I am extremely disappointed in the poor quality of your HDTV signal on 11-1 that resulted from your addition of the Create subchannel on 11-3. During your last fund drive you touted your commitment to bringing us, your members, the best in HD programming. Well with this picture quality your efforts will be in vain if the picture quality remains this poor.
Redundant at the end but hopefully we will get through to them. I had been feeling that with todays choices of programming through cable and satellite that Public Broadcasting was becoming irrelevant. With WTTW's HD programming I reconsidered my position. If they can't fix their PQ they will be irrelevant again. The bitch of it is that I just renewed my membership this past November.
The response I got from them:
http://alvyray.com/DigitalTV/Naming_Proposal.htm . ABC engineering also
As you've seen, WTTW-DT made a change over this past weekend to allow us
to carry an additional program stream on our digital channel. We
didn't do this without thoughtful consideration as to what effect, if
any, this would have on our 11-1, WTTWD HD service.
After careful examination we determined that we were not degrading our
signal by changing from the 1080i to 720p transmission format for HD
Below is a document written by one of the top TV Engineering
Professionals in Chicago, Bob Bruner, who is Engineer in Charge of our
Electronic Maintenance and Design Group. Bob goes through the process
and verifies in written form what we have seen visually.
The over-the-air standard for ATSC transmission is 19.3 MB/Sec. That
includes everything -- pictures, sound, PSIP, and any other meta-data.
The stations can allocate these bits in any way that they choose, the
only requirement being that at least one digital television signal has
to be in there somewhere. There is an additional requirement for a
graduated program of simulcasting with the matching analog station.
This was a misguided directive based on the idea that the simulcasting
would be of new HD programming carried by the analog station. The
framers of this directive didn't take into consideration that
simulcasting is a two-way street and that it is cheaper to up-convert
SD programming than to produce and then down-convert HD programming.
This simulcasting requirement is not a factor for most stations since
all they do is simulcast, carrying some of their shows in HD during
prime-time. In our case, however, our main HD program contains almost
no simulcast material. We met our simulcast obligation by reducing the
number of bits allocated to the HD service to 15 MB/sec, and creating a
3.9 MB/Sec standard definition service into which we fed the
programming for the analog station. That gave us 100% simulcast, end of
discussion. On January 9, we are going to add another standard
definition service, CREATE! from APT. To do so with the same quality
as our other services, we had to get somewhat over 3 MB/sec from
1080i pictures consists of a little over 1900 pixels per line X 1080
lines or about two mega pixels per frame. Multiply this times 30 fps
and you wind up with over 60 Mega Pixels/second. Each Pixel is
represented by a 10 bit byte for luminance, and to keep things simple,
another for Chroma. Actually, there are two bytes, one for R-Y and one
for B-Y, but they are half the resolution of the Luminance. Anyway,
after all the counting is over you come up with about 1.2 to 1.4
GB/second. WAY over 19 MB/Second. If you could put all the bits on the
transmitter into just one picture that would still require 65x
compression of the data. (1300/20 for rough numbers)
If we go through this same exercise for 720p, we get 720 lines X 1280
Pixels or 900,000 pixels per frame X 60 fps = 54 Mega Pixels per second
X 20 bits = 1 giga bit. You can see our base requirement for
compression is now just 50 or so. (1000/20)
So if we switch our HD transmission mode to 720p, we can reduce the bit
allocation by the required 3+ megabits to 11 or 12 and actually be
compressing the original picture information less than were with
1080i at 15 Mb. The argument with this practice would be that you are
giving the customer fewer pixels than formerly.
The 1080i signal consists of 1080 scan lines of about 1900 pixels. The
720P signal obviously has 720 scan lines of about 1280 pixels. This
number of pixels in each case was chosen to maintain square pixels.
Though 1080i has more scan lines, interlace has a figure of merit known
as Kell Factor, which describes a loss of resolution due to the
interlace. The Kell Factor is somewhere between 0.6 and 0.7. If you
multiply 1080 by a number in that range it comes very close to 720, so
the two systems have nearly identical vertical resolution. Few
consumer displays can resolve much more than 1200 pixels, and most
offer considerably less. The highest quality source material in the
station at present is the HDCam tape machine, which sub-samples the1900
pixel input to 1250 pixels and then remaps it to 1900 on output. So the
highest real resolution available right now is about 1200 pixels.
We conducted a small focus group test where editors, cameramen,
maintenance guys and others were asked to compare the two formats on
the same HD source material on identical side by side monitors, and not
only could most not tell there was any difference between the two, but
many picked the 720P picture as the best. So we are confident that
viewer reaction to our picture quality will not be affected by the
There is an interesting web site with a discussion of the relative
merits of 720p and 1080i at
has a good analysis of the trade offs at
. . . . I have glossed over the fact there are intermediate compression
stages that occur at various points in the process in all formats. The
camera or tape
format can sub-sample the chroma more, MPEG obviously squeezes the data
rate considerably and can produce artifacts on its own, and so forth. I
also did a lot of rounding. But the end to end numbers work out and the
concept is valid.
WLS and WFLD both operate at 720p because that is the format their
networks chose to adopt. There are many people who feel that
progressive scanning is inherently better than interlaced. Because the
progressive artifacts are cleaner -- in their opinion -- than the
interlace artifacts. For instance in interlace horizontal motion causes
'jaggies' in vertical lines and causes pixels on adjacent odd and even
scan lines to become uncorrelated. The fact that one field is fading
away as the other is scanned also causes some lack of correlation and a
loss of apparent resolution. The chief motion artifact of progressive
scanning is that the scan time is significantly long enough that
vertical lines become tilted.
Member and Viewer Service Department
5400 N St Louis Ave
Chicago IL 60625-4698
(773) 509-1111, option 6
email@example.com My response to them:
Thank you for your quick response. I have no problem with your going to 720p. The HD picture quality on WLS is very good, with WFLD being excellent. However, you are not even coming close to being very good with the signal you have been broadcasting since Sunday. If your picture quality was as good as WLS or WFLD I would not have written this note. I was hoping that it would be an interim problem resulting from your change over. I find it difficult to believe that you or one of your engineers has watched an over the air broadcast and feel the picture quality is as good as WLS or WFLD.
I know there are zealots out there that favor 720p or 1080i, I am not one of them. The picture quality I receive with my DirecTV receiver is as good or better than what I receive on my HD tuner - that was not the case prior to your changeover.