Archived from groups: alt.tv.tech.hdtv (
More info?)
In article <Pmpje.7843$zd5.6565@twister.nyroc.rr.com>,
"Richard" <rfeirste@nycap.rr.com> writes:
>
> "Test" <mailtest@elcochevy.com> wrote in message
> news:mailtest-95D4D1.09413120052005@news.charter.net...
>> Anything?
>
> There will be no 1080p60 broadcasts in with current standards for broadcast.
> 1080p24 is possible some time down the road; but most HDTV sets will display
> it as 720p60 or 1080i60 anywho.
>
Because of MPEG2 compression that removes many of the redundancies,
after you convert the upconverted 24p film to 60Hz interlaced, then
encode the 60Hz interlaced video into MPEG2, then the redundancies
from the repeated fields are fairly efficiently encoded. Therefore,
the 60Hz interlaced 1080i does fairly closely reproduce the same
image as the 24p original. (Well, not quite, but it is a fairly
honest approximation.)
Much of the evil of 1080i30 (or i60, depending upon your method
of describing 1080line 60Hz interlaced) MPEG2 being squeezed into
the ATSC channel is due to the MPEG2 compression artifacts. These
artifacts occur when too many of the DCT coefficients are removed
or their precision is truncated.
When encoding the 24fps film into 60Hz interlaced video, there is
HUGE amount of redundancy that wouldn't normally occur from a fully
video originated signal (e.g. like an American football game that is
recorded in true 1080i.) The film has LOTS of repeated frames, and
MPEG2 encoding can take good advantage of those redundancies.
So, from an MPEG2 ARTIFACT standpoint, the 24fps film material can
look significantly less MPEG artifact prone than the true 60Hz video.
My guess as to the largest disadvantage of the 24fps film material
that is upconverted to 1080i 60Hz vs. the true 1080p24 is that there
will be SOME loss of vertical resolution that is natural when filtering*
interline twitter (an especially obnoxious artifact that mixes
temporal/spatial aliasing with the perception of flicker.)
*When filtering the interline twitter, a crude (perhaps 5 line)
vertical filter will cause some loss of vertical resolution. This
loss of resolution might be partially mitigated when rebuilding the
24fps frames from the 60Hz interlaced signal, but won't be perfect.
A DYNAMIC and INTELLIGENT interline filter (that doesn't remove detail
unless twitter is found to exist) can help to avoid the loss of
perceptable detail.
**The vertical filter can also be a component of normal MPEG2 artifacts
mitigation also... By doing some horizontal/vertical/diagonal filtering,
some of the information that MPEG2 encoding finds to be troublesome can
be removed. In essence, the vertical filter can be used to mitigate
interlace twitter (which exists only on interlaced material), but also
can be used to mitigate MPEG artifacts.
When 'filtering' is directly done by MPEG2 encoding, that filtering is
mostly done by truncating some DCT coefficients for various reasons. Such
a 'truncation' (either by removing precision or effectively total removal of
certain coefficients) is a NONLINEAR filtering process. This NONLINEAR
filtering can easily produce artificial detail that is outside of the nyquist
sampling limit. This can produce MORE aliasing-like artifacts (than normal
frequency domain filtering) by removal of detail, and can be very very ugly.
*This dct truncation-caused aliasing is one of the causes of the stairstep
or 'aliasing' like artifacts that were fairly common on early DV
camcorders.
So, by very careful filtering in the frequency domain (I mean, normal
linear filters) in the horizontal, vertical and various diagonal directions,
various excess DCT coefficients can be mitigated and the alias-prone
filtering based upon DCT coefficient truncation can be lessened. So
effective PREFILTERING of the video can help the MPEG2 (or DV) encoder
to do a better -- VISUALLY more ACCURATE -- job.
Back closer to the main subject -- it is likely to be advantageous
to do significant DYNAMIC vertical (and horizontal) filtering even
when the video isn't interlaced. It is true that interlace
might force the use of larger amounts of vertical filtering than otherwise
desired, but MPEG2 encoding does encourage the use of filters that might
seem to be superfluous.
Given the significant redundancy in the 24fps originated 1080/60Hz video,
there might be the need for less vertical filtering for the purpose
of optimizing MPEG2 encoding, but the interlace might push the use for
more vertical filtering anyway. True 60Hz interlaced video can be very
difficult due to the potential of much less redundancy (no built in guarantaeed
redundancy that film would provide) and also the interlace issues (both!!!)
John