Closed kimbapslice closed 3 years ago
Hi. heard the same here.
running last avril neogeo core build
Hi again
on the other hand (i don't want to add another thread with sound issu):
running: MiSTer v200429 / NEOGEO v200510 IO board analog out (direct out) / 128Mo ram
thanks a LOT all of you for amazing work!
Hi, this is the author of JT12, which was forked on this repository to handle the sound. If you can produce VGM files of the relevant music tracks where the problem occurs I will check if the problem is in JT12. Thank you.
Hi there Jotego ;)
here are 3 VGM: Blazing Star - Stage 2 FF3 - Terry song Garou - Happy ending NeoGeo-VMG.zip
Thank you!
and as i can't find VGM from Metal Slug X, there is records i done: Metal Slug X_xylophone.zip 0'0" to 0'15" is original 0'15" to 0'30" is from MiSTer
Thank you!
Hi @MonokromFR,
I have simulated the VGM files. Blazing Star blazing.zip and FF3 FF3.zip sound ok to me but I have just fixed one issue (https://github.com/jotego/jt12/issues/57) that may have been causing problems here too.
But I do hear some noise in GAROU.zip, around secons 8-10 there is a rythmic pattern in channels coming from ADPCM-A module. GAROU_adpcmA.zip. Is that what you were referring to? (For reference: the track is here, number 30)
Hi Jotego
thanks for reply!
there is my feed back about your sample above:
Issues @kimbapslice and I were talking about are not glitches or rythmic pattern, we would say "bad VHS tape with audio warble" or "wierd drunked tone/pitch". or same like a vinyl turntable with bad belt.
Pay close attention to FF3 sample: forget about guitar and drum, listen to the 3th synth note: MV2F seem higher and regular, MiSTer seems lower and drunked. FF3_3th_note.zip
Pay close attention to Blazing Star: on MiSTer record i made a (A) shot at 0"4' that affect music. BlazStar_Get_ready.zip
There is an BlazingStar in game record. listen to MiSTer record:
I think:
I played Blazing Star a lot a few years ago and music is stuck in my head, I noticed this issu the first time i played it on MiSTer.
Thanks to trying to figure it out, i hope my explanation are clear enough, it is hard to explain sound by writing (and obviously my english is crappy).
PS: records sounds are very different, MiSTer is crystal clear while MV2F contains less trebles. the point isn't here :)
Sorry I haven't had much time to test recently either, but the last few play times I did get in the latest versions, the warbly-ness of the audio seemed to have decreased a lot. I will try to sit down and track down more specific instances. Will update.
That's a lot of information, @MonokromFR. Thank you.
I was working with some old files. After replacing them I think FF3 and Garou will sound ok to you. Please let me know what you think about them now: FF3-17 Garou-30.zip
About the pitch movement in the Blazing Star in-game recording: Yes, I can hear that between seconds 11 and 12 there is one instrument whose pitch drifts away. But I cannot debug that. I can only debug using VGM files for reference. Once the sound module goes into the rest of the system and we compile it into the FPGA, debugging becomes extremely hard. Please let us focus only at problems we can reproduce with VGM files.
Hi Jotego. Your last FF3 rip sounds bugged to me, while Garou seems ok now...
I'm not a dev at all and trying to understand what you say ;) I just pointed out what i hear at final analog output: tracks are impacted by others tracks/SFX played at the same time. You know better than me what could be wrong ;)
Hello,
Sorry for the slowness on my part. I got out my AES and a NeoSD Pro and loaded up Garou Mark of the Wolves (I also have Garou in MVS if you need me to try). Using the Mister 6-22-2020 build of Neogeo with all Mister files updated.
In Garou Mark of the Wolves: Mister v200814, OS v200618, Neogeo 6-22-2020 build System Type: AES, Bios: Unibios (4.0), Video Mode NTSC In the Unibios Menu JukeBox Music Player, use sound code 0x072A (unknown music) and 0x734 (Hotaru)
They both sound really bad when the flute sample plays.
Compared this to the real AES and it's day and night. I tried to record it on my phone and the sound quality is not good for sharing. I am certain this bug is in other games.
Hi @kimbapslice,
Thank you for your detailed report and sorry for insisting but it does save me a lot of time if I can be pointed directly to a VGM file of the melody in question.
I think the last sound fix I made has not been yet released by MiSTer curators. But if you're talking about sample music it won't be affeced by it as it was related to PSG.
Thank you
Hello @jotego ,
Sorry I misread your last comment and overlooked the request for VGM files (I didn't know what these were - VGM). Anyway, attached are the 2 VGM in question for Garou Mark of the Wolves and verified glitch still exists in Neogeo build 8-20-2020
In the Unibios Menu JukeBox Music Player, use sound code 0x072A (unknown music) and 0x734 (Hotaru) I downloaded the vgms from vgmrips.net garou vgms.zip
The unknown music is bad immediately. The Hotaru stage music sounds off from the beginning, but really gets bad around 23 seconds mark - something to do with the flute sample.
It's really noticeable during the title music for Pulstar as well, about 30 seconds in after loading the game once the gameplay demo starts and there is action on-screen (shots being fired, ships exploding, etc.). The music speed and pitch really starts dragging as though it's being played back on an analog reel-to-reel player with a dirty head.
It doesn't exhibit this behavior when played back in the game's Options screen in AES mode.
If it does not occur when played via the menu, it is likely to be a bus contention issue unrelated with the YM2610 implementation. Commands or PCM data to the audio "chip" must be coming in late.
Anything new on this? seems unresolved.
@furrtek did a decap of the PCM chip so that might provide some insight
Hello same problem here but I noticed the pitch variation in every game I tested : Art of Fighting, Kof98, Snk vs Capcom...
Here is a comparison between the core and a real neo geo
https://drive.google.com/file/d/1szaYSXKVfdZ9bnq_-mSRIFflALAyzSgA/view?usp=drivesdk
https://drive.google.com/file/d/1sqCVhe7P5jOVbZtC4V_RC36oJ95ZnUSW/view?usp=drivesdk
You can sense after pressing the start button how the second note drifts away in some kind of modulation while the computer shows the tutorial.
There are many parts of the game that has the same problem.
Hard to tell in the video but IRL it's easy to spot, sounds like a tape slow down
Cheers
I am trying out the new 09/08/2021 neogeo core update: NeoGeo:
Seems like the pitch variation is gone!!! Amazing work! So far I have only tested Samurai Shodown 3 and Garou Motw. Please someone confirm and we get to close out another bug. Thanks greyrogue and everyone for making this core better everyday!
I have been testing a few additional games and they all sounded very good! Closing ticket!
In several games, the music sounds a bit off.
For example: In Garou Mark of the Wolves during Hotaru Futaba's giant bell/church rooftop stage, the music sounds like a bad VHS tape with audio warble. The flute instrument sound.
I have noticed this in games like Samurai Shodown 3 as well in different stages/character songs like Rimururu or Nakoruru's. This game has a great soundtrack.
Last Blade: Select Yuki and the first fight will be Amano at the bridge cherry blossom stage, the music is affected.
I will try to list more games and examples, but Garou is definitely noticeable and reproducible. It's hardly game breaking. Thank-you! I am running the newest release 11/25/2019 with 128MB RAM.