Canadian TV, Computing and Home Theatre Forums banner

1 - 20 of 21 Posts

·
Premium Member
Joined
·
3,737 Posts
I'm curious: What was the context of that CBC document?
Is it a suggestion to broadcasters and regulatory bodies, or is it simply an observation on how the required signal strength for ATSC reception has been underestimated?
 

·
OTA Forum Moderator
Joined
·
24,867 Posts
The Marcoux Paper: IEEE Broadcast Symposium 2009

downbeat said:
What was the context of that CBC document?
The CBC has been a member of the ATSC almost from the beginning so they are often involved with such technical discussions about DTV. The context of that document is that there has been lots of discussion in the past year about the inadequacies of the formulae being used for DTV power allotments, mostly in the U.S.'s VHF-HI assignments. The process is presently under review on both sides of the border, with Pascal Marcoux's CBC document having been front and centre at a recent IEEE Broadcast Symposium: http://www.digitalhome.ca/forum/showpost.php?p=1020242&postcount=59
 

·
OTA Forum Moderator
Joined
·
24,867 Posts
I think Pascal Marcoux's "Revisiting the Field Strength Requirements for DTV in the Canadian Context" paper should be required reading at this point - the CBC technicians did real world RF Field testing of DTV signals and compared the figures inferred by standard formulae with actual measurements and discovered some common variances that are predictable. Along the way they tested various OTA signal gear devices for loss and came up with some important conclusions.

The significance of the Marcoux Paper is that it is now the basis of ongoing discussions in Canada and the U.S. regarding ERP level allotments for digital OTA stations. The broadcasters have been complaining about less-than-expected performance results, and the Marcoux Paper is their justification.

Keep in mind that the PDF document is in the form of a presentation slide set and not a technical paper.
 

·
Registered
Joined
·
3,396 Posts
Marcoux cited papers prepared by Bendov et. al. in 2001/2004 identifying various shortcomings
in the official FCC/NTIA Longley-Rice propagation prediction model (e.g. ILLR, FREE to all users):
http://www.tvantenna.tv/papers/PFactorsV.pdf
http://www.tvantenna.tv/papers/dtv coverage and service prediction.pdf
Among other factors, these papers calculated the additional desensitization due to antenna
VSWR degrading the EVM (Error Vector Magnitude, e.g. "eye pattern") in a digital signal.
PS: ILLR serves as the calculation engine for antennaweb, Radio Mobile, TVFool and others.

As part of the SHVERA proceedings, the FCC requested comments re how to improve ILLR.
ECHOSTAR submitted a report prepared by a well-known DTV engineering consultant:
http://www.tvtechnology.com/article/11418
http://hraunfoss.fcc.gov/edocs_public/attachmatch/FCC-05-199A3.pdf
And others contributed their opinion (these are only two of many, many, many comments):
http://gullfoss2.fcc.gov/prod/ecfs/retrieve.cgi?native_or_pdf=pdf&id_document=6517882072
http://www.mstv.org/docs/shvera6.17.05.pdf

The FCC only really cared about using the ILLR program to predict INTERFERENCE as part of
determining the best channel allocation and power for a new station, and no one volunteered
to provide a (FREE) update to the ILLR program. Citing a shortage of definitive information re most
of the aforementioned "shortcomings", the FCC decided there was insufficient data to take action:
http://www.fcc.gov/oet/info/documents/reports/SHVERA/SHVERA-FCC-05-199.pdf
Thereby avoiding the inevitable on-going feud between OTA broadcasters and SAT providers
over every fraction of a dB in whatever changes might be proposed to the ILLR program.....
A lot of trouble simply to decide whether a particular SAT viewer was entitled to watch the
NYC/L.A. network feed(s), rather than being restricted to his local network feed....give or
take a roof-top antenna that the viewer may or may not have the wherewithall/approvals....
and trees and other buildings in the way that are "missing" in even commercial models....

Anyone attempting to "predict" indoor signal levels is going to run into a very wide variation
in the amount of signal lost going through the walls (etc) of a building. The VU graphs only
cited a 6 dB Height Gain loss and an oversimplified 8 dB (detached) Building Penetration Loss.

