[opendtv] Re: 20040921 Twang's Tuesday Tribune (Mark's Monday Memo)
- From: Craig Birkmaier <craig@xxxxxxxxx>
- To: opendtv@xxxxxxxxxxxxx
- Date: Thu, 23 Sep 2004 11:51:06 -0400
At 11:13 AM -0400 9/22/04, Manfredi, Albert E wrote:
>The Lechner distance analysis was particularly interesting
>to me, and seems to reflect my reality too. That is, with
>a 25" 4:3 set, at a viewing distance somewhat less than I
>have at home, 480 scanning lines are perceptible. At
>least, with interlace.
Glad you added the part about interlace. It is appropriate to
consider interlace like a filter that is overlaid onto an otherwise
sharp picture, Unfortunately, this filter causes a reduction in both
H & V detail, and the ADDITION of picture impairments that impede the
viewers ability to perceive the available information in the picture.
>
>And the effect of a 31" 16:9 set is the same. A 30" wide
>screen set seems subjectively about the same size as a
>25" 4:3, although of course more appealing because it's
>wide more like a movie screen.
True if the set is interlaced. As noted by another post, the vertical
resolution is identical to the 4:3 set since the number of lines and
picture heights are virtually identical.
>As to "why buy an HDTV set," my answer would be simple.
>If you want to watch TV on a screen bigger than 25" (or
>31" wide screen), which is by no means large set by
>today's standards, you need HDTV to avoid the "bigger but
>proportionally fuzzier too" picture effect. Which is what
>I always notice with anything larger than 20" in an NTSC
>set, at approximately Lechner distance.
You are on the right track here but still are missing the "big picture."
To maintain a constant level of perceived sharpness as the screen
size increases (at a fixed viewing distance), requires more picture
information (high frequency details). One could come up with the
numbers for a continuously sliding scale, but this ignores practical
realities. While it IS possible to send image information in
frequency sub-bands, which can be combined as needed, the practical
solution - seems to be - to send enough information for the
application. It is useful to note that JPEG 2000 is taking the
sub-band approach, and it is being adopted by Hollywood for high
quality theatrical presentations.
So as a starting point, any claim about the need for "HDTV"
resolution must be justified in terms of how much resolution is
needed for a given screen size at the nominal designed viewing
distance. We have been through these numbers before, but they are
worth repeating.
EDTV nearly doubles vertical resolution (versus interlace), and has
the potential to SIGNIFICANTLY increase horizontal resolution as
well. A progressive raster of 854 x 480 or 1024 x 576 is a
substantial improvement over 704 x 480 or 704 x 576. In fact, it is
substantial enough to deliver a sharp picture on screens with
diagonals up to 40 inches. A 1280 x 720 raster can have sufficient
detail to provide a sharp picture on screens of up to 100 inch
diagonal.
These statements are based on the well established reality that
viewers will choose a viewing distance that is appropriate for a
given screen size, and that there are practical limits on what that
viewing distance can be - this is where the Lechner Distance comes
in. Thus, as screen size increases, the viewing distance (measured
in picture heights) typically decreases.
The practical take-a-way here is that screens smaller than 30 inch
diagonal are typically viewed at 7-9 picture heights. Screens between
30 and 60 inches are typically viewed at 4-6 picture heights, Only
VERY LARGE screens are viewed at 3-5 picture heights - most homes are
not large enough to allow this.
>Perhaps EDTV would also solve the problem, just above
>that threshold, but TV transmissions that want to get
>beyond NTSC or SDTV quality might as well jump right to
>HDTV.
Here we completely disagree. EDTV quality is more than adequate for
the vast majority of viewers. Only those with sets larger than 40
inch diagonal need HDTV. And. it has been demonstrated MANY times
that a good EDTV quality program can be upconverted for presentation
at HDTV quality with very little perceptible difference. The most
important factor is not the original source image resolution, but the
quality of the samples delivered to the receiver. It can easily be
demonstrated that the quality of an EDTV bitstream can be
significantly BETTER than an HD bitstream that is over compressed.
Ultimately it comes down to a question of bandwidth. A trade-off
between choice (quantity) versus quality.
Regards
Craig
----------------------------------------------------------------------
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:
