toot_toot wrote: ↑Fri Apr 17, 2020 7:52 pm
But if the Next is
more compatible than those extended variants, why is it I can play the games mentioned in this thread as not working correctly on the Next perfectly fine on my Spectrum +2b with DivMMC, which is notorious for compatibility issues?
Please make sure these things are reported to the team directly as they may not be seen if they are only reported in various forums. The problem is going to be in how the snapshot is loaded rather than any compatibility issue and I will make sure Garry knows about it.
Like it’s been said in this thread, the disappointing thing is that the response seems to be very defensive/argumentative instead of admitting there’s an issue and trying to resolve it. All I wanted was a Spectrum that supported HDMI, played games from SD card and was compatible with the Spectrum’s library. All of those things were promised in the Kickstarter and it would be nice knowing that if a list of incompatible games were published - even if it was over HDMI - the response would be “we’ll fix it” instead of trying to put the blame on the people who actually funded making the Next.
It goes both ways. For example, this HDMI issue has been widely discussed for a year and I have personally made sure people were aware of it whenever they asked about HDMI in discussions I have seen, which must have been dozens of times. It's also been widely discussed that the Next hardware is not finished (the hope is to finish for the second kickstarter) and that this HDMI issue is on the list to be tackled. So when you say that no one's admitting a problem, it's a true wtf moment for me and others
As far as what the KS promised, back then the pi was going to deliver hdmi video. The pi would be operating in the same way as the zxhd does for the spectrum. That's no longer possible as you'd need a full on emulator to generate Next video and a pi zero is not up to it. However, the Next can do something else and that is deliver its own digital RGB signal to the pi for processing. Things have moved far since the KS and using the pi as an hdmi solution would hurt the usefulness of the pi as a co-processor so it's something we don't really want to do.
Pegaz wrote: ↑Fri Apr 17, 2020 11:48 am
I think the same way, but instead of getting an answer to that question, we get tons of technical essays on why hdmi is a bad boy.
Someone had to do a reality check first and not promise the HDMI feature at all, if it didn't work properly.
Or maybe we just expect too much from people, who haven't been able to fix the Next keyboard for three years.
Fortunately, we still have the right to be disappointed, until further notice.
You got a technical answer so that it can help you understand the problem and explain why VGA / RGB works and HDMI does not.
I'm going to take this opportunity to undo a lot of falsehoods that people like to repeat:
1. The ZX Next project is not related to the UNO project which can be seen by anyone who looks at the fpga source code. When Victor transitioned from an Altera part to a Xilinx part (due to cost advantage), Antonio was a great help in getting him started. The pal timing module (that generates spectrum video timing) was taken from the UNO, as is allowed by the open source license, and due to bugs in this module, some display glitches that the UNO had were also evident in early videos of the Next in operation. From this, the UNO developers concluded everything was copied and withdrew their source code from being open. They had never seen the Next's source code and had no evidence that anything was copied.
2. The Radastan mode on the UNO was incorporated as lo-res in the Next. There is a credit in the source code that acknowledges this. Nothing was copied -- it is another implementation. It was deemed a useful feature and like many other useful features (divmmc, multiface, soundrive, turbosound, etc) it became part of the Next's design. It was not deliberately made incompatible. The lo-res mode was under-specced for the Next which has an 8-bit palette so it was extended to 8-bits per pixel initially. Since then lo-res in 4-bits, like Radastan, was added because it makes sense in the Next design. The scrolling that the UNO added to Radastan does not make sense in the Next design so the Next implementation scrolls that layer like every other of its layers. The Next goes much further in its display hardware and some of the things that were done with lo-res on the UNO just don't make sense in the Next context.
3. ULA+ was initially in the Next. However, it too is under-specced as the Next has 256-entry palettes and ULA+ can only do 64 entries. So another method was sought which turned into ULANext. Due to particular political nastiness concerning ULA+, it was removed for a while. It returned recently when public permission was given to put it in.
This dig:
Or maybe we just expect too much from people, who haven't been able to fix the Next keyboard for three years.
When you run a KS there is a fixed amount of funds available. If you run out of money, the project is dead: see Vega+. It was not known if it would be affordable to use the solution that eventually made it into the Next until the end when production was about to start.
The keyboard was envisioned as a 48k replacement keyboard. As such it was designed for the 8x5 membrane. Board compatibility with the 48k was also seen as a backup in case something went wrong with the case. The double layer membrane was found to be unacceptable as expectations for keyboards are higher in the modern day than back in 1982. So a single layer was attempted with the keyboard manufacturer confident that they could produce something that delivered keypress for two keys (a single compound key) within a reasonable time difference. That turned out not to be the case in the time given them. The keyboard manufacture was handled by a reputable manufacturer with many products in the marketplace. They have much more experience than Pegaz or any other back-seat designers in making keyboards.
The three year wait was not due solely to the keyboard. There were many other contributors, principally the loss of the first mold manufacturer who just decided one day they weren't going to do it anymore.