2.6.8

Helmut
Posts: 34
Joined: Sat Jul 15, 2017 3:08 pm

Re: 2.6.8

Postby Helmut » Fri Sep 20, 2019 3:17 pm

Thank you Richie vy much for the work you make us happy with 2.6.8.

Gary, I am testing the Notch in FT8 and have found it working fine as should be,
first i switched off in Setup DSP-MNF the Auto-increase, then with MNF active
shift+MNF for the 100hz Notch which comes on the red line and take the green
notch with the mouse to the station on the FT8 band which I have 3KHz wide.
After releasing the mouse the notch gets active on that station.
Just befor tested the attenuation on a station with +14 (notch off) to -8 (notch on).
Vy useful for strong stations to get deeper in noise for week stations.

I saw the Video, Richie, you draged the bandwith of the notch but could not find
how this works (what key to press) ?

Over all the 2.6.8 works here on my 7000dleMK2 stable, also with my Asio-Interface
and the Bodnar GPDSO in SSB with synchronised PC to 10MHz and Audio on 48KHz.

By the way, it was necessary to build a database from the Reset completly new by setting
all the Setups from screenshots of the 2.6.7 Version. The memory.xml could be exchanged
but first had to let save the file after setup and shutdown and than took the old memory.xml
into the 2.6.8 folder. That saved me a lot of work hi...

73 from
Helmut dl9eri/oe9eri
Joe
Posts: 42
Joined: Tue Aug 22, 2017 5:58 pm

Re: 2.6.8

Postby Joe » Fri Sep 20, 2019 3:29 pm

Richie,
I am amazed at what you have and are doing with Thetis! Also, very grateful, thank you. I can’t wait to get your latest downloaded when you finish with the tidying up. I hope you also get the source files updated as well, I enjoy reading and learning. I know you are quite busy and again thank you very much!

73’s
Joe
WD5Y
DL8LAQ
Posts: 87
Joined: Sun Apr 09, 2017 3:28 pm

Re: 2.6.8

Postby DL8LAQ » Fri Sep 20, 2019 6:17 pm

ramdor wrote:
NC3Z wrote:Should Thetis work with 4K resolution?

Try this version : https://www.dropbox.com/s/srhq9rm6rvlob ... 9.zip?dl=0

I have just quickly added an indicator, that will show top left of spectrum/waterfall. A red indicator will show if the render thread is not achieving requested fps. Turn your fps down until this red indicator turns off 99.9% of the time. I can just about manage 40fps at 4k full screen on a 680GTX when showing panafall (with disable picDisplay enabled and no spot background).

Everything is a trade off at the moment. GDI+ into a picture box is not the fastest, and until the DirectX engine is completed (yes I have started working on it... dont tell anyone), then it will be a case of managing fps vs responsiveness.

Richie.


This version doesn't work on my PC (W10pro, german language, NVIDIA GTX1050).

But it does after moving the database into trash :-)

Thanks Richie :-)
73, Norbert - DL8LAQ - ANAN-7000DLE
User avatar
w-u-2-o
Posts: 2066
Joined: Fri Mar 10, 2017 1:47 pm

Re: 2.6.8

Postby w-u-2-o » Fri Sep 20, 2019 7:57 pm

Richie,

First and foremost, fantastic work, thank you! When I am next in the UK I'm getting on a train and coming to buy you a pint! :)

Second: now running b7 with a from scratch database on a 4K monitor and a ripping fast PC.

Comments on the frame rate calculations:

Things definitely work better now. In fact, by allowing the user to keep the redraw rate below a proper limit, pops and clicks I had on audio when loading certain web pages has completely disappeared. Clearly there was some Thetis thread that had no, or negative, margin for scheduling, but now this does not happen as long as I do my part to keep the refresh rate at an appropriate level. This is an incredibly important improvement.

