microstutter in 6.2.7

[mikesm]

Originally Posted by ToxMox

While things were set to false for me I was noticing a lot of little pixelations and other defects which led me to believe that something is/was wrong with my signal. I moved some splitters around and tweaked my amp and attenuator and switched my HDHRs to QAM from OTA and now it seems I am getting a better feed/signal even though the numbers/stats of the percentages of the signal quality/power and symbol quality aren't much different than they were before. Anyways I went back to setting the property to encode as program stream to true and last night I only had a few slowmo stutters on only WNBCDT and they were much briefer than before. When I was watching 1080i shows on nbc I was getting out of order frames with all my video decoders except for the SageTV Mpeg decoder if you can believe that. Right now my setup seems pretty solid with the exception of NBC not being perfect so I'll continue tweaking that but I'm guessing that most of these slowmo stutter problems are because the stream that Sage is encoding into mpegs has dropouts or corruption due to either signal issues in general or over/underpowering the signal to the tuners. I'll try and give another update at the end of the week after I've watched all my season premieres.

You should be careful about putting too much amplification in as well. Cable amps generally are pretty linear, but they have a point where if the total energy in the input passband gets high enough, the amp will clip on output, causing severe momentary distortion. This shows up in STB diagnostics as FEC failures on the QAM signal.

BTW, if you were seeing nothing but QAM and no analog channels this wouldn't be an issue, as QAM looks like a signal with constant power.

In general, home HFC network design should strive to have the same signal levels at each jack that the MSO is feeding the home with. Increasing power by 10 db or more compared to the normal feed isn't generally helpful and can cause problems. If the level the MSO is feeding you is low, they need to fix that problem, and not you driving to overdrive stuff to compensate.

Thanks,
Mike

[ToxMox]

Thanks for the info.
I am using a 38db gain amp which has a -20db output test port I'm using for output into an attenuator. So in effect i'm getting +18db to the attenuator where I lower the signal a little bit then to my 16 way splitter which is -16db on each port right now so I'm pretty much even near as I can tell. I'll check the STB stats later. The STB was performing much better than usual after I did after all this as well.

[mikesm]

Originally Posted by ToxMox

Thanks for the info.
I am using a 38db gain amp which has a -20db output test port I'm using for output into an attenuator. So in effect i'm getting +18db to the attenuator where I lower the signal a little bit then to my 16 way splitter which is -16db on each port right now so I'm pretty much even near as I can tell. I'll check the STB stats later. The STB was performing much better than usual after I did after all this as well.

Interesting. Most of these amps clip at around 50 dbmv, so if it's a +38db gain, you should make sure you aren't feeding it with more than 10 dbmv from the source. It sounds like you're not triggering this. The test point is usually just attenuated, so if the amp is clipping, it'll still see the distortion.

A decent 16 way splitter should only give you about 13 or 14 db loss (12 db for the splitting and maybe 1-2db for insertion loss), but you're probably fine. The STB stats will tell you if you are seeing problems.

What is the signal level that you see at the drop (from the MSO)?

I wish the HD HomeRun had better diagnostic data output. They should be able to tell you dbmv signal levels, FEC counts, etc... Their demod chip produces all that data.

Thanks,
Mike

[ToxMox]

Originally Posted by mikesm

Interesting. Most of these amps clip at around 50 dbmv, so if it's a +38db gain, you should make sure you aren't feeding it with more than 10 dbmv from the source. It sounds like you're not triggering this. The test point is usually just attenuated, so if the amp is clipping, it'll still see the distortion.

A decent 16 way splitter should only give you about 13 or 14 db loss (12 db for the splitting and maybe 1-2db for insertion loss), but you're probably fine. The STB stats will tell you if you are seeing problems.

What is the signal level that you see at the drop (from the MSO)?

I wish the HD HomeRun had better diagnostic data output. They should be able to tell you dbmv signal levels, FEC counts, etc... Their demod chip produces all that data.

Thanks,
Mike

I'm not home right now so I can't really give exact figures from the MSO but I have a two way splitter. One goes to my amp and the other to my cable modem so in theory the levels my cable modem is reporting should be the same as I am feeding the amp right (which I believe to be -2 dB)?
That being said here are the numbers from my cable modem:

Downstream Value
Frequency 609000000 Hz
Signal to Noise Ratio 37 dB
QAM 256
Network Access Control Object ON
Power Level -2 dBmV
The Downstream Power Level reading is a snapshot taken at the time this page was requested. Please Reload/Refresh this Page for a new reading

Upstream Value
Channel ID 6
Frequency 24000000 Hz
Ranging Service ID 331
Symbol Rate 2.560 Msym/s
Power Level 46 dBmV

