Discussion in 'Amateur Radio News' started by NN2X, Oct 7, 2018.
Your very elaborate and eloquent display of your deep technical insight is very impressive.
Sounds like another XP, or worse, Linux fanboy.
Won't be long we'll all be booting to the internet (of which there will be several, like phone books used to be) and it'll be full-circle back to dumb-terminals.
Ah let the name calling commence. Interesting insights provided here, although none of technical nature. Social sciences are fascinating too though.
Like how some folks like to ride high on their horse?
Not as if you've contributed much of anything but commentary on other people's comments.
I'm happy to accept any comments on JS8 if you're interested in moving back on topic
This appears to be comparing apples and oranges. It is critically important to specify the bandwidth over which the measurement was taken. In the case of the PSK31 article, the bandwidth was 100 Hz. The WJST-X modes use a bandwidth of 2,800 Hz.
With PSK31 referenced to 100 Hz bandwidth and assuming JS8CALL uses the 2,800 BW of WJST-X, this makes JS8CALL look about 14 dB better. In other words, the difference between the two is not 6-(-24) = 30, but 30 - 14 = 16 dB. The formula is 10log(BW1/BW2). It should be noted that PSK31 is one of the poorest digital modes as far as performance.
So if you're using PSK31 at 1000 watts, you'd need 25 watts.
If we used the extra 30 WPM for Forward Error Correction (FEC), the difference would be reduced by another substantial margin. Maybe 10 dB? I didn't do the calculations. FEC is not my thing.
Bottom line, there's no magic here.
You have absolutely no idea of my technical insight and experience. I've been networking and supporting computers and various operating systems for three decades.
Wang, IBM, UNIX, Linux, OS/2 MACos, DOS3 and up, every version of Windows, all of it.
Micro$oft has a bad habit of Beta-Testing their OS's on the paying consumer.
To this day, windows 10 is buggy, features are suddenly removed or downgraded, software disappears, every update is a crapshoot as to whether mission-critical applications will continue working on the platform.
Definitely NOT a good operating system to run in a production environment . AKA Garbage. Thanks for your "input"
Just a small matter of precision (or nitpicking): According to the WSJT-X User Guide § 7.1, the reference bandwidth for FT8 is 2,500 Hz, i.e. 25X the PSK31 bandwidth in the article and 50X the FT8 bandwith (and then also the JS8CALL bandwidth). The difference between 2,500 and 2,800 Hz amounts to a mere 0.5 dB, so it is more than close enough for this discussion.
With FT8 signals at the basis of JS8CALL, I guess they should be compared at identical bandwidths. Assuming even distribution of noise over the entire bandwidth, the 50 Hz bandwidth of JS8CALL and FT8 gives them a 17 dB noise advantage over 2,500 Hz bandwidth, while the referenced 100 Hz bandwidth of PSK31 would yield a 14 dB noise advantage over a 2,500 Hz bandwidth. Considering bandwidth alone, PSK31 should only be -3 dB compared to JS8CALL and FT8, but then we are not considering error correction in JS8CALL and FT8, nor the higher WPM, the free format and no need of time synchronization of PSK31. As you say, Chris, there is no magic, just adding more weight to some features at the cost of some other. All the modes are well thought out, making good use of the available computer capacity of their time. I guess it will be difficult to find a way to compare apples to apples, i.e. in a way that is fair to their various characteristics and primary uses. We may have to settle for comparing oranges to lemons or mandarin oranges, which, although they are close and related, will leave a touch of personal preferences in the (in)equation.
73 Frode LA6VQ
The magic is the following...There is no other digital coding scheme that can decode at -24dB below the noise floor, at 20wpm, while only occupying 50hz...
That is the magic!
With that impressive background of legacy technology, we'd love to have your feedback. If you could spare a few minutes, that would be great.
73 Mike K5TRI
p.s.: Yes, I work at Microsoft