However, it is all too easy to exceed the allowable frame rate. My 4K monitor has a 30Hz refresh rate. By rights I should not be setting the update rate faster than 30Hz. I can actually push it up to 45Hz before I get any significant activity on the red indicator. 60Hz is, of course, of course a solid red indicator and comes with the attendant screen redrawing problems (and audio pops/clicks).

Since the code can detect the limits, would it be possible for the code to simply limit the maximum rate in the setup screen, thereby preventing the user from shooting himself in the foot? Similarly, the waterfall update time also seems to require the same limits.

Of course this will all go away once the spectral display area is done over with DirectX or similar.

Bug in CTUN mode:

When the mouse is placed in the frequency ribbon at the top of the panadapter, normally it would allow one to left click/drag and move the entire IF passband window around. Now it causes the receive passband to move left and right, and in a direction reverse to the movement of the mouse. However, occasionally I can choose just the right spot in the ribbon and obtain a left-right arrow cursor when I left click, under those conditions the entire IF passband can be moved properly.

Thanks & 73 Ritchie!

Scott
NC3Z
Posts: 207
Joined: Sun Oct 29, 2017 8:57 pm
Location: Merritt, NC

Re: 2.6.8

Postby NC3Z » Fri Sep 20, 2019 10:31 pm

Ritchie, are there any software hooks/command line arguments to allow Thetis to auto power on when started?
Gary NC3Z
User avatar
ramdor
Posts: 341
Joined: Wed Jul 03, 2019 3:07 pm

Re: 2.6.8

Postby ramdor » Fri Sep 20, 2019 10:47 pm

NC3Z wrote:Ritchie, are there any software hooks/command line arguments to allow Thetis to auto power on when started?


-autostart but i think it is broken atm
NC3Z
Posts: 207
Joined: Sun Oct 29, 2017 8:57 pm
Location: Merritt, NC

Re: 2.6.8

Postby NC3Z » Fri Sep 20, 2019 10:48 pm

OK, I had that in some notes but it never worked. Figured I would ask while you are neck deep in the code!
Gary NC3Z
User avatar
ramdor
Posts: 341
Joined: Wed Jul 03, 2019 3:07 pm

Re: 2.6.8

Postby ramdor » Fri Sep 20, 2019 10:57 pm

NC3Z wrote:OK, I had that in some notes but it never worked. Figured I would ask while you are neck deep in the code!


having a quick look now, but was about to do a release. If i can sort it in 30 mins or so i'll put it in.
NC3Z
Posts: 207
Joined: Sun Oct 29, 2017 8:57 pm
Location: Merritt, NC

Re: 2.6.8

Postby NC3Z » Fri Sep 20, 2019 10:59 pm

Dont let it delay anything.
Gary NC3Z
SA3ATF
Posts: 72
Joined: Mon Apr 10, 2017 9:42 pm
Location: Solleftea Sweden JP83PE
Contact:

Re: 2.6.8

Postby SA3ATF » Fri Sep 20, 2019 11:05 pm

We are eternally grateful to all the mods for sure, so I just say keep up the good work and it's so damn fun to get something new no matter the size of it all!
As someone previously said: it's like a new radio every day :lol:
I don’t suffer from insanity, I enjoy every second of it!
User avatar
ramdor
Posts: 341
Joined: Wed Jul 03, 2019 3:07 pm

Re: 2.6.8

Postby ramdor » Fri Sep 20, 2019 11:15 pm

Hi all,

Please try : https://www.dropbox.com/s/vuxtfoebucckx ... 9.zip?dl=0

NOTE: You will need to put the System.Buffers dll in the same folder as exe

Code going off to Doug later as I will be unable to do any work on it this weekend. Cheers Richie.

2.6.8 (9/17/19) (b8)
  • change: moved over to array pool system to reduce GC work (requires System.Buffers)
  • fix: dragging with top frequency strip when in CTUN mode (reported by Scott). Need to think about what happens when the filter falls out of sample rate area.
  • tempfix: -autostart flag on commandline will power radio on 2 seconds after Thetis window visible

Multi Manual Notch Tracking

This is first iteration, any other ideas for it, please let me know.