The cited IEEE article shows a wide variation in that 8 dB figure....and apartments are higher.
It also did NOT address VHF penetration loss....that may or may not be more severe than UHF.
There is a shortage of recent data for VHF...esp since Europe, et. al. have gone UHF only....
Data from the 1950/60's isn't valid with Al. Foil backed insulation/wraps and low-E windows.

Here's some additional info re Height Gain and Building Loss:
http://www.digitalhome.ca/forum/showthread.php?p=918798
http://downloads.bbc.co.uk/rd/pubs/reports/1991-11.pdf
http://happy.emu.id.au/lab/info/digtv/files/bbc158.doc
 

·
Registered
Joined
·
4,966 Posts
Ignoring all the technical studies re: propogation models, etc.
To me, common sense says if ya can't use an indoor antenna in the typical Urban/Suburban US metro, without constantly adjusting it for a reliable signal every time ya switch channels, ur not gonna be able to do it in Canada either.
 

·
Registered
Joined
·
25 Posts
I agree with Hol Ands.

I don't believe that this publication was designed to be used as a model for local reception, but was another governments waste of money and time, spending time and money to look at reception issues and trying to determine how to fix those issues.

This forum already has addressed these issues and most people on the forum already understands Fris Transmission equations, Cascaded noise figures, antenna resonance and bandwidth, Fresnel zones, comparing some commercially available antenna's, and the basic's or radio wave propagation, path loss etc.

Home building materials plays a big role in indoor reception. Things such as a tile roof or foil backed building insulation can make a building appear opaque to UHF television. You can have the greatest signal in the world on one side of the house and little or no signal on the other. Or you can have a building 3 or more stories high in front of your building that blocks the signals and you might have no signal at all, yet you can see the light blinking on top of the transmitter from your location.

There is no guarantee's to reception.
 

·
Registered
Joined
·
8,422 Posts
^^^^
My impression was this was a presentation to show the issues to people who are not so knowlegable (politicians?).
 

·
Registered
Joined
·
1,744 Posts
I hope CFTO and Industry Canada sees the interference problem with VHF hi and increases CFTO-dt returning to 9 to the Full 30kw instead of 2.4kw.Why not max power,no one else on 9.
 

·
Registered
Joined
·
25 Posts
The one thing I will say is that I have visited dozens of OTA reception sites and I have never found one as through or as organized as this site.

You people really did a good job setting up this site and maintaining it.

As far as power increases goes, Asking for power increases does not help the cause either.
All radio frequencies reduce at the ratio of the square of the distance,
that is if the distance is doubled the level is reduced to 1/4
if its level. If the distance is tripled is it reduced to 1/9.
If the distance is 5 times it is reduced 1/25.
Eventually the level is so low that it just disappears into the noise.

It would take a proportionally large power increase to increase the range of reception more then a couple of miles.

Height of the transmitting antenna would have more gain then a simple power increase. Most long range VHF transmitter antenna's were placed at the top of their transmit tower 1000 feet or more above ground level.

Here is a list - http://www.economicexpert.com/a/TV:masts.htm

The problem with tall towers is that the signal passes over the people who lives closest to the tower and lands in places where you do not want it to go.

Any station within the Canadian Border Zone - http://www.fcc.gov/ib/sand/agree/files/can-bc/can-dtv.pdf
 

·
Registered
Joined
·
4,966 Posts
Link Budget?

Anyone here have an example DTV / ATSC link budget in spreadsheet form, to allow folks to see what the practical components they have control over
(or not) does to the link reliability?
I've seen plenty fris rundowns of Cascaded Noise Figure, etc. here & elsewhere.

But what about a complete Link budget from XMTR to RCVR?
Been lookin around (not here though) for a while for just an example, to see what assumptions are made, etc. and so far came up with only one, not sure how practical it is though. Especially since it was just a document without much explanation, etc...
 

·
Registered
Joined
·
3,396 Posts
A couple years back I developed some spread sheets that took data from either TVFool
or Radio Mobile and calculated the remaining Fade Margin. I then compared the results:
http://www.avsforum.com/avs-vb/showthread.php?p=10961518
I found a 19.5+ dB "fudge" factor between TVFool and R-M signal levels....which I
initially though was perhaps a spectrum analyzer ViewBandWidth correction factor.

