Multipath and OTA

[Rich A]

Here's some information about my experiences with multipath and OTA digital TV reception.

What is multipath? It happens when a transmitted signal comes into your antenna from more than one direction.

How can that happen? Let's look at my specific problem which is a good example of a really bad multipath situation.

I'm on a little bit of a hill in my town. One of the highest points for a mile or so around. About 400 feet above sea level.

I have multiple Digital TV stations (about 10) available to me because of my clear shot to each. To the South there are a few, the strongest of which is WTNH which is ABC. To the north there are a half dozen or so, of which the strongest is WFSB which is CBS.

I have one Digital HD Tuner in my HTPC. Because I want to use the PVR functionality of Beyond TV to schedule recordings of the OTA digital stations I can not use an antenna rotor to move the antenna from north to south. So I opted to set up two antennas on a tower (I'm a ham and have a small 40 ft tower available to me), one facing south and the other facing north. Then I combined them and fed the end result to the tuner. This makes for a high gain sort of dual-directional antenna. I can now receive both the stations from opposite directions without having to turn any antenna. BUT there was a problem.

For some reason the ABC station from the south (with a 100 percent constant signal) experienced lots of skipping and stuttering. The reason was "multipath".

To my North-North-East is a rather tall cliff. It's just under 1,000 feet. The south side of this small mountain is mostly a flat cliff. And it's only about a mile to my north. So what happens is this:

ABC's single hits my southern facing antenna (where I WANT it) and gets fed down to the tuner. BUT that same signal hits the mountain to my north and bounced back from that direction, hitting my northern facing antenna. That signal is also sent down to the tuner, but because it's traveling a longer path hits the tuner a few micro-seconds later ... out of phase with the original signal. This "dual-out of phase" signal (of the same station) can drive a digital tuner bonkers. (and it does in my case) If this were an analog signal you might just see a slight ghosting or fringe around edges. But with digital signals it's much worse. The digital signal has to be decoded and when it's sees multiple bits where there should be only one .. well the picture gets broken up or not received at all.

Now this can happen with a single directional antenna as well. If the station is really strong, and the reflecting item, (mountain, building etc.) is in just the right position, a bounced second signal could be picked up by the BACK of your antenna. All antennas have what they call a Front to Back ratio. That means how much more gain there is in the pointed direction as apposed to the rear. They can't be perfect and reject ALL rear signals. But they can block the weaker ones. If you are relatively close to the transmitter and it's power is high enough and there is a large object in just the right position behind you, then some of that reflected signal can be picked up by the back of your antenna and you'll have the problem I have. Antennas that are "omni-directional" are especially prone to this phenomenon. That's because they are designed to pick up signals from all directions. A directional antenna is less prone to multipath and the higher the F/B ratio is of the directional antenna the more it will block signals coming in from the back. Usually the larger and higher gain directional antennas will have high F/B ratios.

So how do I intend to solve my problem? First, there are stations in both directions that are "must haves" for me. So I must be able to have as good an antenna I need to receive each of them. My solution (which I haven't set up yet and is a Spring project) is to NOT combine the antennas up on the tower and feed both to one tuner. Rather I will run a dedicated coax feed from each antenna down to it's OWN tuner in the HTPC. This of course requires a 2nd OTA digital tuner. And it will also require that I ONLY use the south antenna and it's tuner for those south stations when recording or viewing. And the North antenna and it's tuner will only be used for all northerly stations. So in effect I have a North tuner, and a South tuner. The North tuner is never asked to view, or record ABC ... And the south tuner only records ABC (and it's southern neighbors) I believe this will eliminate the multipath reception problem I have with that one ABC station. None of the other Southerly stations are strong enough or positioned so as to create their own multipath. I imagine this will require some special set up in my Beyond TV settings.

For example, I can't use "record from any source" when recording ABC. The multipath will STILL BE there on the North antenna. If I choose "any source" then BTV could possibly start a recording on the northern tuner/antenna. So I must be careful to set up all automated recordings for ABC to ONLY use the southern Antenna/Tuner.

