MSX Emulator Comparison: openMSX review update

by snout on 09-03-2005, 12:45
Topic: MRC
Languages:

The openMSX review in the MSX Emulator Comparison has just been updated, putting the recent release of openMSX 0.5.1 to the test. Although the MRC EmuRank has increased from 79.07% to 85.51%, openMSX maintains its 2nd position at the MRC EmuRank Charts.

Relevant link: MSX Emulator Comparison - openMSX review

Comments (18)

By Manuel

Ascended (15602)

Manuel's picture

09-03-2005, 16:37

Thanks for the update!

I was just wondering: are you sure those turboR VDP and CPU benchmarks are correct? AFAIK we haven't changed anything that could influence it, and still it changed quite dramatically.

Gonna test that Uknown Reality stuff again, last time it worked fine on my PC.... :/
What's going wrong with them exactly?

I also didn't hear the hickups you mentioned in SD-Snatcher... Are they really reproducible? (Other CPU activity may trigger hickups...)

The Almost Real glitches are really odd. Some people can reproduce them, some can't. I can't for example, it just runs perfectly on my PC :\

By Manuel

Ascended (15602)

Manuel's picture

09-03-2005, 16:41

Oh, by the way: maybe it's time to modify your conclusion a bit... The distance between the 'good' and the 'best' is only 14% (between paraMSX and MSXPLAYer), while the difference between NLMSX and openMSX is now 21%... So, the way you divided them doesn't make too much sense anymore.

By snout

Ascended (15187)

snout's picture

09-03-2005, 17:19

Could it be that the change in turboR CPU and VDP benchmark results are caused by me using the SDLHi renderer, as requested, this time?

The SCC test can be reproduced, I was running no background tasks whatsoever (except for the bleeding obvious virusscanner ^_^) at the time. I'll send you two mp3s to demonstrate things, it seems the bass-line is interfering with the flute, somehwo...

As for the 256kB samples in Unknown Reality.. just give the demo a go and you'll find out soon enough Wink

If I'm not mistaken a new release of blueMSX is coming very soon, I'll wait a bit with rewriting my conclusion until that one's out. I have two new tests and some extra U&F ratings I'd like to check as well... so... be patient. In general, the good - best - ugly division still works, als NLMSX, MSXPLAYer, openMSX and blueMSX are still the 4 emulators that are capable of running the vast majority of MSX software. The fact that openMSX just got closer to blueMSX doesn't change that fact.

By tcm1998

Expert (120)

tcm1998's picture

09-03-2005, 17:22

And ofcourse GFX 9000 IS there now and so is the debugger interface. There's just no GUI for it yet (for windows). Fudeba for linux (unfortunately it uses a lib not available in windows) uses a similar technique as catapult to control openMSX and debug programs. All sorts of debug commands can however be used through the console Wink

By Manuel

Ascended (15602)

Manuel's picture

09-03-2005, 17:24

I don't think it can be caused by the renderer. The renderer or other performance issues have no impact on the internal timing, AFAIK.

For the other stuff: we'll go to the bottom to find out the cause Wink

By tcm1998

Expert (120)

tcm1998's picture

09-03-2005, 17:24

Hmm, you posted at the same time as I, so I missed that

But UR issue might be a windows only problem Sad

By Manuel

Ascended (15602)

Manuel's picture

09-03-2005, 18:58

I just checked: the UR samples work just great in openMSX on my system Sad

By Manuel

Ascended (15602)

Manuel's picture

09-03-2005, 20:06

snout: are you sure you used an expanded MSX AUdio extension? Like the one in the audio or mbstereo extension or the one in Boosted MSX2?

We can't reproduce your problem at all, not even in Windows as tcm1998 just tried. Can you also send us an mp3 for that one?

By snout

Ascended (15187)

snout's picture

09-03-2005, 20:22

Bugreport: The Philips NMS 1205 MSX-Audio contains 32kB sampleram, while UR reports it has 256kB. (And afaik it plays different parts of the samplekit than a real NMS1205, but that could just be my memory fading). Either way.. congratz, review updated.. EmuRank a tad higher... ^_^

By Manuel

Ascended (15602)

Manuel's picture

09-03-2005, 20:32

Does UR report the amount of sampleRAM? I think not. Before I had my owh Music Module expanded, the demo just worked and didn't report anything about sampleRAM.

Howver, the part you can hear is wrong indeed. We'll investigate it.

By Manuel

Ascended (15602)

Manuel's picture

09-03-2005, 20:33

Oh, by the way, where exactly comes the 70% sound quality come from? Except for the SCC issue you mentioned, is there anything else wrong?

By snout

Ascended (15187)

snout's picture

09-03-2005, 20:51

UR reports the amount of sample RAM it detects in the 'fake boot screen'.

As for the sound quality: Especially the FM drums are really sloppy compared to the real thing. As the max amount of points that can be gained by music chips is 30, every chip getting a 1-5 rating (one extra rating for MBSTEREO, soon 2 separate ones for OPL4 wave and FM), losing one single point decreases the score quite drastically. As none of the soundchips sound 100% exactly like the real thing... you get my drift.

By Vampier

Prophet (2295)

Vampier's picture

09-03-2005, 21:31

By Vampier on March 09 2005, 12:30
no word on Gfx9000, although in a early stage it should have been included in the tests... and imho it should also be included in all future tests on all emu's.

there can be only 1! (or 2 in this case )

The flickering/flashing stuff is indeed a litle bit of a problem...

By Manuel

Ascended (15602)

Manuel's picture

09-03-2005, 21:49

snout: you're right! I completely forgot about that! It's probably related, though. We'll check that out!

About the sound quality: I don't think any emulator sounds exactly like the real thing. A real MSX has a lot of extra sound circuitry which filters the sound. This depends a lot on the specific model also. As for openMSX: we use the same engines as most emulators, so it shouldn't be much worse then the others (e.g.: YM2413 by Jarek Burcynski).

By Edwin

Paragon (1182)

Edwin's picture

09-03-2005, 22:26

You're aiming for perfection! Make good on the aims. Wink

By Grauw

Ascended (8311)

Grauw's picture

09-03-2005, 23:33

The SCC test can be reproduced, I was running no background tasks whatsoever (except for the bleeding obvious virusscanner ^_^) at the time. I'll send you two mp3s to demonstrate things, it seems the bass-line is interfering with the flute, somehwo...
Uh, that sounds awfully much like the difference in playing an SCC+ game on an SCC to me... You sure it isn’t just that? SCC+ is a different cartridge...

~Grauw

By mth

Champion (482)

mth's picture

10-03-2005, 01:15

snout: Which version of SD Snatcher are you using? Some cracks have extra overhead which causes animation and sound to be out of sync, for example the well known StarCracks version. Also make sure your version really uses SCC+ sound, instead of an SCC+ cartridge in SCC mode.

I hope it's one of those things, otherwise we have some heavy bug hunting ahead Wink

By Vampier

Prophet (2295)

Vampier's picture

11-03-2005, 07:42

Don't mind me when I slowly will go totally insane with bug hunting Wink