Signal Stats Value
Total Unerrored Codewords -498890986
Total Correctable Codewords 2420
Total Uncorrectable Codewords 78

[ToxMox]

BTW here is a PDF of the amp I am using.
http://channelvision.com/pdf/product.../CVT-38BID.pdf

[mikesm]

Originally Posted by ToxMox

I'm not home right now so I can't really give exact figures from the MSO but I have a two way splitter. One goes to my amp and the other to my cable modem so in theory the levels my cable modem is reporting should be the same as I am feeding the amp right (which I believe to be -2 dB)?
That being said here are the numbers from my cable modem:

Downstream Value
Frequency 609000000 Hz
Signal to Noise Ratio 37 dB
QAM 256
Network Access Control Object ON
Power Level -2 dBmV
The Downstream Power Level reading is a snapshot taken at the time this page was requested. Please Reload/Refresh this Page for a new reading

Upstream Value
Channel ID 6
Frequency 24000000 Hz
Ranging Service ID 331
Symbol Rate 2.560 Msym/s
Power Level 46 dBmV

Signal Stats Value
Total Unerrored Codewords -498890986
Total Correctable Codewords 2420
Total Uncorrectable Codewords 78

Ok, that's a very nice amp. It has tilt adjustment too which is nice. Personally, I would use the main output of the amp, and just turn the gain down. That makes the tilt of the amp less of an issue, lower noise, and it will run cooler too, etc...

The modem is showing -2 dbmv, so that means the drop is probably at +2dbmv (3 db for the split, and 1 db for insertion loss), which is a little low, but fine, esp. since you are seeing 37db in SNR (more than adequate to support 256 QAM). That also implies the SNR at the drop is 41 dbmv which is good.

Obviously, if you are feeding 16 jacks or so from the signal you are going to need a distribution amp. However, if you fed the STB direct from the splitter at the drop, you should not be seeing any improvement by feeding it through the distribution network with the amp, etc...

For example, if you lowered the signal by 16 db for the splitter but the noise was stable, you'd end up with a SNR of 21 dbmv, which is dicey for 64 QAM, but no good for 256 QAM (you need > 30dbmv SNR for 256 QAM at 10E-8 BER).

I generally like to see an SNR in the 35 dbmv range at the STB. You can't do much about noise, hence the need for amplification, but the amp tends to add noise too (7 db for your model at max gain max tilt). The noise figure goes up with the gain and tilt adjustment, so backing off the gain on the amp and using the main output should trim 3-4 db of noise, and improve your SNR as compared to your set up today.

Make sense? This RF stuff is rarely explained by the MSO or vendors, but it's not all that complicated when you think about it.

Thanks,
Mike

[mikesm]

PS Your stats on the modem look pretty good. The un-errored FEC count is negative because the counter wrapped. A few FEC uncorrectables are to be expected if the modem has been up a long time, which is my bet since the counter wrapped.

If you were seeing a lot of correctable errors soon after booting the modem, that would be a sign of trouble, as in general, the plant should not need the modem to do FEC to assure reliable operation.

Power levels in the upstream below 50 is acceptable. If you were over 50 dbmv signal level in the reverse, that would be a bad sign, as that would mean the modem is having to "yell" at the top of it's lungs for the CMTS to hear it.

Thanks,
Mike

[ToxMox]

Thanks for all the great info. I rebooted the cable modem a couple/few days ago so those numbers are a for a few days of uptime. I originally went the way you suggested of using the normal output instead of the -20dB test output and backing off the gain but I had worse results. After some reading online about my amp it turns out the amp is always boosting 38dB and the gain adjustment backs down the incoming signal before boosting.

I read that in the 2nd post of this thread on another forum:
http://www.epanorama.net/phpBB2/view...03119cc91046b1

[ToxMox]

I'm sorry. I just remembered that it isn't a two way split at the beginning. Ok now I know this is a little strange but I did this:
First split is a 3 way splitter that has two -7dB inserts and one -3.5dB jack. I then plugged one of the -7dB splits into my cable modem and then I connected the -3.5 and the other -7 to a 2way in reverse in essence combining them and then I went into the amp.

[mikesm]

Originally Posted by ToxMox

Thanks for all the great info. I rebooted the cable modem a couple/few days ago so those numbers are a for a few days of uptime. I originally went the way you suggested of using the normal output instead of the -20dB test output and backing off the gain but I had worse results. After some reading online about my amp it turns out the amp is always boosting 38dB and the gain adjustment backs down the incoming signal before boosting.

