that goes on in the encoder or player, that is what you are going to get. We know that both formats produce this resolution using test signals in VC-1 when driving 1080 displays using actual discs authored this way. So there is absolutely not the limited factor.
Next in line (putting aside DI and animations as the thread rightly mentions doesn't have this problem) is the D-5. The D-5 tape has true 1080p resolution in both axis. Indeed, its chroma (color) bandwidth) is double what we need (4:2:2 format versus 4:2:0 we use in BD/HD DVD). While not a subject of discussion, D-5 is also a 10-bit format, providing better dynamic range than what we need, assume the scan was done at that depth (a very good assumption). There is mild DCT compression used in D-5, and more so when 10-bit mode is used. That could disturb fine texture on difficult material (something without a lot of easy segments to encode in a frame). But those would be more like artifacts, and not show up as a frequency roll off seen on a spectrum analyzer. Indeed, the compression artifacts create artificial edges which tend to push the high frequency components (false resolution) than lower it.
Next is the telecine transfer. Here we are talking about an optical scan of the film at high speed with an anti-aliasing filter required for digital recording. There are a number of telecine machines but the typical Spirit Telecine with its 1920 Luma CCD will resolve that much bandwidth, subject to roll off from its anti-aliasing filter. Put in English, the filter is not going to have a brick-wall response so it is liable to roll off high frequencies some, require some sharpening to compensate (which can create some artifacts but not frequency cut off).
Higher resolution telecine equipment exists (e.g. scanning spot CRT) to scan at higher resolution but again, the thread acknowledges that 4K scans and such, would not have this issue so we won't go there. But suffice it to say, I hope the future is 4K scans as it solves many problems, including the filtering one just mentioned. Indeed, at the Home Theater Cruise last year, Joe said experiments he had done using 4K scans, showed that even DVDs had a better picture quality when using higher resolution scans than 1080p. He is now pushing to test 6K scans! OK, so the man really cares about picture quality
Now, just because the equipment scans at that resolution, it doesn't mean the output will be at that resolution. Film is notorious for curling and it must be held flat for it to properly scan so operator care and equipment maintenance is very important to achieve the rated resolution. But I suspect the post houses doing scans for major studios have high enough skill and operational practices to avoid issues here. But someone else needs to confirm it that is closer to it than I.
Next in line is the film itself. I know from my photography world that standard 35mm film has resolution far in excess of HD formats we are talking about here. While people will argue about the cross over point between film and digital (there are a lot more to picture quality than just pure resolution) I am sure we can all agree that 35mm film has a minimum of 3X to 5X more resolution than 1080p. There is a reason Casablanca looks so good after 60 years in VC-1/HD DVD
. So we are good to go as far as the recording medium is concerned.
Now we get to where the film came from. If it is printed from other sources such as digital creations, then we have another link in the chain as that equipment could cause roll off but I really doubt it. if the film is copies of other films, then there could be some issues here. Alas, I am not at all an expert in this area and others surely would need to chime in to provide comfort, or make us concerned about these steps. And of course, whether any concerns relate to major studios producing films, as opposed to low budget productions.
So net, net, there is no reason for horizontal resolution limit here at the technological level that I can see. But others need to chime in to confirm the above with more hands on experience than I.
What I can say is that in practical experience I do not agree at all with the final assertion that a 720p display resolves all that there is in HD DVD (and presumably BD although my tests do not include that format).
When HD DVD titles first came out, I compared the picture on a 1080p LCD flat panel (Sony Qualia) and 720p HD Plasma (Panasonic). There was no doubt that the LCD had higher resolution, especially in static scenes (the LCD smears detail in high motion images due to slow response time although this is much improved these days). The Penny was definitely softer although it had other advantages over the LCD which we won't go into here. But there was no question that there was more on the source than a 720p display can resolve. Note that I was viewing them from 1 foot away so this comment is not representative of people watching these sets at farther distances. But since we are talking about the resolution of the format, this should not be a point of contention.
In addition, we have tested the new 1080p Marantz DLP projector against the Sony Ruby using HD DVD. As you may know, the Ruby's lens does not have high enough quality, suffering from lateral CA (Chromatic Aberrations - not all the colors focusing on the same plane, causing some to bleed) and poor focus control, wasting its potential to truly resolve 1080p signals. We see this in test sequences where full Nyquist limit (1920) is not resolved on the Ruby. The Marantz however, fully resolves 1080p and as such, has a noticeably sharper image than the Ruby. This is why Kevin Collins on my team, who outfitted the HD DVD truck, insisted on using it there and frankly, scuffs at anything less in our HD DVD meets
. The new Sharp DLP btw, has the same resolution when we used it in our Dallas meet.
Please note that I am not trying to put down the Ruby as that is the display I own in my house. Just that we are able to resolve even the difference between two 1080p projectors with HD DVD so there is clearly information there on the disc or we would not be able to do so. At the same time, I would be remiss without mentioning that I believe there is some false sharpening that occurs with digital displays with good lenses in that the pixel edges are sharp compared to say, a CRT display, giving the illusion of sharper images than what is there. But again, that is the subject of a different topic and does not invalid the observation that a true 1080p set provides the sharpest images from HD DVD (and presumably BD).
So where does all of this leave us? I think the chain is fully capable of resolving more than 720p. While 720p displays show the best images they have ever produced with HD DVD/BD, they can not rival the resolution of a set with 1080p resolution when driven by these formats. Indeed, I highly recommend that people who are in the market for new displays, get a 1080p set and not settle for 720p. Newer projectors with this resolution are becoming very reasonable in price (cost of 100-200 HD DVD/BDs). Note that it is criminal to get a 1080p set without 1:1 pixel mapping as that clearly lops off a good 10 to 20% resolution from the source. So shame on display vendors who still do not provide this mode
. But even with overscan, these sets outperform 720p displays.
Thanks in advance for your answer.
You are welcome. I hope I didn't ramble too much
. Up early in the morning jetlagged with little to do.