The difference I found between TVFool and R-M turned out to be TVFool
assumed F(99,99) statistics (worst location, deepest fade), whereas R-M
was F(50,50) statistics (median location, median fade) which requires
a 15-20+ dB Fade Margin...so they were somewhat different approaches....
NOW, both of them yield a signal level only somewhat short of the max value.

Consequently, TVFool changed from F(99,99) statistics to something more meaningful....
which you see in TVFool today....at which point I and everyone else changed over to
simply using TVFool and then applying various correction factors, such as Clutter Loss,
Antenna Gain, Indoor Loss, System Noise Figure, ad nauseum....

Note that I explained and provided references to all of the pertinent parameters....and some
day I might/should update the spread sheet so data from the NEW and Improved TVFool
can be directly input....without the "fudge" factor. In the mean time, zero out "Lsa"....I think...

Groan....now I'll have to, yet again, take a shot at updating/validating that monster.....
 

·
Registered
Joined
·
4,966 Posts
Another N.F. Calc Spreadsheet

Hollands..
Thx,I'll have to sign up to avs to grab that old file of yours...
Been lookin for sumthin like that, to consider margins and what not...

Stampeder,

I posted yet another N.F. Calc spreadsheet the other day in Google Docs...
It's basically the same thing mentioned above,
Just a little simpler to follow (IMHO). One can also plug in their downlead length, splitter losses, drop lengths, etc and see what it does to ur system N.F (e.g. exchanging RG-6 atten from various mfrs, etc. one thing I can think of..,zero out the Balun for 300 ohm input pre-amps, etc. Of course just zero out some stage that isn't there. If ya happen to see anything in it that can be improved or needs modification, lemme know and I'll update it. One Mod I was gonna make was to assume the average NF from the subject Paper by CBC, where the average NF of the consumer TVs they tested was like 7.2 dB. So was gonna plug that in as default, should anyone else decide to use it.

https://docs.google.com/fileview?id...jktNmM5MS00YmM2LWExMWItYjJiNGQ0ZDJlNmQ0&hl=en

new Link:

http://docs.google.com/fileview?id=...WItMjE2OS00ZWQ4LTk0YjYtZWM2NDlkOTFlMTNj&hl=en

added a 2nd scenario, to illustrate the importance of keeping the LNA close to the antenna,
as opposed to inside the house.
 

·
Registered
Joined
·
2,579 Posts
What got me about the Marcoux paper was how bad the NF of the stand alone tuners that they sampled were. I'm wondering if it is simply that the TV's tested had later gen chipsets.

Slide #4 (with the wind turbines in the background) reminded me of this previous CBC study.

cheers.
 

·
Registered
Joined
·
6,257 Posts
What got me about the Marcoux paper was how bad the NF of the stand alone tuners that they sampled were. I'm wondering if it is simply that the TV's tested had later gen chipsets.
The RF tuner is separate from the decoder chipset so what chipset is used should be irrelevant to the NF.
 

·
Registered
Joined
·
4,966 Posts
Slide #4 (with the wind turbines in the background) reminded me of this previous CBC study.
Yeah no doubt...Anyone picking up the Buffalo stations, keep an eye
on the weather radar shots, look for the two colorful blobs of interference caused by the Wind Farms in Wyoming County (SE of Buffalo).
They have no way of cancelling out the echos from the blades with existing radar technology.

What got me about the Marcoux paper was how bad the NF of the stand alone tuners that they sampled were. I'm wondering if it is simply that the TV's tested had later gen chipsets.
Ya, if ya believe them, a Zenith DTT-901 is bad.
In My experience, it's the best ATSC tuner I've run in to yet.
Have heard the original Zenith DTT-900 had a bad tuner.
So does that mean the DTT-901 had some crappy
tuner also at some point? Or was something flawed in their test setup?

Not like I have a ton of tuners and TV's to compare, but I have no problem going out on a limb
to imply the Tuners in today's TVs are crap vs the Zenith DTT 901 I have.
Also have no problem saying the USB tuner I have is crap too, but it works.
Too bad we can't borrow their test equipment for our own tests. Folks here
more apt to have a variety of equipment to compare to.
 