An aside note: One might think that I could just use ONE antenna facing north and therefor only get the reflected signal from the mountain and no multipath. The problem with a received "bounced" signal is that unlike a perfectly curved dish that receives, and bounces the signal to the dish pickup (like Satellite dishes) the item in question is not perfect. The bounced signal off a mountain or building can have it's polarization mixed up, or it can come and go or vary greatly in strength. But the more important factor is that there are OTHER Southern stations I still need to receive as well from the South that "don't" bounce off the mountain. So I think my "multiple" tuner/antenna setup is going to be the only way to solve this unique problem.

I hope that gives you some insight to the wonderful world of "multipath".

[Fonceur]

Another option would be to put a filter to cut off that bounced signal from the southern antenna. Which is basically what cable companies do to remove the analog frequencies that their customers don't subscribe too. Now weither you could actually buy such a filter to only remove the specific frequency used by ABC in your area, I have no idea.

[Rich A]

Originally Posted by Fonceur

Another option would be to put a filter to cut off that bounced signal from the southern antenna. Which is basically what cable companies do to remove the analog frequencies that their customers don't subscribe too. Now whether you could actually buy such a filter to only remove the specific frequency used by ABC in your area, I have no idea.

Oh yes . I investigated that in the beginning. For the filter to work it would have to be a deep "notch" filter. Something with steep skirts so as to not attenuate any close-by other signals. They do exist .. and are mostly used in the example you gave. But the PRICE .. egad. They are quite expensive.

I've seen some setups that are not so expensive where instead of using a single notch filter they used a low pass and high pass filter in parallel. Where the over-lapping end of each was the center frequency you wanted to notch out. A little less expensive but because they are cheaper, the skirts are not as sharp and you end up attenuating channels that are close by.

But that IS a way to do it if one wanted to spend the bucks. Which reminds me of something I had forgotten about. That would be a "stub" attenuator. Where you would attach a stub tuned to exactly the frequency of the signal you wish to get rid of. Hmmmm. That might be a way to go. Here in CT, in the old days when the TV front ends weren't so selective, we had a problem with TV channel 8. A lot of people like to get the next channel (9) out of Ney York. But if they lived anywhere near CT's Channel 8 the signal was so strong it would bleed over to channel 9 and mess up the picture. Way back then I used to work in a TV repair shop and a trick we used to do was to put a 300 ohm tuned "stub" on the TV's antenna terminals. This was simply a shorted length of 300 ohm twin-lead that was cut to the frequency of channel 8. I forget what the length was but 9 inches sounds familiar. I'd start at that length and while watching channel 9 keep cutting a 1/4 inch at a time and shorting it until the interference from the strong channel 8 went away. It didn't have any effect on the actual channel 8 as the signal was so strong to begin with. Wow .. thanks for jogging my memory. That may work out pretty good.

[pdaughe]

Wow, another great post by Rich A!

I suspect there many be many readers who read this post and say to themselves, "doesn't apply to me" or "way over my head."

Actually, the base subject matter of this thread, I believe, DOES apply to most.

I don't understand this subject matter the way Rich does, but multipath reception is indeed a concern and can impact almost anyone.

One of the important conclusions to be drawn (allow me to translate for those of us way below your level Rich) from Rich's post is that with respect to OTA reception,

MORE IS NOT BETTER.

That is, a bigger antenna is not necessarily better than a smaller one; a multi-directional antenna is not necessarily better than a directional antenna.

When I decided to "experiment" with OTA, antennaweb.org recommended a small, directional antenna. I purchased a $50 Winegard antenna adhering to the recommendation.

THAT was difficult for me -- I AM susceptible to the "more (bigger) is better"; "pay more and it will function better" kind of thinking. I was fully aware there were much larger antennas available.

I didn't cave though and simply bought the smaller, cheaper RECOMMENDED antenna -- afterall, it was just an experiment.....

I'm glad I did. I receive thirteen digital channels with 100% signal strength on all but one, which is 98% signal strength. The experiment was a huge success; all of the channels play beautifully in BTV (if you are wondering, HDTV on BTV simply ROCKS!)

