lr-mame2003 driver improvement and backport

Riverstorm

You mean add hash values and then create a set like a regular ROM set so it adds/removes samples automatically? Samples then would probably look like a ROM set when processing.

I've never tried adding hash values or seen them added to the actual sample set in the DAT to see what it would do like automate it. Usually Twisty's has all the current correct samples for downloading.

maxbeanz

Do we have to enable cheats manually somehow in 2003plus ? If so, could you turn them on by default because you have to enable individual cheats in the menu anyway. Also, why is there a 'Generate XML Dat' in the MAME menu ? Is this a bug that's replaced the Cheat menu ? The Dat file is on Github anyway.

markwkidd

@maxbeanz said in lr-mame2003 driver improvement and backport:

Do we have to enable cheats manually somehow in 2003plus ? If so, could you turn them on by default because you have to enable individual cheats in the menu anyway. Also, why is there a 'Generate XML Dat' in the MAME menu ? Is this a bug that's replaced the Cheat menu ? The Dat file is on Github anyway.

You still need to turn on cheats via the core options for mame2003-plus. Then you'll see them in the MAME Tab menu as with mame2003. Sometime soon-ish I will get around to enabling cheats by default like mame2003 which will skip the core options step.

The XML DAT generator is there for people who are doing up-to-the-minute testing of newly supported mame2003-plus romsets.

Unless you are testing nightly mame2003-plus builds you are fine to use the DAT that is already on github.

maxbeanz

@markwkidd Thanks for the info. Is the Datfile on Github much different from a manually generated one then ? The Github dat says I have 3 games missing (4 roms) (hsf2a, kof2003d, svcchaos) which is what was stated elsewhere in this thread.

Riverstorm

@maxbeanz said in lr-mame2003 driver improvement and backport:

The Github dat says I have 3 games missing (4 roms) (hsf2a, kof2003d, svcchaos) which is what was stated elsewhere in this thread.

I've purposely been avoiding adding those few missing ROMs (4 total) also as they are not MAME official. Once Mark or Grant say they will become official ROMs or they will always be hacks then I'll add them in.

@maxbeanz - I did some research last night but I have no idea why samples don't receive hashes. It's been that way since the beginning I believe. None of the tools even attempt to create hashes for samples. I don't know if they assume samples are easy to obtain and don't change much. Maybe someone with better MAME institutional knowledge could answer it.

dankcushions

i guess samples are not static files like roms. they could be re-recorded at better fidelity, etc. the names of the files should remain the same, though.

maxbeanz

@riverstorm Or if you know that you have a perfect samples set for 2003plus you could do a DIR2DAT using CLRMamePro which scans a selected directory and adds CRC's/SHA's (if selected) and spits out the DAT file.

Also, did somebody make a Datfile of just the new and updated roms for 2003plus that are needed so I can add them to my filtered 2003 set ?

markwkidd

My understanding about samples is that mamedev considers them like bezel images or any other assets which can't be validated digitally against the original like can be done with ROMs. The logic being that samples are therefore supported but no one set is 'endorsed' by adding it to the source in the driver.

That being said, I personally think it sounds like a good idea to consider adding recommended sample crcs to mame2003-plus if there is a consensus about what sample sets are good.

Riverstorm

@maxbeanz said in lr-mame2003 driver improvement and backport:

Also, did somebody make a Datfile of just the new and updated roms for 2003plus that are needed so I can add them to my filtered 2003 set ?

Here's the samples from Twisty's I used with mame2003-plus (processed with the DAT so all of them are named and packed in the proper zip). I am using enhanced unofficial samples where possible.

It has complete CRC and SHA1 information. I imported the ROM header information from the main DAT so a missing ROM from a set will show as "Astro Blaster (version 3)" vs. "astrob".

The year/manufacturer is based on the ROM set and not who actually made the sample or what year. Process it like a ROM set (Set and ROMs checked) and not samples. I would build it in a new folder and cross link to your original set until you feel comfortable it's correct before merging.

I am ok with endorsing Twisty's samples but I do use the unofficial enhanced versions of some samples like (dkong, berzerk, invaders, qbert, rallyx, starcas, etc. ) he has listed on the site. I am ok with either though.

Also, did somebody make a Datfile of just the new and updated roms for 2003plus that are needed so I can add them to my filtered 2003 set ?

I've been thinking about this and I don't think it would be to hard if done in non-merged mode.