·
Registered
Joined
·
4,966 Posts
had to look at them tuner measurements again, and noticed
what stands out realtive to the others, is the zenith dtt901's
large signal C/N threshold. So they are saying C/N can be as low
as 14 dB and still decode a signa (ATSC spec says at leasy 15 dB??).
So doesn't that mean it's one of the best they tested in the presence
of interference? So, if there's a lesson to be learned from their tests, can we infer how well a tuner deals with adj. ch interf. (Selectivity) may
be more important than it's Sensitivity (N.F.)?
Likely to be a lot more adj ch hits in any given market/area than co-chan.
Maybe that's why that dtt901 (it's low large signal C/N threshold) seems prettty good to most people in the real world?
 

·
Registered
Joined
·
3,396 Posts
PREAMP NOISE FIGURES:
Marcoux's paper cites Wetmore+Schnelle's IEEE Trans on Broadcasting Jun2004 article,
"The Performance of Antenna Amplifiers Used for Terrestrial DTV Reception"
(fol. link requires subscription..or a library):
http://ieeexplore.ieee.org/search/f...er&openedRefinements=*&searchField=Search+All
Although the article did not name manufacturers and part numbers, they did list the
manufacturer's spec value for Gain and Noise Figure in the same table as their Gain
and Noise Figure measurements.
I took an educated guess at matching up these results...which was fairly easy in many cases:
http://imageevent.com/holl_ands/files/ota
[The Preamp Gain table also includes measured NF values found in HDTVPrimer.]
Since worst Preamps appear to be antique B-T Suburban III and Voyager III models, I've always
wondered where Wetmore obtained the Preamps....perhaps after some were in use for many years???
After all, they were trying to assess what "typical" viewers were actually using....
But I think they might have been testing some "junk bin" Preamps removed by a local installer!!!!!

============================================================
CECB/STB/DTV SENSITIVITY/NOISE FIGURE:
NAB sponsored a CECB (converter box) test:
http://www.nabfastroad.org/NAB-STV Digital Converter Box Evaluation/Manuals-summary-report1.html
Sensitivity was compared to the MSTV/NAB Funded Prototype.
Note that actual performance (number of stations actually received) DID NOT CORRELATE
to either the measured sensitivity....or the number of ATSC A/74 Multipath Test Ensembles
(actual "worse case" field captures) successfully decoded.

LG (Zenith/Insignia) was the ONLY manufacturer that submitted "Noise Figure" test data (pg28)
in the (mistaken?) belief that they had to verify conformance to the FCC's max NF<14 dB spec:
https://fjallfoss.fcc.gov/prod/oet/forms/blobs/retrieve.cgi?attachment_id=820977&native_or_pdf=pdf
https://fjallfoss.fcc.gov/oetcf/eas...N&application_id=406918&fcc_id='BEJ9QKE00710'
And, of course, all CECB manufactures had to verify conformance to -83 dBm sensitivity spec,
which is NF=8 dB away from theoretical sensitivity (-106.2 dBm Noise + 15.2 dB CNR = -91 dBm).

FCC OET (Engineering Group) conducted tests, including Noise Figure (implied from Sensitivity),
for 2005/2006 vintage STB/DTVs (manufacturers unidentified):
http://www.fcc.gov/oet/info/documents/reports/TR-05-1017-ATSC-reception-testing.pdf

BTW: When I find an unexpected sensitivity (aka NF) result, I immediately suspect that there
might have been a low level RF leakage path which can either improve sensitivity...or degrade
it due to multipath....which is why it is important to have confirmation from alternative labs.
Attenuating the main signal to "prove" the leakage path doesn't exist is invalid...the leakage
path might be only one or two dB below threshold...which still affects sensitivity measurements.

=========================================================
ANTENNA SPECSMANSHIP:
BTW: There is probably a lot of "specsmanship" re Marcoux's antenna "tests". For example,
HBU-22 "specs" simply say 4.1 dBd Hi-VHF and 7.3 (not 7.2) dBd UHF Gain without
saying which channel nor whether it was guaranteed minimum (fat chance), some
sort of "average", "typical" (whatever that is supposed to mean) or the max Gain
found somewhere in the band (very likely).....YMMV....

And since Marcoux didn't explain what HE was measuring, perhaps it was the
WORST Gain found somewhere in the band.....maybe, maybe not....YMMV....
 
1 - 20 of 21 Posts
Top