In my case I live in a Detroit suburb and am VERY close to the transmitting stations (I think the farthest one away is around 12 miles).

The bottom line: Follow the guidelines provided by antennaweb.org!

Suggestion: if you haven't exploited HDTV in BTV 4, DO!

[Rich A]

Just an added bit of information. There are "filters" designed for home use that are tunable and run from 30 to 60 dollars. However being "low cost" they are designed to be rather wide. (the tighter the "notch" the more it will cost) In my case the best I could find would notch the offending channel by 20 db or more, but also affect the next 12 channels from 7 to 12 db above and below the notch frequency. If you don't have any channels close by the frequency you need to filter or they are strong, these consumer notch filters may work. In my case the offending channel is a VHF channel and those filters don't work as well as the UHF ones.

Also, something I found during my search for a solution was that some OTA digital tuners claim to have advanced circuitry to cut down multipath problems.

If this is true, and you think you may have this problem, you might want to ask the manufacturer of any tuner you are looking at, about how it handles multipath problems. Just how it works I haven't found out. Very possibly some kind of DSP or digibtal signal processing that is becoming so comon in the radio receiver world. I believe the tuner I have (ATI HDTV Wonder) is pretty basic and has nothing like that.

[merrypig]

Any modern rf receiver circuit has an adaptive filter on the input which performs a correlation of the incoming signal with a time delayed version. This helps spot 'multipath' signals that are coming in after the original (or even ahead of it...)

Unfortunately, the correlation is quite expensive in terms of hardware resources so there is a finite delay that they can accomodate. Typically this will be tuned to handle multipath signals caused by a reflection coming off neighbouring buildings etc and down the same antenna. We're probably talking very small delays in the path - microseconds probably. (not sure to be honest). A signal that is travelling off to a nearby cliff face as in your case however, will incur quite a delay - normally a signal travelling much further would likely be too weak to be dominant - I think the use of a second antenna tuned to that direction is making your problem worse certainly, particularly with strong digital OTA transmitters in use.

I think your'e probably on the right tracks to consider the notch filtering on your 'non preferred' antenna - you can pretty much be assured that the hardware is already working to reject the worst case multi path signals that it's capable of.

What I do find amazing though, is that you need a signal strength reading of '90%' or more to get a good signal with OTA. I'm suspicious of the metric used here. Digital receivers will work with a much greater attenuation than that - I suspect this % is more a signal of bit error rate than signal strength. So a 90% signal is perhaps a much lower % level of reception but is falling outside the CRC/BER correction mechanism in the transport itself. Kind of irrelevant I guess. Just been niggling me

[Baylo]

Originally Posted by merrypig

What I do find amazing though, is that you need a signal strength reading of '90%' or more to get a good signal with OTA. I'm suspicious of the metric used here. Digital receivers will work with a much greater attenuation than that - I suspect this % is more a signal of bit error rate than signal strength. So a 90% signal is perhaps a much lower % level of reception but is falling outside the CRC/BER correction mechanism in the transport itself. Kind of irrelevant I guess. Just been niggling me

I think you might be right. The Fusion HDTV software gives both a signal strength (%) and level (db) reading. My guess is that the dbs reflects the level of attenuation you refered to.

Mark

[Rich A]

Originally Posted by merrypig

Any modern rf receiver circuit has an adaptive filter on the input which performs a correlation of the incoming signal with a time delayed version. This helps spot 'multipath' signals that are coming in after the original (or even ahead of it...)

Unfortunately, the correlation is quite expensive in terms of hardware resources so there is a finite delay that they can accomodate. Typically this will be tuned to handle multipath signals caused by a reflection coming off neighbouring buildings etc and down the same antenna. We're probably talking very small delays in the path - microseconds probably. (not sure to be honest). A signal that is travelling off to a nearby cliff face as in your case however, will incur quite a delay - normally a signal travelling much further would likely be too weak to be dominant - I think the use of a second antenna tuned to that direction is making your problem worse certainly, particularly with strong digital OTA transmitters in use.

