GoldenEye 007 Nintendo 64 Community, GoldenEye X, Nintendo 64 Games Discussion

[SubDrag]

Is there any noticeable difference? Or just might avoid an out of memory crash.

[zoinkity]

Since there was a bit of a question about VI regs, here's a link to a document covering the more obscure ones. It tries, and fails, to leave out the technical end of signal generation.
https://pastebin.com/pJG5SBnW

Y & X scale are two halfwords: the upper is a subpixel offset, the lower a scale factor.
The scale factors are a 2.10 value used to convert the image you're trying to display to the actual output size (640x480 for NTSC). Note they apply to the generated field, not the original image (that gets confusing with y scale because interlacing effectively halves that).
The values are inverted to create the scale factors. 0x400, converted to 2.10 format, is 1.0. 1.2 is the common factor for lazy NTSC -> PAL output. Deflickered mode uses 0.5 (0x800) to blend two scanlines together, then uses subpixel offsets to overlay the two interlaced fields in a 1:3 ratio.

Guessing what you're calling vstart is the Vertical Video registers. Again these are two halfwords, the number of halflines to start drawing active video and the duration counting from vsync. Under normal circumstances, the difference should match the expected actual vertical resolution (474 in NTSC).
However...if you want to letterbox, your video isn't 4:3, or you want to shift video further up or down the screen for whatever reason, the start and end can be adjusted.
The example given was gameplay in Donald Duck, running (roughly) 400x300. Vertical Video was reduced to 002D 01F3, shaving 8 from the top and 12 from the bottom. What's actually displayed is 454 instead of 474. Everything else is black--whatever the console believes is the black reference IRE.

...Which kinda also makes the point that plenty of other games used weird video modes, Rare certainly didn't have a monopoly on that front.

[Jonaeru]

Awesome! I'd love to see the Japanese version too, but it seems like it might be difficult.
I'm planning to get an Analogue 3D in January next year...
_________________
Perfect Dark N64 Lab (JP): http://pdlabo.knowhow.jp/
YouTube PD: https://youtube.com/@jonaeru-pdlab
YouTube GE: https://youtube.com/@jonaeru-gelab
X: https://x.com/jonaeru

[zoinkity]

Japanese? Shouldn't be too hard to port from USA. They aren't very different, would just be address adjustments and maybe 1 or 2 changes to menus.

Patching PAL directly would mean completely redoing everything from scratch. That version was refactored, its directory finally cleaned of unused old code.

[edit]
Looked at your username and realized you're talking about Perfect Dark, not GoldenEye. Sorry about that!

[Jonaeru]

Thanks for reply!
Even if it's GoldenEye 007, I'm still happy!
_________________
Perfect Dark N64 Lab (JP): http://pdlabo.knowhow.jp/
YouTube PD: https://youtube.com/@jonaeru-pdlab
YouTube GE: https://youtube.com/@jonaeru-gelab
X: https://x.com/jonaeru

[bmw]

Just wanted to report that this 480i patch works great on my new Mister setup (Multisystem2 console). Even works in the turbo core, boosting framerates around 15 to 20 percent.

HOLY CRAP does this look better at this resolution! And even better at 480p which the core supports for 480i content. I never knew the menu text in this game could be so crisp.

Once I get a proper controller hooked up and configured I will do some video capture.

[bmw]

Screenshots incoming....

This is straight from the analogue output of my Mister, at 480p.

[SubDrag]

How is the performance compared to the Analogue 3d video I posted? I was under the impression Mister FPGA is a little underpowered.

[bmw]

I'm hoping to find that out later tonight.

Does the Analogue 3D have a framerate counter overlay option? If so it would be easy to compare as the Mister N64 core does. If not, I will count the total frames that displayed in your video during the Dam animation the moving camera intro) and compare that to my Mister stock N64 core and the turbo core.

[bmw]

Here are my testing results. I used the 2.7 seconds it takes in the Dam intro to pan across the guard tower, starting the first frame where the right-most part of the roof exits to the right, and ending when the last bit of the tower disappears. This gave me a consistent test point.

I hand-counted the number of visible frame changes by advancing the video one frame at at ime.

43 frames: Your clip (480i)
41 frames: Stock N64 Mister core (480i)
44 frames: Turbo N64 Mister core (480i)
44 frames: Turbo N64 Mister core (480p)
54 frames: Turbo N64 Mister core (240p original rom)
45 frames: Stock N64 Mister core (240p original rom)

This amounts to 15 to 20 FPS depending on the test (the best framerate being on the stock rom at 240p on the Mister turbo core). The 480i test results were basically identical between the Mister and the Analogue 3d, when using the turbo core on the Mister. Slightly lower framerates when using the stock core.

I would note, however, that the interlaced version has at least the ILLUSION of smoother motion because full frame update takes 2 frames to actually happen - the odd update and then the even update. And while this causes interlacing artifacts (I call it double-vision) in up-close objects, the distant ones (such as the mountains in dam) appear to move each time. Kind of a quasi-interpolation effect.

Lastly, as much as I like the 480p, it is far from perfect as the Mister adds a lot of image errors, usually in the form of a few pixels being out of place every few frames. Nothing significant but noticeable.