Unseen wrote:
I would guess 25MHz (on the wire) because that is the minimum TMDS clock for DVI and HDMI. Even if the scaler can go lower, the receiver would need to support it too.
I don't know the minimums, but that sounds reasonable.
The 301 HD does have a "DVI" mode that I tried hoping to get better compatibility. Some other video processors and displays are a bit more forgiving with signals marked as DVI.
I don't have general display compatibility concerns when I test downscaling. My DAC doesn't mind the standards and gladly converts just about anything to RGB. I also have a projector hooked up that is very accommodating and accepts just about anything that I can input.
Unseen wrote:
The correct way to fix it would be to use pixel replication and flagging it as such in the HDMI metadata. The relevant CEA standard specifies pixel doubling for 720x240p/480i/288p/576i for an effective horizontal resolution of 1440 pixels, but it also defines a few special modes with 2880 horizontal pixels and up to 10x pixel replication to support pixel-exact output of content from low-resolution game systems.
That sort of makes sense if I planned to feed a standard digital display, but I can't think of a use case for downscaling doubled vertical lines while doubling color clocks on a modern display. Additionally, once I start feeding 1440 color clocks with only 240 lines, I've already abandoned the standards and picky displays wouldn't accept it, anyway.
I find that the best timings for the horizontal color clocks are system specific, but I chose 1920 as a convenient starting point. Standards don't really matter much here, because the goal would be to feed a CRT through a DAC (that gladly accepts off spec signals). Naturally, a CRT has no concept of horizontal resolution, so I only have to remain within the mechanical abilities of the display. (All academic, of course, because I have no use for super resolution and it's much easier--and more practical--to use a Corio2 for custom timings.)