Version 1.1.1
- overlay makes better use of local cache for resources

Version 1.1
- the buttons to force input resolution can be mapped to different resolutions (ctrl + click)
- certain settings like the overlay are saved on exit and recalled on launch
- output base resolution spinboxes properly track the resolution when disabled
- alias list is properly cleared on re-load
- in theory better recovery if invalid signals occur, have not seen em to test

Version 1.1.2
- ability to change capture bit depth
- less artifacts when forcing input resolutions
- possibly fixed a rare qtimer-related crash on exit
- implements rgbeasy callback for unrecoverable errors, never seen em but this was the last callback not yet implemented

Benchmarked vcs against datapath's own viewer that ships with the newest (2011) drivers for the visionrgb-pro. Left image is for 16-bit color, right for 24-bit, vertical axis is fps, capturing a 60 hz source. No up/downscaling, 1:1 with source. Vcs is black, datapath viewer is dashed grey. Running in a qemu/kvm virtual machine on a haswell 3.3 ghz machine with win xp as guest and linux as host. Vcs is faster in both color modes, notably so in 16-bit. In 24-bit the difference is small and could be due to other factors. The capture card is likely the limiter in 24-bit, probably also in 16-bit at 1024 x 768.

Just ordered this card and when I get it working, I will report back on my experience with your software. Looks very promising.

Version 1.1.4
- anti-tearing. implements a local triple buffer to reconstruct frame coherency when capture data is fragmented. read the readme for more info. efficacy is demonstrated there with the visionrgb-pro: https://www.youtube.com/watch?v=_gNywFJFRWo, with tomb raider 1 and f-22 lightning 3. needs a decent cpu. has issues/corner cases for you to discover.
- various dialog window sizes are now saved on exit and recalled on launch
- frame drop indicator in the overlay now stays on for longer even if the frame dropping is transient

Version 1.1.7
- custom output filter chains. demonstrated there: https://www.youtube.com/watch?v=XHCoYcRQdQ8. this feature is undocumented, refer to the video for basic usage.
- magnifying glass
- speedup to anti-tear
- loading video/color parameters automatically updates the current signal
- fixed possible crash with anti-tear
- anti-tear has a new setting, step size. keep it at 1 unless you have a good plan.
- button to restore default settings for anti-tear

Known new issues
- latency display is no longer reliable. look at the frame rate instead.

The problem i've been having with this card is when the source changes resolutions. How is this handled in VCS? (i'm downloading now) I'd be okay if it started a new file every change. You can always stitch it together in vegas or something. Or should I just get comfortable with setting it to the max and hoping for the best? Please forgive me, I just got this card a bit ago and havent had a lot of time to finger it out.

Depends on what the problem is. The main thing to know is vcs doesn't record footage, only draws it to screen. You'd set vcs to use or force a particular output size and/or filtering and record the window with a screen recorder.

I have a problem to run the latest version 1.1.7

When launching the program, I get this error message.
https://abload.de/img/vcs-errorj7sz3.jpg

And after that this is written in the console.
https://abload.de/img/vcs-error229s12.jpg

I'm using the VisionRGB E1S on windows 10 PC with the latest driver. Version 1.1.2 works perfectly fine on the same machine.

For reasons of memory usage I lowered the maximum output size from 4096 x 4096 to 4096 x 2160, so the error suggests the e1s supports an input resolution higher than that. I didn't think it would. If you use version 1.1.2 and go to the about tab, what does it give as the maximum input resolution for the card?

It says 2048x4096, so definitely less than 4096x2160. Just a wild guess - perhaps switched width and height?

The code for detecting and printing out those parameters appears correct and reasonably reports 1920 x 1260 for my visionrgb-pro. Whether it's the rgbeasy api having changed for newer cards is hard to say.

Version 1.1.7b
- latency indicator works again
- maximum output size is 4096 x 4096 again
- ability to delete filter complexes (with right-click)
- removed spam in the console when using post-scaling filters

Thank you for the quick patch!
Now VCS works again on my PC. 😀

Strange thing about the resolution still. The e1s isn't a 4k capture card and the datapath docs for it mention 2048 as the maximum width so I'd be inclined to think 2048 x 4096 is a valid reading, to the extent that width and height aren't reversed there. Still, the datapath docs don't always tell the full story and a height of 4096 is a bit odd. Same for the 1 x 2 minimum since for my pro it's 1 x 1.

I've received and installed my Datapath DGC103 PCI capture card. Downloaded the latest drivers from their site and manually installed them on Windows 10 x64. Started OBS to see that the drivers work correctly. And they do. The image that I get is good, but not great. It's way to soft for my likes.

I just installed your program. I'm experiencing some problems, which I hope to sort out, because the image appears to be miles ahead!
- I sometimes have to re-open the program a view times before I can see what's on the screen. What I see reminds me of my old CRT monitor set at a too high refreshrate
- the 1024x768 resolution looks great, the 1280x1024 has problems on the top and bottom of the screen

- I cannot use the mouse and/or keyboard (in OBS they work fine). Nothing seems to happen
- o, and something I feel stupid about..... how do I record video?

Is there anything I can provide?

I've seen certain issues at high resolutions i.e. 1024 x 768 and above but haven't looked into them too much. For instance if the source is 1024 x 768 and the capture card doesn't initialize it correctly on startup, the screen may appear frozen. This has been an issue with datapath's own viewer as well. Simply forcing the correct input resolution should solve it. As for capturing 1280 x 1024 full-size, it's unclear whether you can with the visionrgb-pro. Datapath's own viewer doesn't let you output at that resolution and doesn't say why. It works if you apply hardware downscaling first but vcs doesn't implement this (always asks the card to submit frames full-size).

For recording you'd use a screen recorder over the window.

Thank you for the quick reply. I will try 800x600.

In 800x600 I get a scrambled screen in VCS. Both input and output report the correct resolution. OBS has no problems showing this, but still a soft (allthough lesser then on the higher resolutions) image.

edit: when I boot up the captered PC while still using VCS, it shows the bios and startup screens just fine. When it enters Windows XP it's like this

Hard to say. Possibly you won't be able to use vcs with that combination of hardware and/or software. The only thing that comes to mind is you may not have the correct 32-bit rgbeasy dll if you fudged them in manually.

Version 1.1.8
- basic temporal denoising filter
- some small changes