I think your'e probably on the right tracks to consider the notch filtering on your 'non preferred' antenna - you can pretty much be assured that the hardware is already working to reject the worst case multi path signals that it's capable of.

What I do find amazing though, is that you need a signal strength reading of '90%' or more to get a good signal with OTA. I'm suspicious of the metric used here. Digital receivers will work with a much greater attenuation than that - I suspect this % is more a signal of bit error rate than signal strength. So a 90% signal is perhaps a much lower % level of reception but is falling outside the CRC/BER correction mechanism in the transport itself. Kind of irrelevant I guess. Just been niggling me

Great info .. thanks. Remember back when we were first testing the BTV and digital tuners? At that time there was no conversion factor for the signal strength and they were just reading that alone. We'd get figures like from 0 to 20,000 or more mostly dependant on the particular manufacturer's way of presenting it.

Of course now we have the final polished result which is a "scale" from 0 to 100. I'm pretty sure they are no longer just sampling signal strength but (to their credit) they are analyzing the signal by sampling it and checking for bit drops along with the signal strength. This means a more real-world scale not only based on signal but also quality (due to drops) Multi path reception, is one case where you can have very high signals but given the right circumstances still have bit drops due to multipath reception.

I wonder just how much digital processing is done on these tuner cards to reduce that problem. After all, they are relatively cheap. But to verify what you have said about typical modern receivers having that feature, I have seen the subject in some advertisements or in technical white papers. I believe the Divco Fusion card manufacturer touts it.

You'd love to chat with my ham buddy who works for Motorola and handles all the communications hardware for Connecticut and surrounding states. He was just telling me the other day about the local State Police and their setup. They were testing some antennas that were part of a digital repeater thing .. Normally they ran at 100 watts output which was the design and were configured to be able to fill in some dead areas around valleys and mountains. They have equipment I can only drool over .. <grin> and first tested at 100 watts and checked bit error percentage. (this is a fully digital voice system) Then they went down to 10 watts and fully expected a greater percentage of bit errors. Surprisingly they found NO difference. Which I guess means their antenna sites were well positioned. He says they use a repeater system and at the transmitter/receivers there are computers that do automatic bit error correction. He was trying to explain all this to me .. but frankly it was a little over my head. Basically he said that if an officer said a word like "traffic" and the word was glitched because a bit or two being dropped made it sound like "traff" that their digital system could correct it so it still came out right. Pretty amazing stuff. I have other ham friends in a lot of communications areas like Civil Air Patrol and just about all of them are going to 100 percent digital voice and getting rid of their old analog voice systems.

I think I can shed some light on why it may take such a high signal to give reliable and steady digital video reception. I have played a lot (in my youth) with ham radio dealing with frequencies up to about 1296 Mhz. When you get into the very high frequencies and are dealing with digital signals, (video in particular) the RF is indeed "flaky". There's a lot of signal rising and falling and so many other things that can affect it. Even ground fog can cause "ducting" and in some cases actually increase your range. Video in particular requires a lot of strength as even with a digital signal it's using a considerable band width. I built a small TV station back in the 70s and was using the 70 cm band (430 mhz) and even with a 100 watt signal had a very limited range (about 5 to 10 miles line of sight) This was of course all analog and I was not running any kind of side band filter so the signal was 3 mhz wide. That meant to stay legal with the FCC I had to stay in the UHF ham bands or higher where there was enough bandwidth available.

But the bottom line was my range for 100 percent clear voice reception on the same band was 10 times that of the video. It was fun though. All this digital stuff is sorta new to me. Yeah, gimme the days when "real radios" glowed in the dark.

[dlandrum]

From a cost and quality perspective, it seems buying a 2nd OTA tuner is the best option. Setup a 2nd digital lineup on snapsteam.net. The result is that you have Digital North Lineup (DN) and Digital South Lineup (DS). I know you can figure-out the setup.

