@gf: So are you able to provide any actual useful feedback on the performance of the 6xx4BC/BD? That's what the OP is interested in. The only threads that I've managed to turn up have been negative in terms of triggering and display stability. Your actual usage experience would be most helpful.
P.S. I agree that to be achieving triggering at the claimed sample rates then it must be doing something in the FPGA, even if it isn't working very well.
Well, it basically does capture and display waveforms

.
I do not know what functionality beyond this is needed by @shobo.
The specs are public anyway, so I see no need to repeat them.
But I can list a couple of points which bother
me most.
I don't know the VDS1022I, so I can't compare.
* There are significant offset errors, even after running the calibration procedure. Hantek recently did admit a software bug in the calibration procedure and promised a correction - let's wait and see...
* Edge trigger basically works OK (for waveforms which permit edge-triggering) as long as the noise level is sufficiently low. With increasing noise level the edge trigger becomes indeed unstable. IMO the edge trigger should use a larger hysteresis by default, or offer an option in the GUI to adjust it. So when the noise level is too large then it is necessary to switch to pulse trigger (which provides some kind off holdoff functionality when using the ">" operator). Adjusting the parameters needs some fiddling, though , but it enables triggering of signals which cannot be triggered with the simple edge trigger any more. I also managed to trigger on glitches after some fiddling - but that's not "plug and play". There are also other trigger options (video, SPI, CAN,...) but I never needed them. The availability of these advanced trigger options suggests that the FPGA does even implement a configurable, stateful trigger engine, so trigger performance is possibly just a matter how this engine gets parametrized by the software.
* The white noise floor is about 350-400 uV RMS. BW limit does not reduce it. This is IMO caused by operating the HMCAD1511 at 50x digital gain, which enables a simpler (and therefore cheaper) anlog frontend design. There is no hi-res acquisition mode, so averaging is the only option to improve the noise floor.
* At 200ns/div, the captured waveform is sometimes truncated. I did not find out yet, which border conditions need to apply in order that the problem occurs.
* FFT is calculated from only 1024 points, although up to 64k can be captured by a single acquisition
* At <= 200ns/div FFT is calcaulated from upsampled (interpolated) data of up to 12.5GS. This is IMO useless, as the interpolation does not augment the signal with additional information, but in fact it reduces the number of "true" samples included in the FFT. If the data are sampled with 1GSPS with Nyquist at 500 MHz, what sense does it make to display the spectrum up to 6.25Ghz?
* The same applies to captured data when they aresaved to a file. When they were captured at <= 200ns/div, then the upsampled (interpolated) data are saved, and not the 1GSPS raw data.
* There are a couple of bugs in the software/GUI logic. Example: The interpolation mode can only be changed when the timebase is set to <= 20ns/div, although it also affects the waveform display at larger timebases. Some setting changes are not sent to the device before other settings are changed as well. Most issues can be worked around if you are aware of them, but they may make you wonder if you don't know them yet.
Yes, there are a couple of thing I'd like to see improved, but I also don't want to make the device bad. Given the price I paid for my 6074BD (i.e. the model with AWG), it would likely not be fair to expect much more than I got (for a Rigol DS1054Z, I would have needed to pay three times as much at that time).
EDIT:
I just did take a brief look at the Owon web page and my understanding is that Owon's 4CH/1GSPS model would be rather the VDS2064, and not the VDS1022I.
EDIT: Just noticed that I did mis-read the spec - the VDS2064 does not have 1GSPS either, but only 500MSPS (and for 1GSPS the VDS3104 were required).