Date
1 - 14 of 14
Locked How does transmit time impact decoding effectiveness with FT8? #decode
|
Tim K9WX
How late can I start transmitting an FT8 signal in the 15-second cycle before it becomes a waste of time?
I can score PSKReporter hits by starting my TX6 CQ message as late as 8 seconds into a cycle, and WSJT-X will let me start a TX at least as late as 10 seconds into a cycle, although I have not seen any PSKReporter hits with a delay of longer than 8 seconds. How is the likelihood of a message being decoded reduced with by a delayed start? Tim K9WX |
|
|
|
Alan B
pskreporter hits come from other stations successfully decoding your
toggle quoted message
Show quoted text
transmission. So you have already determined your answer. 73 de w6akb On Thu, Jul 14, 2022 at 12:29 PM Tim K9WX <timk9wx@...> wrote:
How late can I start transmitting an FT8 signal in the 15-second cycle |
|
|
|
Michael Black
Hamspots.net is bit better/faster if you're trying to track this.
toggle quoted message
Show quoted text
Mike W9MDB On Thursday, July 14, 2022 at 04:08:17 PM CDT, Alan B <alan.biocca@...> wrote:
pskreporter hits come from other stations successfully decoding your transmission. So you have already determined your answer. 73 de w6akb On Thu, Jul 14, 2022 at 12:29 PM Tim K9WX <timk9wx@...> wrote: How late can I start transmitting an FT8 signal in the 15-second cycle |
|
|
|
Tim K9WX
Thanks, Alan. Yes, it seems an 8 second delay can still result in a successful decode on the RX end. But I was hoping to gain a bit more insight. Can it be done at 9 seconds? Ten seconds? Where does it go to zero? How quickly does the success rate fall off as delays are introduced? Does the content of the message itself influence success? Are free-format messages more likely to fail than standard messages for a given delay?
Tim K9WX |
|
|
|
Reino Talarmo
Hi Tim
Many nice questions. I try to give some generic information. But I was hoping to gain a bit more insight. Can it be done at 9 seconds? Ten seconds? Where does it go to zero? How quickly does the success rate fall off as delays are introduced?WSJT-X uses a powerful forward error correction and synchronization patterns and in a S/N situation decoding probability is high as long as there are enough bits received both information containing bits and synchronization bits. If either fails, then no decoding is possible. That happens as you noted at about 8 seconds clipping from the start of the message. Does the content of the message itself influence success?Simple answer is no. Are free-format messages more likely to fail than standard messages for a given delay?The same answer. Situation changes, when S/N goes down as some of the received bits are no more received correctly and the error correction or synchronization fails. The failing error correction is detected using an error detection part of the message and also that may fail and you will receive ghost messages. You may study that a bit more in article https://physics.princeton.edu/pulsar/k1jt/FT4_FT8_QEX.pdf 73, Reino OH3mA |
|
|
|
JTAlert Support (VK3AMA)
On 15/07/2022 12:27 pm, Tim K9WX wrote:
Are free-format messages more likely to fail than standard messagesFYI, If you're looking at PSKReporter only CQ decodes are uploaded, so if you're testing using free-format messages they wont get spotted. de Laurie VK3AMA |
|
|
|
On Thu, Jul 14, 2022 at 07:27 PM, Tim K9WX wrote:
... Yes, it seems an 8 second delay can still result in a successfulDefinitive measurements of the probability of decoding an FT8 message when the transmitting station starts transmitting after the beginning of a 15 s interval were published in QEX about two years ago. See https://physics.princeton.edu//pulsar/k1jt/FT4_FT8_QEX.pdf Figure 9 and associated text have the details. Yes, the content of the message matters, because it determines whether AP decoding can be used. -- 73, Joe, K1JT |
|
|
|
Alan B
I find that responding to CQ's gets logged into PSKReporter, as well as
CQ's. 73 de w6akb On Thu, Jul 14, 2022 at 11:11 PM JTAlert Support (VK3AMA) < vk3ama.ham.apps@...> wrote: On 15/07/2022 12:27 pm, Tim K9WX wrote:Are free-format messages more likely to fail than standard messagesFYI, If you're looking at PSKReporter only CQ decodes are uploaded, so |
|
|
|
neil_zampella <neilz@...>
Um .... PSKReporter doesn't log anything, it just reports what you've
toggle quoted message
Show quoted text
decoded, which includes CQs and replies. Neil, KN3ILZ On 7/15/2022 9:37 AM, Alan B wrote:
I find that responding to CQ's gets logged into PSKReporter, as well as |
|
|
|
Michael Black
If the message contains a grid -- i.e. TX6 or TX1 -- then pskreporter will allow reporting the transmitting station.
toggle quoted message
Show quoted text
If you skip TX1 or don't have a grid in TX6 (e.g. special call) it won't be uploaded to pskreporter as there is no grid. Pskreporter relies on the grid for the maps. Mike W9MDB On Friday, July 15, 2022 at 09:37:58 AM CDT, Alan B <alan.biocca@...> wrote:
I find that responding to CQ's gets logged into PSKReporter, as well as CQ's. 73 de w6akb On Thu, Jul 14, 2022 at 11:11 PM JTAlert Support (VK3AMA) < vk3ama.ham.apps@...> wrote: On 15/07/2022 12:27 pm, Tim K9WX wrote:Are free-format messages more likely to fail than standard messagesFYI, If you're looking at PSKReporter only CQ decodes are uploaded, so |
|
|
|
Gary - AG0N
On Jul 14, 2022, at 8:27 PM, Tim K9WX <timk9wx@...> wrote:It’s my understanding that the transmitted signal contains a certain amount of FEC. FEC doesn’t work correctly without all of the data, so if you start late, there’s no “backup” to get the message through. What it boils down to is if you start later than the point at which all of the “package” gets sent at least once during the sequence, you won’t be decoded. But, if you are strong enough at the receive end that at least one complete package is received by him you will be decoded. Yes, you can start a little less than half way through a sequence, but you’ll only be decoded by those stations that are receiving you with good signal. You need to know something about how data is transmitted to fully grasp it. If you dissected a data “packet” or package, you would see it isn’t just a simple text message. Gary - AG0N |
|
|
|
William Smith
While WSJT decoding works if you start late (or stop early, or drop a few seconds in the middle, or some combination), the error correction is more "not all of the transmission was copied, but I can figure it out with the error correction". For a perfect signal you don't need to get the entire message through, but if you are going to have dropouts of varying types, you'll want to send as much of it as possible, so that the receiver will have a better chance of getting the message.
toggle quoted message
Show quoted text
73, Willie N1JBJ On Jul 15, 2022, at 10:48 AM, Gary - AG0N <wb0kkm@...> wrote:On Jul 14, 2022, at 8:27 PM, Tim K9WX <timk9wx@...> wrote:It’s my understanding that the transmitted signal contains a certain amount of FEC. FEC doesn’t work correctly without all of the data, so if you start late, there’s no “backup” to get the message through. What it boils down to is if you start later than the point at which all of the “package” gets sent at least once during the sequence, you won’t be decoded. But, if you are strong enough at the receive end that at least one complete package is received by him you will be decoded. Yes, you can start a little less than half way through a sequence, but you’ll only be decoded by those stations that are receiving you with good signal. |
|
|
|
Randy, WS4C
Very helpful discussion here. Thanks to all!
Randy, WS4C |
|
|
|
Ummm.. Actually you CAN download a 24 hour or 1 week spot log of any specific callsign from PSKReporter if you know where and how to look.
73 -Jim NU0C On Fri, 15 Jul 2022 09:40:52 -0500 "neil_zampella" <neilz@...> wrote: Um .... PSKReporter doesn't log anything, it just reports what you've |
|
|