  • +MNF button adds 200Hz notch at VFOA frequency, <shift> +MNF will add 100Hz
  • when in right click tune mode (cursors), middle mouse will add 200Hz at mouse pos, <shift> middle will add 100Hz
  • hover over a notch and it will turn green and pop out text details
  • shift middle click when hovering will delete the notch
  • middle click when hovering will set notch to active/inactive and turn grey to show incative, text also turns red
  • shift left drag while hovering will resize (often quicker than finding the edges)
  • mouse wheel while hovering will resize +- 10Hz, <shift> wheel will resize +- 1Hz (***)
  • while hovering drag edges of notch to resize
  • MNF turns on/off notch system (turns them all feint yellow, even all the inactive ones which were previously grey)
  • GUI notch can not be used if setup notch filter is in Add/Edit mode
  • dynamic hit test so that a notch can be highlighted when 0Hz wide, and maximum zoom out

Notch Known issues
  • notch can be dragged below 0hz (wont apply the notch, but will look like it has)
  • notch can be dragged above max_freq (wont apply the notch, but will look like it has)
  • need to look into xvtr implications/implementation
  • no colour options for notches

Notch Future
  • It would be nice to see effect of notch(s) on waterfall and panorama. Whilst I could 'fake' this, it would be great if there was an option to activate this in WDSP.

(***) this resize is done on the notch instantly, all other notch changes are made on mouse button release

UI Known issues:
- using spot backgrounds has undesired effect on background when disable pic display used
- new link spectrum does not use correct values when band changed
- band panel will not resize correctly when app started in collapsed view. Due to panel.visible check being done before window itself is visible. Needs full rework of visible code
- if pan is used, and then TX without DUP enabled, active TX area can be off screen. Check zoom etc
- gap appears on top of filter when in panafall, related to step/top of grid
- size difference in compressed view of rx1/rx2 meters
- Special Panafall Mode (from spot) no longer works. (NC3Z) --- re-implement
- filter when ouside sample rate area needs to do something with the stream. Happens when CTUN scroll top area moves filter out of sample rate

ToDo:
- grab top/bottom areas of spectrum range sliders to only change the upper or lower limits
- inspect code paths for the SWR part of EDGE PWR_SWR
- re-work of VHF/HF/GEN panel(s) show/hide method
- peak blob(s)
- remove all dead/commented code (console.cs)
- hide date/time option
- improved puresignal enabled/disabled notification?
- add offset to waterfall range if link spectrum grid used
- further work required on 'spectrum scale drag'
- rework all band edge/info into single class, with DB

Other reported issues:
- 200D running 1.8 Orion firmware om Windows 7. TX bug ... crashes when tx/tune/mox/ptt, ok on RX (G3ZQH)
Last edited by ramdor on Sat Sep 21, 2019 12:11 am, edited 1 time in total.
NC3Z
Posts: 207
Joined: Sun Oct 29, 2017 8:57 pm
Location: Merritt, NC

Re: 2.6.8

Postby NC3Z » Fri Sep 20, 2019 11:28 pm

Cool on the autostart, very nice.

And mousewheel notch resize is awesome!!!
Gary NC3Z
User avatar
ramdor
Posts: 341
Joined: Wed Jul 03, 2019 3:07 pm

Re: 2.6.8

Postby ramdor » Fri Sep 20, 2019 11:47 pm

w-u-2-o wrote:Since the code can detect the limits, would it be possible for the code to simply limit the maximum rate in the setup screen, thereby preventing the user from shooting himself in the foot? Similarly, the waterfall update time also seems to require the same limits.

Of course this will all go away once the spectral display area is done over with DirectX or similar.

