On Sat, 27 Apr 2002 22:50:22 GMT, Loren Amelang
>On Sat, 27 Apr 2002 15:20:00 GMT, "Gunnar Thalin"
>
>..
>> > >And where on earth did you hear that Canon cameras
>interpolate
>> > >the fields? I just got my Canon MV30i (European
>version of Optura Pi)
>> > >yesterday, and you really scared me for a while.
>> > >But I just made some tests, and I can assure you
>you're wrong;
>> > >it does NOT interpolate the fields. Not spatially and not
>> > >temporally.
>> >
>> > Check out www.adamwilt.com for more on this -
>..
>> Well, I didn't find anything on that site, but it really doesn't
>> matter. My Canon clearly doesn't interpolate the fields.
>> It's so easy to test, it's ridiculous we're having this debate.
>> Maybe other Canon models do this but not mine.
>
>The best link explaining Canon's 3CCD "Frame Movie"
>mode:
>http://www.dv.com/magazine/2000/1100/wilt1100.html
>has apparently been obsoleted by the DVmag site.
>
>Google's cache of the text of the article, but not the
>graphics which were the really enlightening part, is still
>available at:
>http://www.google.com/search?
>q=cache:wZwrqVgaiakC:www.dv.com/magazine/2000/
>1100/wilt1100.html+wilt1100+dv&hl=en&ie=utf-8
>(Looks like that won't all highlight - you'll have to cut and
>paste it...)
>
>The Internet Archive "Wayback Machine" still has the
>graphics as well, at:
>http://web.archive.org/web/20020219112308/http://ww
>w.dv.com/magazine/2000/1100/wilt1100.html
>(But it only appears to work with Internet Explorer...)
>
>--> Adam - Any chance you can get DV to let you archive
>this _with_ the graphics on your site? It is still very relevant!
>
>The best current info (but without any graphics) is at:
>http://www.adamwilt.com/DV-FAQ-etc.html#filmlook
>
>So much for the Canon 3CCD models, which sort-of
>interpolate to get to "frame movie" mode.
>
> On Sat, 27 Apr 2002 16:27:32 GMT, lawrence@vex.net
>(Lawrence Kwan) wrote:
>..
>> This is not true for all Canon camcorders. What you have
>said only
>> applies to Canon 3CCD camcorder such as XL-1; it does
>not have true
>> progressive scan CCDs. However, Canon's single CCD
>camcorders such as
>> Optura, Optura Pi, Elura, Elura2 etc, does have true
>progressive scan CCD.
>> Absolutely no interpolation is used to produce the
>progressive scan video.
>
>I believe this is true. But...
>
>-----
>Display all headers
>From: bigrocketman3@webtv.net (Steve McDonald)
>Subject: Re: progressive scan (Peter)
>Date: Sun, 7 Oct 2001 18:56:39 -0700 (PDT)
>Organization: WebTV Subscriber
>
>There was an erroneous reference in Canon's original Elura
>specifications about the number of frames it shoots in
>progressive mode and this is still being repeated. This
>incorrect reference continues to cause confusion regarding
>later models with these same modes. ...
>
>In the switchable progressive mode that is present in many
>of Canon's DV camcorder models, 30 full-frame scans per
>second are made. There are not 60 full-frame scans being
>made each second and none of the scans are discarded.
>When these models are switched to the standard or movie
>mode, there are 60 interlaced fields being scanned per
>second which each contain every other line of the picture.
>The Sony DV models that have a progressive scanning
>mode, use only 15 full-frame scans per second. This 15 fps
>mode can produce good still pictures when they are
>captured later from the video recordings, but when played
>back in realtime, they appear jerky and are generally
>unusable for motion video.
>
>After the analog output from the CCD is processed and
>digitized, it is in an interlaced form for direct output or
>recording, regardless of whether the CCD acquired it in
>progressive or interlaced movie mode. These models can
>acquire an image on the CCD in progressive mode, but they
>can record and output it only as an interlaced signal.
>
>The advantage of these progressive scanning cameras is
>that they can acquire images from the optical source with
>almost no elapsed time between adjacent lines, thus
>eliminating the blurring or fluttering that can occur in still
>pictures derived from interlaced scanned recordings. With
>interlaced scanning, almost 1/60 of a second can pass
>between the acquisition on the CCD of adjacent lines and
>especially with fast moving subjects, some shifting in the
>image content between adjacent lines takes place. This
>has an adverse effect mostly on still images captured from
>video footage and doesn't affect the perception of a
>realtime display from normal rolling videotape, as this is
>compatible with the way our vision processes moving
>images. In fact, if it weren't for the complex and amazing
>way our visual system processes and cleans up aberrations
>in what our eyes take in, video and movies would not be
>workable.
>
>The main point is that the way that progressive scan
>cameras acquire an image is what's important to be able to
>later capture clear still images from the recordings. The
>interlaced mode of recording doesn't negate the effects of
>the progressive scanned CCD image. For this reason, the
>Sony system of using a progressive shutter with an
>interlaced scanning CCD, gives much the same effect as a
>camera with a CCD that scans progressively. The
>progressive shutter simultaneously exposes both sets of
>interlaced scanning lines on the CCD to the same optical
>image. Although these two sets of lines are scanned
>sequentially, the image on them was acquired at the same
>instant. The 2nd set of CCD lines retains its image until its
>turn comes to be scanned. This progressive shutter is
>mechanical, rather than like the electronic shutter used for
>rolling footage and it can be used only for still pictures, not
>for motion video. There is a big advantage to using a
>progressive scan CCD or a progressive shutter for
>producing still images, so these features are not wasted if
>you know how and when to use them.
>
>Steve McDonald
>-----
>
>So when
>
>-----
>>"Neuman - Ruether"
>message
>>news:3cc95089.1696897@newsstand.cit.cornell.edu...
>>> - PS-mode/"frame-mode" operates at 30fps for Canon
>cameras,
>>> 15fps for Sony, but the Canon interpolates fields to get
>>> the higher frame rate, reducing spatial resolution (which
>>> in any case cannot be greater than it is for interlaced
>>> video...)
>
>>Slightly incorrect. Motion is recorded at higher spatial
>>(but lower temporal) resolution in progressive mode than in
>>interlaced mode. Static video is the same for both modes.
>
>And where does that extra spatial resolution come from...?
>Sorry, you cannot exceed within a frame the resolution
>provided by the interlaced video, unless you are talking
>about stills and proper time alignment of the fields for
>these - but for motion, as I said, there cannot be increased
>spatial resolution with PS mode...
>-----
>
>I'll have to agree with him. You get exactly the same
>amount of information from Canon (1CCD) progressive scan
>as you do from regular movie mode. The only difference is
>whether 720X480 images are acquired every 1/30 (or
>1/25) second, or 720X240 images are acquired every 1/60
>(or 1/50) second.
>
>I guess you could say grabbing something every 1/60 sec
>is higher temporal resolution, and grabbing 720X480 all at
>once is higher spatial resolution. But once it is on tape or on
>your hard drive it is the exact same number of bits.
>
>To return to the original question (!), I suspect it is a simpler
>task for a video mode converter to make a high-definition
>image from the progressive scan capture, which gives it
>internally consistent 720X480 images to interpolate among.
>
>But it would not be out of the question to design a
>converter which recognized the interlaced fields and
>intelligently included the extra time information in its output.
>Reverse Telecine already splits one of every three
>interlaced frames into its component fields and divides the
>information between two output frames - it shouldn't be too
>hard to split and process every frame intelligently. Then you
>have the same amount of spatial data overall, but you have
>added more time resolution.
>
>Loren Amelang