Thetis v2.8.11 - UPDATES and current issues

[W1RHS]

Hi Richie,

Running Version 21K9(rc5b). No instances of 0x8899000C in my error log file. THANK YOU for all your fine work.

73

Rick, W1RHS

[w-u-2-o]

Nothing to report in my error log.

[Joe]

Richie,
Nothing to report in my error log as well. I had reported the vac buffer run-away earlier but today went and re-loaded (rc5b) and all is smooth now, I even could lower both ring buffers to 5 without any glitches - sample rate at 19200 with buffer size at 64, working very stable with this update.


Thanks and 73's,
Joe
WD5Y

[vk1hx]

Something to brighten my very dreary mood recently, Ernst OE3IDE has sent over a nice new splash, and it looks great with progress bar and eta to completion time. It will be in rc5 at some point before/during the weekend.

splash1.png

splash1.png (229.03 KiB) Viewed 10653 times

Richie.

This new splash looks great and nice work Ernst. Only one comment, can the text under Thetis be updated to something less obvious, like "ANAN SDR Software" or just "SDR Software"? IMO.

Also no error log here either.

[K1LSB]

Stupid question here, where would I find the file ErrorLog.txt? I can't find a file of that name anywhere.

Thanks,
Mark

Either

C:\Users\<you>\AppData\Roaming\OpenHPSDR\Thetis-x64
or
C:\Users\<you>\AppData\Roaming\OpenHPSDR\Thetis

depending if you run 64bit or 32bit versions

Richie.

Hmm, no Error*.* found anywhere. Does that possibly mean I've never logged any session errors?

[vk1hx]

Stupid question here, where would I find the file ErrorLog.txt? I can't find a file of that name anywhere.

Thanks,
Mark

Either

C:\Users\<you>\AppData\Roaming\OpenHPSDR\Thetis-x64
or
C:\Users\<you>\AppData\Roaming\OpenHPSDR\Thetis

depending if you run 64bit or 32bit versions

Richie.

Hmm, no Error*.* found anywhere. Does that possibly mean I've never logged any session errors?

No errors?

[K1LSB]

I can find no ErrorLog.txt anywhere on my computer.

Mark

[K9CDR]

Good afternoon Richie,

Does RC5 run on win7 ? A friend logged in to my computer last night and uploaded via teamviewer, backed up last version 2.8.11 and installed it. Upon opening got an error message. Shut program and rechecked everything then started again with same result. Restored to previous version.

Thanks in advance,
Harry W2QAQ

I don't know about WIN7 compatibility but you do have hold down the control key when starting Thetis after unpacking the update zip file into the install folder and overwriting files with same name.


***Richie no errors in my log with that description***

[w-u-2-o]

Is it time to remove "report -ve errors" from the seq log?

[ULTIMAX]

Is it time to remove "report -ve errors" from the seq log?

Better yet how about giving the operator the option to turn that off if they wish, Once in a while I get that sequence errors for some reason in 80Meters and apparently is not necessary to worry about it, so it would be nice to be able to turn any error messages from showing up totally.

mho

[Chiron]

Hello Richie,


I appreciate your good work on Thetis. It works very great. It is fast, low latency, stable, nice new features etc.
Also RC5 works like a charm.
Your large amount of video's do help to use (for me) new features. Some features are there already for a long time, but I never used it. And also the new features introduced by you are great. It makes the usage more practical for me.

It is a joy to see that you are so motivated to improve existing and add new features. It motivates me to test more features I never have used or new where existing.
I'm not sure if you have seen my previous post. It might be not for this forum 'subject' because it is a feature/change request concerning puresignal. You where doing some improvements/changements on puresignal, and I thought that it would be good to add something to be able to combine it with Anan and Hermes-lite2.
It would be great to be able to switch between Anan 7000DLE and HermesLite2 in Thetis without manually change the SetPk value when I change between the sets.
So far all is also perfectly working with the HermesLite.
The timing of the audio (buffer latency) can be set to 12ms, with stable sound. It makes using VOX possible, without notice of others.

I’m looking forward to more updates from you.

Regards,
Chiron

@ramdor Richie

PureSignal usage is great.
For the anan with 16 bit ADC is the default value for SetPk = 0.4072

When switching over to hermesLite2 with has a 12 bit ADC the SetPk = 0.4072 is not the propper value. It must be something like 0.238

In order to get 'PureSignal correction' to work the SetPk needs to be set to a value like 0.23. After doing this manualy all works fine. But after starting Thetis again, the default value is 0.4072. Thats perfect for the Anan. But not working for HermesLite2.

Is it possible to save the SetPk value together with the IP address of the radio. And at starting Thetis, recalling the SetPk value corresponding to the IP address of the selected radio?

Chiron

[K9CDR]

Is it time to remove "report -ve errors" from the seq log?

In case that might have anything to do with me, my notation was pertaining to Richie's request for a specific errror a few posts back. If not, then obviously disregard

[W2QAQ]

Good afternoon Richie,

Does RC5 run on win7 ? A friend logged in to my computer last night and uploaded via teamviewer, backed up last version 2.8.11 and installed it. Upon opening got an error message. Shut program and rechecked everything then started again with same result. Restored to previous version.

Thanks in advance,
Harry W2QAQ

I don't know about WIN7 compatibility but you do have hold down the control key when starting Thetis after unpacking the update zip file into the install folder and overwriting files with same name.


***Richie no errors in my log with that description***

Thank you for taking the time to reply to my inquiry. The control button was held down. The installation was attempted twice.

[K9CDR]

Good afternoon Richie,

Does RC5 run on win7 ? A friend logged in to my computer last night and uploaded via teamviewer, backed up last version 2.8.11 and installed it. Upon opening got an error message. Shut program and rechecked everything then started again with same result. Restored to previous version.

Thanks in advance,
Harry W2QAQ

I don't know about WIN7 compatibility but you do have hold down the control key when starting Thetis after unpacking the update zip file into the install folder and overwriting files with same name.


***Richie no errors in my log with that description***

Thank you for taking the time to reply to my inquiry. The control button was held down. The installation was attempted twice.

Well I think the most important information is missing from you yet. I think the actual error message would be quite relevant in the matter ! That would certainly point the group in the group in the right direction at least if not the specific cause.

[W2QAQ]

I don't know about WIN7 compatibility but you do have hold down the control key when starting Thetis after unpacking the update zip file into the install folder and overwriting files with same name.


***Richie no errors in my log with that description***

Thank you for taking the time to reply to my inquiry. The control button was held down. The installation was attempted twice.

Well I think the most important information is missing from you yet. I think the actual error message would be quite relevant in the matter ! That would certainly point the group in the group in the right direction at least if not the specific cause.

I agree Cory, unfortunately the computer was under remote control at the time and my friend deleted it both times before I was able to screen cap it. I'll see what I can do to regenerate it and when I do I'll post it.

Thanks again,
Harry

[K9CDR]

Thank you for taking the time to reply to my inquiry. The control button was held down. The installation was attempted twice.

Well I think the most important information is missing from you yet. I think the actual error message would be quite relevant in the matter ! That would certainly point the group in the group in the right direction at least if not the specific cause.

I agree Cory, unfortunately the computer was under remote control at the time and my friend deleted it both times before I was able to screen cap it. I'll see what I can do to regenerate it and when I do I'll post it.

Thanks again,
Harry

I dont know if any kind of startup errors like you had would be logged or if that only happens when the program is actually running after successful startup, but you may check this folder for an errorlog.txt file, if one would exist yet in your case--C:\Users\--YOURUSERNAME---\AppData\Roaming\OpenHPSDR\Thetis-x64 assuming 64 bit machine

Was kind of hoping someone would've chimed in by now as to your original question RE:WIn7 if would even work or if this is a dead end...

[DL2XY]

WIN7 SP1 64bit is fully supported.

[ramdor]

Anyone want VAC var ratio 'Forced' as part of the TX Profile. It would require a DB update, and be the first since 21f. I always run them forced when using Muzychenko's virtual cables and datamodes.

Richie.

[w4cm]

Multi-Meter Eye Candy would be nice.

[ramdor]

Multi-Meter Eye Candy would be nice.

So would some time to do it, they wont be in 21k9 sadly.

Richie.

[kc2rgw]

Anyone want VAC var ratio 'Forced' as part of the TX Profile. It would require a DB update, and be the first since 21f. I always run them forced when using Muzychenko's virtual cables and datamodes.

Richie.

One of the snags with a lot of the sub features tied to TX profiles comes in when a user has multiple profiles and has to toggle back through to re-set them all. An example recently, I was playing with the DSP tab and all the params there, so I had to toggle each TX profile and make the multiple tweaks to each of them and save again. It's magnified if changes span tabs and you then want to apply to multiple profiles.

I use profiles for 3k 3.5k 4k 5k AM etc with varying TX widths, but they are all common to SSB so all the underlying things like DSP tab settings, gain settings, the fundamentals are all common across them.

To the point of digital, I run 48k and 2048 for digital as I found fldigi doesn't like sampling or buffer conflicts.

With using VAC for both mic input and also for digital, you get a kind of complex web of the saved profiles settings.

This kind of logic sieve as to how to structure all these relationships is precisely why I never became a software developer beyond scripting LOL, my working memory doesn't deal with trees beyond about two steps deep.

Would it make sense to tie the collection of options to the mode of operation, SSB, DIGU/L etc to maybe avoid as much "ohhh I forgot that my 5k profile didn't get the updated gain changes or filter types that I made a few weeks ago."?

[DL2XY]

Anyone want VAC var ratio 'Forced' as part of the TX Profile. It would require a DB update, and be the first since 21f. I always run them forced when using Muzychenko's virtual cables and datamodes.

Richie.

Personally i would prefer an option to completely bypass the SRC (as it was in PowerSDR-MRX).
Reason: Even if "Forced" the converter is creating minor artifacts (phase jitter) due to the involved interpolation between two very similar but slightly different samplerates. Also the ringbuffer is creating unneccesary latency.

Another topic related VAC:
I cannot use my wireless headset directly with Thetis VAC. It has an unsymmetric interface with MIC: 16000 S/s 1 CH and Phones: 48000 S/s 2 CH.

Yet another topic related VAC:
Implementation of ASIO interface is quite "minimalistic". It supports only a single instance and is restricted to only the first input or output pin.
But an ASIO interface can have up to 32 inputs and output channels and the related driver pins.
It would be fine if one could select specific pins and share them between VAC1 and VAC2 simultaneously.

[w-u-2-o]

Personally i would prefer an option to completely bypass the SRC (as it was in PowerSDR-MRX).
Reason: Even if "Forced" the converter is creating minor artifacts (phase jitter) due to the involved interpolation between two very similar but slightly different samplerates.

When "forced" AFAIK there is no conversion, and no interpolation done. There are only the input and output buffers, which will under or over-run occasionally depending on the difference between audio sample rate in the ANAN hardware and the audio sample rate of the PC hardware, since neither will be at exactly 48KHz unless they are both phase locked to a common reference oscillator. Thus there would be no functional difference between "forced" and no resampler at all. There are, however, actions that are taken in the event of an under or over-run which endeavor to smooth the transition associated with the lost data. Is it not better to have those transitions smoothed?

Also the ringbuffer is creating unneccesary latency.

I know that both Warren and Bryan have looked at this extensively. They have even compared the approaches used in highly regarded audio software like Ardour. For a variety of reasons that I honestly don't remember they were unable to improve upon the approach used.

[K1LSB]

Walt,

What's an SRC?

Thanks,
Mark

[John]

In current Thetis RC version 21k9rc5b programs like WSJT-X that send CAT commands to VFO B for frequency shift on/during transmit - Thetis assigns VFO B freq also to VFO A. Reverting back to an older version like 21k7 the "issue" is not exhibited. VFO Sync is not enabled. VFO A lock does keep VFO A from changing. Perhaps I am missing something?

[ramdor]

In current Thetis RC version 21k9rc5b programs like WSJT-X that send CAT commands to VFO B for frequency shift on/during transmit - Thetis assigns VFO B freq also to VFO A. Reverting back to an older version like 21k7 the "issue" is not exhibited. VFO Sync is not enabled. VFO A lock does keep VFO A from changing. Perhaps I am missing something?

works fine here, as there have been no changes made to CAT. VFOb is duped to VFOa for TCI outbound messages if the setting is set like that, but has no relation to CAT.

Something else going on, or some other external program is causing it.

Tested here : https://youtu.be/mJ9rp3K1RuE (hit play if that video is not working correctly)

Richie.

[John]

Hel Richie,

Yes your are correct. I was using Ham Radio Deluxe as a layer between WSJT-X and Thetis. I did this because I use multiple rig types, simplifies switching rigs. Something in HRD is causing the issue. Thanks and sorry for the waste of your time!
John

[KA5KKT]

Walt,

What's an SRC?

Thanks,
Mark

When using acronyms that may not be part of many people's everyday language, it would be polite to once, early in a post or perhaps OP (Original Post) to actually spell out the verbiage of the acronym.

VFO is an example of one that probably never needs to be spelled out. Per K1LSB, I appreciate his question. SRC seems to be one that should be.

[DL2XY]

Walt,

What's an SRC?

Thanks,
Mark

Sorry for confusion, SRC means Sample Rate Converter.

[DL2XY]

Personally i would prefer an option to completely bypass the SRC (as it was in PowerSDR-MRX).
Reason: Even if "Forced" the converter is creating minor artifacts (phase jitter) due to the involved interpolation between two very similar but slightly different samplerates.

When "forced" AFAIK there is no conversion, and no interpolation done. There are only the input and output buffers, which will under or over-run occasionally depending on the difference between audio sample rate in the ANAN hardware and the audio sample rate of the PC hardware, since neither will be at exactly 48KHz unless they are both phase locked to a common reference oscillator. Thus there would be no functional difference between "forced" and no resampler at all. There are, however, actions that are taken in the event of an under or over-run which endeavor to smooth the transition associated with the lost data. Is it not better to have those transitions smoothed?

Also the ringbuffer is creating unneccesary latency.

I know that both Warren and Bryan have looked at this extensively. They have even compared the approaches used in highly regarded audio software like Ardour. For a variety of reasons that I honestly don't remember they were unable to improve upon the approach used.

Here are some high resolution measurements of VAC output. Test signal is 1kHz from Thetis receive path test generator, so 0.001 Hz offset is 1ppm samplerate error.

First setling from forced to normal operation( big scale):

setling.jpg (105.49 KiB) Viewed 10076 times

Normal operation in detail (zoomed Y):

Normal.jpg (152.23 KiB) Viewed 10076 times

Forced to 1.000020 (the mean error on my setup, to maximize time intervals between clicks):

forced20ppm.jpg (144.66 KiB) Viewed 10076 times

as we can see, the "resolution" of the converter is about 50ppm, it is switching (dithering) between two values. Each slope maybe a relevant phase jump for some digital modes.
Due to the integer behavior of conversion resolution it is not capable to dither around zero, so a mean error of 25ppm remains.

Forced to 1.0000000:

forced1.jpg (134.9 KiB) Viewed 10076 times

Indeed Spot on with only 1ppm offset (which may be rounding error of my PLL demodulator, its integer calculation too).
But we still have half ringbuffersize/samplerate latency, which will change over time to 0 or full ringbuffersize/samplerate until next overflow or underflow.

Walter