Bug in CTUN mode:


Hi Scott,

Cheers for the feedback. Yes we could have a dynamic fps option, in-fact I did quickly have a play around with the idea when adding the red indicator. I think before adding that I will need to do some more directx testing.

I have been testing using a GDI compliant DX2d surface. I will do some more testing but it looks like this method is not optimal. Direct2d specific line/geometry/textures will hopefully be the way to go, but will need time to convert/implement.

The bug in CTUN mode. Thanks for that, I didn't even know it did that, lol. Hopefully fixed now, but there is an issue as far as I can see, and not sure what the correct solution would be. It is possible to slide to a frequency where a previous filter has now moved outside of the sample area. You can try this easily at 48k sample rate. It doesn't make any sense to me to stop the slide drag, because it is nice to keep going and going. Other than pulling the plug on the stream and then reinstating when back in view, not much else we can do?

Cheers, Richie.
User avatar
K2BU
Posts: 7
Joined: Sun Apr 09, 2017 4:07 pm
Location: Tabernacle, NJ

Re: 2.6.8

Postby K2BU » Sat Sep 21, 2019 12:46 am

I have just noticed a bug when running any version of Thetis with either Commander in DXLab or DDUTils. If I try and turn on the CWX tab while running those utilities, Thetis will freeze and crash. If I only load Thetis and not my logging program or DDUTils, CWX works normally. I have had several other users confirm this situation. If I go back to Protocol 1 and PowerSDRmrx 3.4.9, the CWX tab works just fine. Anyone have any ideas? de Chas, K2BU
User avatar
ramdor
Posts: 341
Joined: Wed Jul 03, 2019 3:07 pm

Re: 2.6.8

Postby ramdor » Sat Sep 21, 2019 1:53 am

K2BU wrote:I have just noticed a bug when running any version of Thetis with either Commander in DXLab or DDUTils. If I try and turn on the CWX tab while running those utilities, Thetis will freeze and crash. If I only load Thetis and not my logging program or DDUTils, CWX works normally. I have had several other users confirm this situation. If I go back to Protocol 1 and PowerSDRmrx 3.4.9, the CWX tab works just fine. Anyone have any ideas? de Chas, K2BU


noted, will look into it next week when get some free time. Cheers.
User avatar
W1AEX
Posts: 273
Joined: Sun Apr 09, 2017 6:17 pm
Location: Connecticut, USA
Contact:

Re: 2.6.8

Postby W1AEX » Sat Sep 21, 2019 3:02 am

The B8 release is running well here Richie and the notch GUI implementation is working perfectly. This feature is a really wonderful addition on top of all the other things that have magically appeared lately! Thanks for all the time, effort, and expertise you have invested in this project. It is very much appreciated!

73,

Rob W1AEX
"One thing I am certain of is that there is too much certainty in the world."
DL8LAQ
Posts: 87
Joined: Sun Apr 09, 2017 3:28 pm

Re: 2.6.8

Postby DL8LAQ » Sat Sep 21, 2019 10:07 am

Why does check / uncheck Semi Break-In change the sidetone level?

Press MOX, let your memory keyer generate cw and then check / uncheck Semi Break-In.

I use Thetis 2.6.8 b8 and found this behaviour while setting up cw keying with UCXLog.
73, Norbert - DL8LAQ - ANAN-7000DLE
User avatar
w-u-2-o
Posts: 2066
Joined: Fri Mar 10, 2017 1:47 pm

Re: 2.6.8

Postby w-u-2-o » Sun Sep 22, 2019 1:40 am

ramdor wrote:
I have been testing using a GDI compliant DX2d surface. I will do some more testing but it looks like this method is not optimal. Direct2d specific line/geometry/textures will hopefully be the way to go, but will need time to convert/implement.
Direct2D sounds like an interesting way to go, rather than going full DirectX.
The bug in CTUN mode. Thanks for that, I didn't even know it did that, lol. Hopefully fixed now, but there is an issue as far as I can see, and not sure what the correct solution would be. It is possible to slide to a frequency where a previous filter has now moved outside of the sample area. You can try this easily at 48k sample rate. It doesn't make any sense to me to stop the slide drag, because it is nice to keep going and going. Other than pulling the plug on the stream and then reinstating when back in view, not much else we can do?
Conceptually, not necessarily code-wise, the solution is simple.

