[opendtv] Re: Pan-scan-zoom
- From: Tom Barry <trbarry@xxxxxxxxxxx>
- To: opendtv@xxxxxxxxxxxxx
- Date: Thu, 08 Oct 2009 06:40:59 -0400
Stessen, Jeroen wrote:
> Hello,
>
> Kilroy Hughes wrote that every source should be transmitted as-is, and the
> receiver should render it according to the capabilities of the display.
> I like the logic of his arguments. I would like to add one component, as
> I have proposed in patent US20040130664. Once it occurred to me that ATSC
> assumes that HD sources will be rendered on SD displays by down-sampling
> the entire image. The composition is preserved (unless the display is 4:3,
> of course), but the high resolution is lost. I was thinking that the whole
> purpose of HDTV is to put meaningful information in the details, and that
> these details are visible because of the larger displays and wide viewing
> angle. In other words: HD cameras can capture wider scenes, whereas SD
> cameras must sooner zoom in on the details. An original SD production will
> look different from a down-sampled HD production. This goes further than the
> cropping that occurs in pan-and-scan. Therefore I invented "pan-scan-zoom",
> where the source provides some metadata about the area of interest, whose
> details should be preserved. The SD receiver can (dynamically) zoom in on
> the details, instead of down-sampling the whole image. It is like having
> several virtual SD cameras on the scene. A method exists for down-sampling
> with a variable ratio and automatically adapting anti-aliasing filtering.
> It is explained in patents US5892695 and US6963890, and in the PhD thesis
> of Ad van den Enden (ISBN 9066746505). That is a solved problem.
>
> I would very much like to contribute my idea for display framing to a
> "standard for multi-standard". It is useful for small displays, far away
> displays, or even bad eyesight.
>
> PS, interlace is a form of "super-resolution", it can be your friend. ;-)
>
> Groeten,
> -- Jeroen
>
> Jeroen H. Stessen
> Specialist Picture Quality
>
> Philips Consumer Lifestyle
> Advanced Technology (Eindhoven)
> High Tech Campus 37 - room 8.042
> 5656 AE Eindhoven - Nederland
>
>
>
>
>
>
>> -----Original Message-----
>> From: opendtv-bounce@xxxxxxxxxxxxx [mailto:opendtv-bounce@xxxxxxxxxxxxx]
>> On Behalf Of Kilroy Hughes
>> Sent: Wednesday, October 07, 2009 17:24
>> To: opendtv@xxxxxxxxxxxxx
>> Subject: [opendtv] Re: 70th Anniversary Blu-ray and standard DVDs of the
>> Wizard of Oz
>>
>> I know you share genetic material with TV sets and see the problem as
>> adapting different content aspect ratios and frame rates to whatever the
>> most popular sets like, but I want to share with you the pain I live in
>> the world of "three screens and a cloud".
>>
>> In my world, not only is content all over the map, with more combinations
>> of picture shape, resolution, and frame rate than broadcast (but
>> thankfully less interlace and non-square samples) ... but devices and
>> viewing conditions are all over the map.
>>
>> The architecture that works best is encoding the actual picture height and
>> width and frame rate, and making the decode/display rendering device adapt
>> to its current display circumstances. That means you might want to watch
>> a 2.35AR movie on your iPod with a 1.5AR screen "center cut" because of
>> the small screen, but when you plug it into a TV connection going to a
>> 1.78AR display using 1.78AR signal format you might tell your pod to
>> letter box it with common sides or a slight side crop. If you had a built
>> in display that was 2.35AR you could display full screen/wide screen, but
>> most displays don't support wire formats that will take a 2.35AR signal,
>> so you have to pad to a 1.33 or 1.78 frame.
>>
>> The basic architecture that works is to encode the content with its
>> original source characteristics ... don't hide 24p content in 60i encoding
>> (e.g. Blu-ray vs. DVD), and don't hide 1.33, 1.85 - 2.35AR movies in
>> 1.78AR encoded frames. Then the device doing the decoding has all the
>> information it needs to frame the content for a directly connected
>> display, or a wire format that restricts the output options allowed for an
>> external display. In my world, we have to assume the same content file
>> will be viewed on lots of different devices and displays, so there's no
>> option to encode it as though every display is a 16:9 TV and let the
>> others suck (the Blu-ray AR problem: Blu-ray is locked to 16:9. DVD was
>> smarter because it was designed to handle a mixed 4:3 and 16:9 TV world
>> with adaptive framing in the device).
>>
>> We are going through a painful transition where there are conflicting
>> distribution architectures resulting in both rending devices and displays
>> fighting to control picture framing and refresh, resulting in postage
>> stamp pictures surrounded by letter boxes and pillar boxes and other
>> nonsense. Happily, the frame rate part of the problem is underway with
>> most modern displays supporting 24P wire formats and many even have a
>> "leave it alone" setting to avoid scaling up the image and cropping the
>> edges off to simulate a cathode ray tube with a bad power supply.
>> Interlace encoding is fading into the sunset very slowly along with the
>> dinosaurs. Now we need to make progress on picture framing in a way that
>> works in a world with an increasing variety of video devices, where the
>> display framing is done by the decoding device with accurate information
>> about the encoded picture and whatever wire format it might be currently
>> connected to (or not, with direct display control such as internal or
>> VGA multisync).
>>
>> Kilroy
>>
>
>
> The information contained in this message may be confidential and legally
> protected under applicable law. The message is intended solely for the
> addressee(s). If you are not the intended recipient, you are hereby notified
> that any use, forwarding, dissemination, or reproduction of this message is
> strictly prohibited and may be unlawful. If you are not the intended
> recipient, please contact the sender by return e-mail and destroy all copies
> of the original message.
>
Jeroen -
My own contribution to a solution from a few years ago, see:
<www.trbarry.com/Displaying_4x3_on_16x9.htm>.
The idea is you can get get away with only so much cropping, so much
stretching, so much warping. So you use only small amounts of each,
together, and then add smaller unobtrusive variable color curtains on
the sides that follow the color and brightness of the edges of the source.
I'm on my way out for a class right not and don't have time to read your
patent but I assume it would also be best to select the source area of
most interest before doing all the above and it would be nice if the
source specified that.
- Tom
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at
FreeLists.org
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.
Other related posts: