Welcome, Guest
Username: Password: Remember me

TOPIC: JTDX with TCI

JTDX with TCI 24 Feb 2021 07:43 #1

  • g0cgl
  • g0cgl's Avatar
  • OFFLINE
  • Erik EI4KF / G0CGL
  • Posts: 390
  • Thank you received: 206
Thanks to Igor UA3DJY (and EE who helped make this happen), the current beta JTDX and the next official release has TCI CAT and audio! Information given to me is that providing you keep the SDR filter latency low (medium level filter taps), testing shows there is now almost no latency (the killer of digital especially FT8/4) and the use of TCI removes the errors introduced by VAC software. Testing has been quite exhaustive with various settings checked and there is certainly a worthwhile performance uplift.

Thanks to all concerned, I'd been hoping to see TCI implemented for a long time.
The administrator has disabled public write access.
The following user(s) said Thank You: Rome, iw7dmh, G4ENZ, LA3QMA, PD0DH, UA4FX

JTDX with TCI 24 Feb 2021 10:30 #2

  • M0CVX
  • M0CVX's Avatar
  • OFFLINE
  • Posts: 18
  • Thank you received: 3
Great stuff! I've been checking the JTDX website for weeks :-)
Well done all involved!
I'm new to SSDR, so what would you regard as a 'mid' filter taps setting, Erik?
Last Edit: 24 Feb 2021 10:42 by M0CVX. Reason: Added a question
The administrator has disabled public write access.

JTDX with TCI 24 Feb 2021 10:53 #3

  • g0cgl
  • g0cgl's Avatar
  • OFFLINE
  • Erik EI4KF / G0CGL
  • Posts: 390
  • Thank you received: 206
Hi - for TX Filter Taps on DIGL/DIGU I have set 2000. I think the default is 1537 so 2000 gives a sharper filter shape without any discernable increase in latency. I have the same 2000 for receive DIGL/DIGU filter taps. For SSB amd CW, latency is not such as issue so it is quite reasonable to increase taps well above 2000, the higher the number the more 'sheer' the filter side but testing on FT8 decodes shows that high numbers and steep sides makes the number of decodes per receive period on average less which you do not want. 2000 is somewhat arbitary but seems a resonable compromise based on some hours of testing.
The administrator has disabled public write access.

JTDX with TCI 24 Feb 2021 12:35 #4

  • M0CVX
  • M0CVX's Avatar
  • OFFLINE
  • Posts: 18
  • Thank you received: 3
Brill. Thanks Erik, I'll try digu taps at 2000 then.
Also, a public thanks for your Addendum document; I've already set up VAC and Com0Com plus two JTDX decoders (ready for 50.313 and 50.323), but would have struggled to do that without that document. Hopefully the new TCI interface in JTDX will make a new twin-decoder setup even easier!
Regards and see you on Six before we know it.. ;-)

Paul M0CVX
The administrator has disabled public write access.

JTDX with TCI 24 Feb 2021 12:54 #5

  • g0cgl
  • g0cgl's Avatar
  • OFFLINE
  • Erik EI4KF / G0CGL
  • Posts: 390
  • Thank you received: 206
Time flies doesn't it Paul?! 6m is nearly here again. Let's hope the season is as good as in 2020.

Thanks for your comments re the UMA. A big re-write is being done to coincide with ESDR3.

2x JTDX, one on each receiver, is as you say vital for 6m. It should be a breeze since TCI will accept any number of clients. I haven't tried it yet though - will audio channels on one client be separate from those on the 2nd client? My ageing brain cannot compute hi - will have to test it.

Erik.
The administrator has disabled public write access.
The following user(s) said Thank You: M0CVX

JTDX with TCI 24 Feb 2021 22:42 #6

  • Gian
  • Gian's Avatar
  • OFFLINE
  • HB9FAR
  • Posts: 8
...is the jtdx rc 156 release already available..?
HB9FAR
The administrator has disabled public write access.

JTDX with TCI 25 Feb 2021 06:57 #7

  • M0CVX
  • M0CVX's Avatar
  • OFFLINE
  • Posts: 18
  • Thank you received: 3
I don't think so (yet) Gian. Maybe it is currently a Beta release for those more closely involved in JTDX. I suspect a new RC release isn't far away...
73
Paul M0CVX
The administrator has disabled public write access.
The following user(s) said Thank You: Gian

JTDX with TCI 25 Feb 2021 07:40 #8

  • g0cgl
  • g0cgl's Avatar
  • OFFLINE
  • Erik EI4KF / G0CGL
  • Posts: 390
  • Thank you received: 206
Yes I am testing the beta and the official release will follow soon, though I do not know exactly when that will be. There is a small issue to be corrected for TCI CAT/Audio operation but otherwise it is working well. One of the testers is checking the 2 instance, one per receiver, scenario today so more news on whether that works will hopefully be forthcoming.
The administrator has disabled public write access.
The following user(s) said Thank You: Gian

JTDX with TCI 25 Feb 2021 09:06 #9

  • g0cgl
  • g0cgl's Avatar
  • OFFLINE
  • Erik EI4KF / G0CGL
  • Posts: 390
  • Thank you received: 206
So, actually the 2nd receiver does not work. ESDRx requires a second TCI client with TCI to include separate audio stream.

In the latest ESDR3 promotional video, Vasily claims JTDX with TCI "just works, no issues" but there seems to have been an oversight here.

So therefore in practice, unless there is a change to TCI implementation, the 2nd receiver will need VAC.
The administrator has disabled public write access.

JTDX with TCI 25 Feb 2021 09:20 #10

  • M0CVX
  • M0CVX's Avatar
  • OFFLINE
  • Posts: 18
  • Thank you received: 3
Ah..well at least it has been found. I hope a fix is in the offing. If not, I know that my 2 x VAC/Com0Com set up works...but I think you were implying that TCI audio was more performant than VAC Erik?
The administrator has disabled public write access.

JTDX with TCI 25 Feb 2021 10:13 #11

  • g0cgl
  • g0cgl's Avatar
  • OFFLINE
  • Erik EI4KF / G0CGL
  • Posts: 390
  • Thank you received: 206
There is less latency with TCI, there is no intermediate VAC software in the audio chain. Also with VAC, which was not designed for our purposes, it is nearly impossible to remove latency especially on transmit. Using TCI the TX is perfect. On receive the performance is at least as good as via VAC and some testers say there is small improvement in FT8 decode numbers per reieve cycle on average so long as the SDR filter latency is at a reasonable minimum. This was after some hours of testing. Another issue that the new JTDX resolves is that historically subsequent Windows updates have led to more and more issues with the VAC software so being able to dispense with it is a bonus.
The administrator has disabled public write access.
The following user(s) said Thank You: M0CVX

JTDX with TCI 26 Feb 2021 12:28 #12

  • N8SDR
  • N8SDR's Avatar
  • OFFLINE
  • Posts: 173
  • Thank you received: 141
Erik- Hope all is well and thanks for the info- However I experience the opposite effect with filter taps- also note a cleaner signal is produced with a higher filter tap as well- my personal observation so far is that using 7168 on all filter taps TX and RX in any mode results in cleanest produced signals- as the cleaner signals easier to decode as it is a more tighter signal making it easier to decode and not being interfered with by closer by stations using the same or very close digital frequency especially on the received end. Again this is what I have experienced and I'm sure there can be exceptions to this- ( also I use 7168 as it is the highest a multiple of 1024 that is currently available in the EESDR application- multiples of 2 128,256,512,1024 are faster for a CPU to process )

My thought in this is a cleaner signal with less IMD is easier to decode than one with IMD
That said- if a users PC isn't up too the task of running tighter/sharper filtering because it uses more system resources then the PC can handle then yes stepping down in size of filtering would be better as that would not introduce more lag and possible stuttering of the VAC and and other applications involved because you have reduced the load on the CPU.
Please do not email me and ask for help I no longer will be part of this, If I could delete /remove my profile I would do so.
Last Edit: 26 Feb 2021 12:30 by N8SDR.
The administrator has disabled public write access.

JTDX with TCI 26 Feb 2021 13:34 #13

  • g0cgl
  • g0cgl's Avatar
  • OFFLINE
  • Erik EI4KF / G0CGL
  • Posts: 390
  • Thank you received: 206
That figure may be valid for SSB and CW. There is an EE document that shows the shape of the filter as the taps number is increased and effectively such a number is a rectangular shape. What that document lacks though is the associated increase in latency which is why EE chose 1537 as the default. So some testing was done to maximise FT8 decodes on crowded bands and circa 2000 is the proven optimum.

I passed this on as information and not a prescription and I dare say that casual operators will not care very much either way. That said, reducing latency is quite important when using TCI audio because JTDX has to buffer the audio stream in the TCI interface (WebSocket audio stream) and this buffering brings some delay of its own.
Last Edit: 26 Feb 2021 13:59 by g0cgl.
The administrator has disabled public write access.

JTDX with TCI 06 Apr 2021 09:39 #14

  • S58WW
  • S58WW's Avatar
  • OFFLINE
  • A half-truth is the most cowardly of lies.
  • Posts: 74
  • Thank you received: 26
Hello all!

Can anyone tell me how to get involved in testing new JTDX 2.2.0-RC156? with TCI support?
I am member of JTDX group and I know that some of the members from here are testiing new JTDX.
It is not the "must" issue, but I am just wondering.

Rome can you put me on list for EESDR3 testers please?

SRI for my English - not native speaker. :pardon:
73 Pibo S58WW
SUNSDR2DX
Main shack PC:
AMD RYZEN 7 2700 (8 cores ; 16 threads) 32 GB RAM
on Gigabyte Aorus Elite B450 with 1TB Samsung 980 NVMe M.2
NVIDIA GeForce GTX 1660 SUPER
The administrator has disabled public write access.

JTDX with TCI 06 Apr 2021 10:57 #15

  • g0cgl
  • g0cgl's Avatar
  • OFFLINE
  • Erik EI4KF / G0CGL
  • Posts: 390
  • Thank you received: 206
Email Igor UA3DJY and ask him, saying that you are an ESDR2 user.

This email address is being protected from spambots. You need JavaScript enabled to view it.

Most testing is done now, it is working very well. But I am sure he will be pleased to get another tester as you never know what oddity can appear with different computers, radios, etc.

Meanwhile I have emailed you the link for downloading the latest test version with 32bit sample size audio. I cannot post the link here.
The administrator has disabled public write access.
The following user(s) said Thank You: S58WW
Time to create page: 0.189 seconds