First, consider that the current behavior when tuning the RX/TX passband in CTUN has the spectral display window start to scroll in frequency when the RX/TX passband is tuned outside of the window.

Now just take that paradigm and turn it on its head. If the spectral display window is dragged such that the RX/TX passband would fall outside of the overall IF bandwidth then scroll the RX/TX passband to stay within the edge of the IF passband.

In practice, this will allow the RX/TX passband to sometimes not be on the part of the spectral display that is shown, depending on the pan/zoom state of the display, but it will always be within the current IF bandwidth.

The alternative is, of course, just as you say: as soon as the IF bandwidth no longer covers the RX/TX passband then receive and transmit must be disabled until the RX/TX passband is brought back within the IF bandwidth.

Cheers!

Scott
K9RX
Posts: 289
Joined: Fri Apr 14, 2017 3:47 pm

Re: 2.6.8

Postby K9RX » Sun Sep 22, 2019 12:56 pm

Re CTUN mouse shift of the window ... when the filter window hits the edge of the displayed window, stop its motion! This of course is accomplished by changing its frequency to keep it positioned at that edge of the screen. That way it is always on the screen - just at the edge as long as you are tuning in that one direction. You're never 'lost' as to where it is and there's no issues of turning off TX/RX because its no longer visible.

Gary
K9RX
K9RX
Posts: 289
Joined: Fri Apr 14, 2017 3:47 pm

Re: 2.6.8

Postby K9RX » Sun Sep 22, 2019 1:00 pm

and while we're on the subject of window manipulations: suggestion - correct the water fall so that it is always 'correctly' displaying the signal independent of tuning.

Currently when one tunes the waterfall is smeared until the tuning has stopped ... since you know when tuning how much frequency change you've made, and you know the relationship of that to the amount of pixels traversed - you can use that value to move the waterfall whenever tuning thus keeping it always 'in sync' with the panadapter. Obviously the one side you are tuning away from would have a blank (background) screen for all of the displayed time and would only start displaying the new panadapter data/signal from that moment on ... but since it was off screen anyway (previously, before you tuned) its irrelevant. Yaesu I believe does this in the nex 101.

Gary
K9RX
User avatar
ramdor
Posts: 341
Joined: Wed Jul 03, 2019 3:07 pm

Re: 2.6.8

Postby ramdor » Sun Sep 22, 2019 3:08 pm

So, two couple of possibilities :

1) Move the RX/TX passband when the IF edge is hit, but it will mean that whatever frequency you were on before is now lost. The RX/TX passband will now be on either edge depending which way you are tuning.

2) RX/TX disabled if outside IF, and the RF/TX passband left as tuned and will not move (this is how it is at the moment but rx/tx is not disabled)

Any other possibilities?

Also Gary, with ref to moving the waterfall data. At the moment the whole texture data is shifted down by one pixel, and then new row added at the top. This is very fast, when analysed the waterfall function completes in half the time of the panadaptor. Keeping the waterfall 'in line' will mean pushing each line of that texture by the tune delta, which whilst ok on an ftdx-101d with 480 pixels ( :lol:) vertical if they were using the whole screen is somewhat quicker than someone running at 4k. Would it not result in something like this with a big chunk of black appearing as you tune up in the band in this case?

j.jpg
j.jpg (173.04 KiB) Viewed 2779 times


Cheers Richie.
User avatar
w-u-2-o
Posts: 2066
Joined: Fri Mar 10, 2017 1:47 pm

Re: 2.6.8

Postby w-u-2-o » Sun Sep 22, 2019 3:37 pm

ramdor wrote:So, two couple of possibilities :

1) Move the RX/TX passband when the IF edge is hit, but it will mean that whatever frequency you were on before is now lost. The RX/TX passband will now be on either edge depending which way you are tuning.