I read that in the 2nd post of this thread on another forum:
http://www.epanorama.net/phpBB2/view...03119cc91046b1

Boy if that's true it's a pretty stupid design. I would have expected better from channelvision!!

Oh, I really would not do what your are doing about splitting and combining like that. I would take your two-way splitter and remove the three-way one. That three-way is also 2 2-way splitters in one package. Odd.

Thanks,
mike

[ToxMox]

Originally Posted by mikesm

Boy if that's true it's a pretty stupid design. I would have expected better from channelvision!!

Oh, I really would not do what your are doing about splitting and combining like that. I would take your two-way splitter and remove the three-way one. That three-way is also 2 2-way splitters in one package. Odd.

Thanks,
mike

Maybe that is why it was discontinued.
You were right about getting rid of the 3 way splitter and just put in a 2 way. On my STB I am getting around +7dB and 35 S/N and on my cable modem here are the new numbers:
Downstream Value
Frequency 609000000 Hz
Signal to Noise Ratio 37 dB
QAM 256
Network Access Control Object ON
Power Level 1 dBmV
The Downstream Power Level reading is a snapshot taken at the time this page was requested. Please Reload/Refresh this Page for a new reading

Upstream Value
Channel ID 6
Frequency 24000000 Hz
Ranging Service ID 331
Symbol Rate 2.560 Msym/s
Power Level 43 dBmV

Signal Stats Value
Total Unerrored Codewords 45365306
Total Correctable Codewords 2
Total Uncorrectable Codewords 0

I'll report back after tonight if my NBCDT problems are fixed. Thanks again Mike

[rsagetv99]

My stuttering is definitely not the HDHR, because I was watching a recorded show and it would start to stutter, most noticeably audio, but the video did slow down too.

What I found was that I could skip back in the recording, or pause and start playing again and the stuttering would subside for a couple minutes then come right back.

I tried different video decoders including Nvidia (what I normally use with no problems), PowerDVD, and the SageTV decoder - all had this problem. I tried rebooting, and stopping all background processes such as show analyzer but nothing worked. Finally I switched my audio setting from the Nvidia decoders to the SageTV Audio decoder and it went away.

I will submit a bug report.

[amac]

My experience has been the longer Sage has been running, accumulating history, and stocking up shows (I've got over 9 days of programming stored up), the more likely that Sage as a stutter at a fixed interval (every x minutes).

Sage support fixed this for me over a year ago with an optimization in their sort routines used in their scheduling algorithm (which runs every few minutes at the same fixed interval) I suspect that I've accumulated so much now that another optimization is needed.

[mikesm]

Originally Posted by ToxMox

Maybe that is why it was discontinued.
You were right about getting rid of the 3 way splitter and just put in a 2 way. On my STB I am getting around +7dB and 35 S/N and on my cable modem here are the new numbers:
Downstream Value
Frequency 609000000 Hz
Signal to Noise Ratio 37 dB
QAM 256
Network Access Control Object ON
Power Level 1 dBmV
The Downstream Power Level reading is a snapshot taken at the time this page was requested. Please Reload/Refresh this Page for a new reading

Upstream Value
Channel ID 6
Frequency 24000000 Hz
Ranging Service ID 331
Symbol Rate 2.560 Msym/s
Power Level 43 dBmV

Signal Stats Value
Total Unerrored Codewords 45365306
Total Correctable Codewords 2
Total Uncorrectable Codewords 0

I'll report back after tonight if my NBCDT problems are fixed. Thanks again Mike

Hey, did this fix your stuttering problem?

thx
mike

[ToxMox]

I wish I could say yes. Most of my problems are gone. I am still having problems with some of my WNBCDT recordings. Some are fine others I have huge stuttering issues. Like last night's Biggest Loser is not even an HD program but the digital recording of it had several of my slowmotion sadness as I've begun calling it. I am still playing around. I think having a perfect signal fixes these stuttering issues but Sage should be able to play through the problem areas without the playback going crazy. Other apps can play these recordings just fine so right now I'm at a loss. I'm also stupid because I was planning on saving my copy of The Biggest Loser so I could have it for testing but I instinctively deleted it right after watching it.

[kevine]

Originally Posted by ToxMox

I'm also stupid because I was planning on saving my copy of The Biggest Loser so I could have it for testing but I instinctively deleted it right after watching it.

This is 1 feature I wish they had - Undelete. This is especially useful when for the people that keep their drives clean, like I believe you do. It would be nice if things were just marked for deletion instead of being deleted immediately. Then they would be the first to go when room is needed.

Even better would be to take the classes and prioritze those as well like favorites. For example, I someone wished manual recordings got deleted before favorites, they could set that.

[ToxMox]

Ok...who here with this problem is running Vista?