Sirius blunder
Last edited by Blaxer; Nov 16, 2005 at 05:31 PM.
PS It's worth it!
JMC
Trending Topics
In other words, the car was ordered with Sirius, charged against the car for Sirius, but when it arrived there was no Sirius so we got a $500 credit against the original cost of the Sirius.
We don't miss it
The Best of Mercedes & AMG
It cant be added on by the dealer on the SLK as of now, like the CD changer
Do what i did and get a i-pod link it should run $500 with parts and labor.
It's ok and all, but it cuts out far far too often for my taste. Every overpass, in the garage, etc.
Sorry, but it doesn't take a friggin' genius to install a few bytes of ram, and then cache the signal so it wouldn't cut out unless you lost the sat for like 15 seconds.
Instead, I'm going with the iPod kit. A $300 player that holds 60gb worth of music? That works for me.
Personally, I woudn't consider that worth the trade for no drop-outs.
Personally, I woudn't consider that worth the trade for no drop-outs.
Of course they could make it so that you could dial the delay/cache to anything from 0 to 60 seconds, or whatever you'd like. It would also be fairly trivial to have it start caching as soon as the alarm button is pressed, so that by the time you're belted in it's already cached 60 seconds. It really would be quite easy if they just sat down and thought it through from the start.
This is 2005, not 1985. The thing is just very poorly engineered when they know from the beginning that they'll get dropouts. Does Microsoft own stock in Sirius or something?
Also, on logic 7.....sirius tends to come out of the center for music and hisses with speech
Of course they could make it so that you could dial the delay/cache to anything from 0 to 60 seconds, or whatever you'd like. It would also be fairly trivial to have it start caching as soon as the alarm button is pressed, so that by the time you're belted in it's already cached 60 seconds. It really would be quite easy if they just sat down and thought it through from the start.
This is 2005, not 1985. The thing is just very poorly engineered when they know from the beginning that they'll get dropouts. Does Microsoft own stock in Sirius or something?
The rad sat datstream is realtime. When you lose signal (tunnels, buildings obstructing, etc.) the data is gone. If your sat rad cached, then you just wouldn't know it until 15 seconds later.
For your request to work, Sirius would have to have a complicated scheme where they time compress the signal and send in chunks, repeating each chunk multiple times.
Not gonna happen!
The better solution which XM seems to be ahead on, is terrestrial repeaters, particularly in denser cities. When I pull into my garage I can watch my radio switch from sat to terrestrial.
They can't be using absolutely 100% of the available bandwidth, so use the rest for error correction.
Simply compress the signal as usual, and chop it up and send it in rotating 10 second bursts. Send 10 seconds of channel 1, then 10 of channel 2, etc up to channel 10, and multiply this by say 15 to get 150 channels. No more bandwidth is being used, but you decrease your dropouts by a factor of ten, yet you increase their length to 10 seconds.
Personally I'd rather have one dropout for 10 seconds in a drive, rather than 10 dropouts of 1 second, which is what I get now.
Then take the extra bandwidth, whether it's 10%, 20%, or 50%, and send error correction (redundant) data down, also in burst mode, as far away from the original signal (timewise) as possible. You ignore channels like the weather channels and those which are infrequently used, and stick to the main channels that people listen to the most.
Now almost all of the dropouts are gone on the main channels, and when you do get one it will be on a fairly unused channel, and still it will be far far less than we get now with the current stream because the error correction data will replace as much as possible.
Cost to them? Oh, a couple hours of programming on the upstream and downstream. No cost for repeaters.
Of course, repeaters are a 100x better idea and Sirius should have them too. But they can do quite a bit to decrease their dropouts by a huge huge margin without costing a penny.
Yes, it would incur a very slight startup time of a couple seconds before your first packet arrives (whether it be the main packet or the error correction packet it should be there in under 5 seconds), just put up a message 'Acquiring Satellite Signal' and nobody will notice.
They can't be using absolutely 100% of the available bandwidth, so use the rest for error correction.
Simply compress the signal as usual, and chop it up and send it in rotating 10 second bursts. Send 10 seconds of channel 1, then 10 of channel 2, etc up to channel 10, and multiply this by say 15 to get 150 channels. No more bandwidth is being used, but you decrease your dropouts by a factor of ten, yet you increase their length to 10 seconds.
Personally I'd rather have one dropout for 10 seconds in a drive, rather than 10 dropouts of 1 second, which is what I get now.
Then take the extra bandwidth, whether it's 10%, 20%, or 50%, and send error correction (redundant) data down, also in burst mode, as far away from the original signal (timewise) as possible. You ignore channels like the weather channels and those which are infrequently used, and stick to the main channels that people listen to the most.
Now almost all of the dropouts are gone on the main channels, and when you do get one it will be on a fairly unused channel, and still it will be far far less than we get now with the current stream because the error correction data will replace as much as possible.
Cost to them? Oh, a couple hours of programming on the upstream and downstream. No cost for repeaters.
Of course, repeaters are a 100x better idea and Sirius should have them too. But they can do quite a bit to decrease their dropouts by a huge huge margin without costing a penny.
Yes, it would incur a very slight startup time of a couple seconds before your first packet arrives (whether it be the main packet or the error correction packet it should be there in under 5 seconds), just put up a message 'Acquiring Satellite Signal' and nobody will notice.

First principle is to allow absolute lowest cost for the receivers (this is the part they have to build and sell in the tens of millions). It is not feasible to have enough memory to cache all the channels (or even the more popular ones). When you tune to a channel you are decoding that channel only.
My XM does have a 30 minute "cache". You can replay in case you didn't hear it correct (i.e. news) or even pause it while you run in the store. However, if you lose the signal, you have lost the signal. Further, if you change the channel then you've dumped the 30 minutes of save. It can't possibly cache all the channels.
Take a minute to actually read what I wrote, I didn't say ECC correction. I said that there must be some bandwidth left on the table, and to use that to send redundant data down.
And no, it wouldn't take a huge amount of memory, in fact, if your XM radio can store 30 minutes of one channel, that would be enough to cache about 10 seconds of 150 channels, it's a pretty trivial amount by todays standards.
At least take 60 seconds to read what was written before dismissing it out of hand. It wouldn't increase their costs a penny to do it properly from the start. Memory is dirt cheap these days, otherwise they wouldn't have sprung for the memory to store a channel for 30 minutes, thanks for proving my point without realizing it.