2) RX/TX disabled if outside IF, and the RF/TX passband left as tuned and will not move (this is how it is at the moment but rx/tx is not disabled)

Any other possibilities?
Hi Richie,

WARNING: heavy philosophical design thoughts follow! :D

I thought of one more. It represents a more ideal, object oriented version of the radio. In such a radio:

1. Each element of the radio: RX passband, TX passband, and spectral display, would really be completely independent objects.
2. One could have as many spectral displays, RX passbands and TX passbands as one wanted, within the limits of the number of DDCs and DUCs implemented in the FPGA.
3. If an RX or TX passband instance happened to be within the coverage of a spectral display, it would be represented graphically in ways that are similar to how they are represented now.

However, this would tend to be wasteful of DDCs, since people generally (but not always) tend to want to watch a visual representation of the energy in the passband. And there are only so many DDCs to go around. So for now it seems reasonable to require an actual RX passband be within a given spectral display IF bandwidth (but again it might not be visible depending upon the pan/zoom state of the display). And because RX passbands are essentially a second mix in the software, one can have as many as they like within a given spectral display.

Anyhow, this is probably going WAY further afield than anyone wants to. But in an ideal client I've always envisioned a spectral display window per DDC (first mix to first IF), with small windows (windoids?) that represented the UI for each individual RX passband (second mix to second IF) or TX passband instantiated in the spectral display window. Each windoid would have it's own AGC, demod, bandwidth, S-meter, DSP features and other controls and displays. One could drop them onto the display in a manner similar to the way you are engineering for the manual notch filters.

Cheers!

Scott
K9RX
Posts: 289
Joined: Fri Apr 14, 2017 3:47 pm

Re: 2.6.8

Postby K9RX » Sun Sep 22, 2019 10:58 pm

I don't know that it matters that if you tune to the edge - in CTUN - you lose where you are! Its the nature of hte beast. You ARE tuning afterall AND asking to tune outside of where you are... NOT showing where you are, i.e. having it off screen to me is a non-starter as it means I don't know where it is or how far away it is if i happen to get carried away with it.... but again - it is what one intuitively would expect, get to the edge and now the window moves vs. the filter passband moving.

When I partnered with a Planetarium program way back in the 90's for astronomy and I asked to provide a means to have a digital setting circle tie it to the program (no one, at that time, was doing this - it was all new by us) the planetarium guy said "but I can't refresh the screen fast enough" to which I said "let the position of the telescope move around the screen - say if it were tracking or if it were being moved ... and if it gets to the edge of the screen redraw it with the telescope centered at that point in time... then repeat. You could do that here but the downside of that you are still losing your TX/RX passband position and b) you keep the edge near to the edge ... another option.

As for doing a pixel shift - you couldn't just move a pointer to the display array? You have an array the size of the window and a pointer for where new data enters and one for where it exits ... and when moving those shift relative to the array. If this were data in general this would be an easy thing to do - how video is handled, IF it is even handled or accessible this way I don't know.

Gary
K9RX
Posts: 289
Joined: Fri Apr 14, 2017 3:47 pm

Re: 2.6.8

Postby K9RX » Sun Sep 22, 2019 11:02 pm

That stated what you had said - the 'black area' is what I had said would be a result of doing this - although I was thinking it would be a line which was the left (or right) edge of the display - the edge you are moving towards.