On the DN lineup, remove ABC. On the DS lineup, remove CBS. The when you search for a show on CBS, only one channel appears. If you did not remove CBS from DS, then you would see duplicate shows in your search results and not know which to pick. For this reason, I think you would remove channels from both lineups to have one distinct set of channels across the two lineups.

When you do choose to record a show, you leave all of the defaults as they are. BTV will map the channel to the lineup and the lineup to a tuner. For your case these are all one-to-one. Yet, you could easily add a third OTA tuner to use your DN feed and DN lineup, etc.

[Rich A]

Originally Posted by dlandrum

From a cost and quality perspective, it seems buying a 2nd OTA tuner is the best option. Setup a 2nd digital lineup on snapsteam.net. The result is that you have Digital North Lineup (DN) and Digital South Lineup (DS). I know you can figure-out the setup.

On the DN lineup, remove ABC. On the DS lineup, remove CBS. The when you search for a show on CBS, only one channel appears. If you did not remove CBS from DS, then you would see duplicate shows in your search results and not know which to pick. For this reason, I think you would remove channels from both lineups to have one distinct set of channels across the two lineups.

When you do choose to record a show, you leave all of the defaults as they are. BTV will map the channel to the lineup and the lineup to a tuner. For your case these are all one-to-one. Yet, you could easily add a third OTA tuner to use your DN feed and DN lineup, etc.

David,
Thanks. That indeed looks to be the best all around solution. PLUS, it will give me the ability to record two shows at once (as long as the stations are not from the same direction. I was really getting caught up with the single tuner/dual antenna and notch filters etc. I think your post may have cleared my thinking. Thanks ..

[merrypig]

Someone's been thinking outside of the box! Splendid idea

[davefred99]

Why dont you order a channel Master Jointenna tuned to filter out just the problem channel. They are basicaly a combiner with a single channel notch filter on one input for joining two antennas and must be ordered for a specific channel but can be stacked for multiple channels.They cost about $30.00 each and can be ordered from warren electronics and take a about two weeks to get as they are factory tuned.
I have a similar problem where as I have one channel that is only 7 miles from me and everthing else is over 50 miles out. The close channel is over powering the distant signals and needs to be attenuated so I am going to use two antennas one small antenna for the close channel and a bigger CH 4228 for the distant ones with the filter on the input of the bigger antenna.
There was a pretty good discussion about them on the AVS forums in there OTA antenna section.

[Rich A]

Originally Posted by davefred99

Why dont you order a channel Master Jointenna tuned to filter out just the problem channel. They are basicaly a combiner with a single channel notch filter on one input for joining two antennas and must be ordered for a specific channel but can be stacked for multiple channels.They cost about $30.00 each and can be ordered from warren electronics and take a about two weeks to get as they are factory tuned.
I have a similar problem where as I have one channel that is only 7 miles from me and everthing else is over 50 miles out. The close channel is over powering the distant signals and needs to be attenuated so I am going to use two antennas one small antenna for the close channel and a bigger CH 4228 for the distant ones with the filter on the input of the bigger antenna.
There was a pretty good discussion about them on the AVS forums in there OTA antenna section.

Yes, I'm familiar with the Jointenna thing. That too was something I was considering. But adding a 2nd hdtv turner really wouldn't be that more expensive AND as I mentioned it would give me the ability to record ABC and either CBS or NBC at the same time. But not CBS and NBC. But I could do ABC and PBS or .... who's on first? Arrghhhh.

[dlandrum]

Leave NBC on both lineups, but shange the callsign to nNBC and sNBC. Then you can smart schedule recordings for ABC and nNBC OR CBS and sNBC ... etc.

[Fonceur]

Originally Posted by Rich A

Yes, I'm familiar with the Jointenna thing. That too was something I was considering. But adding a 2nd hdtv turner really wouldn't be that more expensive AND as I mentioned it would give me the ability to record ABC and either CBS or NBC at the same time. But not CBS and NBC. But I could do ABC and PBS or .... who's on first? Arrghhhh.

Or do both, if it's really just a 30$ component + another HDTV tuner, best of both worlds...