My thought is to make a subset of mame2003-plus. This set would be around 1.5GB or so. You would be able to dump it in the ROM directory and play any new games or changed games from mame2003-plus only.

Then if you want dump this subset into your main lr-mame2003 folder it would match exactly the mame2003-plus full set.

One minor issue is lr-mame2003 has these two sets which don't exist in mame2003-plus. So it wouldn't hurt anything except they would be removed if processed with the mame2003-plus DAT. That way instead of like 28GB for the two sets it would be like 16GB or so.

Only exist in lr-mame2003:

Cabal (US set 2) (cabal2)
Urashima Mahjong (urashima)

I would rather wait until those 4 non official ROMs are decided on what to do. It would require processing the set a few times to do this but I don't think it would be to hard.

@markwkidd - Do you know if you can add hash information to the sound sample sections of a DAT?

<sample name="run01.wav"/>
<sample name="run02.wav"/>
<sample name="run03.wav"/>
<sample name="jump.wav"/>
<sample name="dkstomp.wav"/>

maxbeanz

@riverstorm I ran the samples dat you posted and noticed that you are using 'Gorf (0.125u2 & Newer)' from Twisty's. Is this compatible with MAME 0.78 ? I looked inside the zip and it seems to contain 11 samples at 0 Bytes ?

Riverstorm

@maxbeanz said in lr-mame2003 driver improvement and backport:

@riverstorm I ran the samples dat you posted and noticed that you are using 'Gorf (0.125u2 & Newer)' from Twisty's. Is this compatible with MAME 0.78 ? I looked inside the zip and it seems to contain 11 samples at 0 Bytes ?

The hazards of no hash but I understand how it allows you to use samples that sound like you remember. Memory is funny thing.

Yeah I have the newer samples that have some merging. I will replace the new Gorf samples with the older Pre 125u2. If you see any other sets you prefer just let me know. I am using all the enhanced versions from Twisty's (Retrocade and Gyrovision versions) but if you like the official ones or the older ones (like Wizard of Wor has current and older) just let me know.

maxbeanz

@riverstorm Not sure I understood you. I just wondered if the Gorf samples you are using (Gorf (0.125u2 & Newer))' are actually compatible with mame2003plus ? If you look inside the zip, 11 of the files are 0 bytes. Is it corrupted ?

dankcushions

@maxbeanz yes. no mame version is going to be able to do anything with a sample file of 0 bytes :)

i just downloaded the same file as you and all the files had sizes. try redownloading.

maxbeanz

@dankcushions There are 98 files in the 'Gorf (0.125u2 & Newer)' from Twisty's. 11 are 0 bytes.

Riverstorm

@dankcushions said in lr-mame2003 driver improvement and backport:

@maxbeanz yes. no mame version is going to be able to do anything with a sample file of 0 bytes :)

i just downloaded the same file as you and all the files had sizes. try redownloading.

I think he was talking about "Gorf (0.125u2 & Newer)" has some zero byte ROMs and "Gorf (0.125u1 & Older)" has none. When I built the DAT it was with the newer set and should have been the older set to work with 0.78.

I am not sure but I think it is intentional (no longer needed) maybe the newer MAME versions are using dumped working ROMs so they no longer need those samples. Either way the older set is for 0.78, it has the same samples and they aren't zero byte.

maxbeanz

@riverstorm The older Gorf samples are OK apart from the 'emperor' sample. It's 1kb and doesn't play anything. Not sure if it's supposed to be like that ?

dankcushions

sorry, yes, i rechecked and i have the same 8 files at 0 bytes. perhaps they haven't been recorded, or aren't used by the game, i don't know. there are no checksums or 'official' mame samples. we just work with what people have recorded and circulated.

i've began to lose the thread of this conversation - can't you just install the samples and see if they work in mame2003plus? all it will care about are the file names.

Riverstorm

@maxbeanz said in lr-mame2003 driver improvement and backport:

@riverstorm The older Gorf samples are OK apart from the 'emperor' sample. It's 1kb and doesn't play anything. Not sure if it's supposed to be like that ?

There's only two choices either the newer 0 byte or the older 188 byte sample. If you have another sample we can drop that one in place?

markwkidd

@mediamogul these folks are trying to get some Gorf samples sorted out. Do you know anything about this emperor sample?

Riverstorm

@dankcushions said in lr-mame2003 driver improvement and backport:

i rechecked and i have the same 8 files at 0 bytes.

Hey Dank, a double recheck, it's actually 11 just to verify we are looking at the same sample! ;)