If however you are saying timing-wise it is a different animal - one to the other the pan vs. the WF then indeed possibly there isn't a clean solution and the limited window of the Yaesu is what allows them to get away with it.

g.
User avatar
w2ner
Posts: 60
Joined: Wed Sep 11, 2019 12:14 am

Re: 2.6.8

Postby w2ner » Sun Sep 22, 2019 11:02 pm

Guys, lets not make this a blog, lets keep this post clean and to the point so Richie can get information he needs to fix problems. Take anything else off line.
ik7ytt
Posts: 48
Joined: Tue Mar 12, 2019 11:48 am

Re: 2.6.8

Postby ik7ytt » Sun Sep 22, 2019 11:09 pm

Hi Richie

great work. now Thetis become more interesting.

The Ctune implementation will be super.

if possible for me will be nice also re-organize position of band/filter and other button in better way.
there is a lot of un-used space, and when RX2 is activated , become really difficult to understand. i made a screenshot and moved some buttons just for general idea.

Thetis desing idea.jpg
Thetis desing idea.jpg (419.75 KiB) Viewed 2725 times


thanks again
Andy ik7ytt
cLicari
Posts: 82
Joined: Mon Apr 10, 2017 9:33 pm
Location: North Texas

Re: 2.6.8

Postby cLicari » Mon Sep 23, 2019 3:59 am

K2BU wrote:I have just noticed a bug when running any version of Thetis with either Commander in DXLab or DDUTils. If I try and turn on the CWX tab while running those utilities, Thetis will freeze and crash. If I only load Thetis and not my logging program or DDUTils, CWX works normally. I have had several other users confirm this situation. If I go back to Protocol 1 and PowerSDRmrx 3.4.9, the CWX tab works just fine. Anyone have any ideas? de Chas, K2BU


When launching DXLab If in USB or LSB mode, several times the mode switches to CWU and the occasionally the frequency has changed a few hundred cycles.
Carl
NX5T
Carl Licari
NX5T
Anan 8000DLE, FW v2.0 2b, Thetis 2.6.9 c2_1,
Windows 10, dedicated network subnet, Intel i7 8700K, 32GB, NVIDIA GeForce GTX 1060, Samsung 40" 4K display
User avatar
ramdor
Posts: 341
Joined: Wed Jul 03, 2019 3:07 pm

Re: 2.6.8

Postby ramdor » Tue Sep 24, 2019 7:56 pm

Hi all,

Thanks for the reports for various things/issues, all have been added 'to the list'.

..... Direct X conversion

Needing a change of 'scenery' from the rest of the code, I thought I would have a play with the SharpDX (2d) library. The following video shows current progress, running on my brothers rather nice PC (mine is a bit crusty with 8+ year tech).

https://youtu.be/N5bMWviKivs

(sorry not a 4k video, but hitfilm express is unable to export in 4k@60fps)

His system can reach around 38fps@4k with GDI+ with the picDsiplay background turned off before the red indicator fps issue starts showing. Switch it to DirectX and boom, 4k@60fps no problem with 2 Rx both in panafall mode. The responsiveness of the whole thing is in a different league when running in DirectX as can be seen just by the redraw speed of the setup window. You can see from the vid that in GDI+ mode the primary thread is 100% capped, however when switching to DirectX all that effort is offloaded to GPU and additional directx controlled cpu threads.

It is probably a couple of weeks from being put to out test.... coming soon (tm) :D

Cheers,

Richie.
User avatar
w-u-2-o
Posts: 2066
Joined: Fri Mar 10, 2017 1:47 pm

Re: 2.6.8

Postby w-u-2-o » Tue Sep 24, 2019 10:21 pm

Drooling over that video! My god, I can't imagine what it is going to be like living in the 21st century, finally!

Fabulous work, Richie, really outstanding!

73!

Scott
User avatar
W1AEX
Posts: 273
Joined: Sun Apr 09, 2017 6:17 pm
Location: Connecticut, USA
Contact:

Re: 2.6.8

Postby W1AEX » Tue Sep 24, 2019 10:30 pm

Impressive performance boost with DirectX engaged!
"One thing I am certain of is that there is too much certainty in the world."

Return to “Thetis”