Donkey Kong on MAME - sound quality not accurate

[apocalypto]

Hi guys, I wanted to get some opinions on MAME and Donkey Kong.

In current versions of MAME the sounds are distinctly muffled. It does not sound like an authentic DK arcade machine. The older versions of MAME that used the add-on wav file samples did sound more authentic, but that's very old MAME technology and not a good solution to downgrade to those.

MAME devs swear it is accurate audio emulation to the arcade machine. I find this hard to believe. Every recording I have heard of a real cabinet has had a loud, crisp sound, not muffled at all.

Please listen to these 2 recordings for an example:

This is a recording of DK Arcade Game:
http://www.mikesarcade.com/store/prod/ddk/sounds/dkwalkjump.wav

This is a recording of DK on MAME .141 I just captured direct from my pc.
http://fuelhousedemo.com/dkong-mame.wav

Are any of you running MAME (without samples) and it is sounding like the original arcade game? Or is yours muffled also? What version of MAME are you running?

Thanks!

[ahofle]

If you think that's bad, try firing up Mario Bros.

[DeLuSioNal29]

The 1st link is broken.  Here's the correct link:  http://www.mikesarcade.com/store/prod/ddk/sounds/dkwalkjump.wav

[SavannahLion]

I'm just speculating here.

Does the game sport an amplifier and it's simply not emulated here? I checked the manual, but I'm too tired to read the schematics right now. It looks like there is one.

I know MAMEDev sometimes ignore some hardware configurations on the premise of accurate emulation. I recall one article that described a change in a ROM that accurately rendered the video out on a game to include blue when, in fact, the only known existing cab of that particular game has had the blue feed disabled at the monitor (rather than at the PCB). An accurate emulation would take this fact into account and disable the blue channel accordingly. MAME did disable the blue channel in earlier versions but this was "corrected" in later revisions.

Could that be happening here? An external amplifer that's simply not accounted for?

I dunno, it's a thought.

[apocalypto]

I'm just speculating here.

Does the game sport an amplifier and it's simply not emulated here? I checked the manual, but I'm too tired to read the schematics right now. It looks like there is one.

I know MAMEDev sometimes ignore some hardware configurations on the premise of accurate emulation. I recall one article that described a change in a ROM that accurately rendered the video out on a game to include blue when, in fact, the only known existing cab of that particular game has had the blue feed disabled at the monitor (rather than at the PCB). An accurate emulation would take this fact into account and disable the blue channel accordingly. MAME did disable the blue channel in earlier versions but this was "corrected" in later revisions.

Could that be happening here? An external amplifer that's simply not accounted for?

I dunno, it's a thought.

That's an interesting thought. I wonder if there's a way to software amp that rom somehow. I'm not very experienced with MAME, could it be as simple as a patch file/DIFF file that for that rom??

Any MAME devs use this forum?

[VanillaGorilla]

Honestly, and I don't know how old you are Apocalypto, but the mame audio that is currently in place sounds pretty much like I remember it from back in the day. The one from mikes arcade sounds EQ'd, or of a poor recording quality, and rather shrill to me. I agree that the old DK SAMPLES were much 'brighter', but I never thought they were more 'accurate'. I always remember Donkey Kong sounding like it was inside a shoe box or something, very bass-ey, and very muffled. That machine really rumbled....

[ark_ader]

Honestly, and I don't know how old you are Apocalypto, but the mame audio that is currently in place sounds pretty much like I remember it from back in the day. The one from mikes arcade sounds EQ'd, or of a poor recording quality, and rather shrill to me. I agree that the old DK SAMPLES were much 'brighter', but I never thought they were more 'accurate'. I always remember Donkey Kong sounding like it was inside a shoe box or something, very bass-ey, and very muffled. That machine really rumbled....

Yeah the Mamedevs say its 300% accurate.  Better than sliced bread.

I like the older versions better with samples, and I do not need a cray to play it.

I'm sure Haze will pop up and say its wonderful and perfect, etc.

But I have played it enough times in the arcade, to tell the difference.   

[VanillaGorilla]

I like the older versions better with samples, and I do not need a cray to play it.

I agree totally. I wasn't trying to say the way it is now is 'better', but it is how I remember it standing in an arcade. But I'm sure a bunch of people on this forum have original DK's that could confirm or deny this fact. I liked the version back in 066 just fine (I think that was like 1996). It would be sweet if someone would just add an EQ to the settings menu in MAME.

[apocalypto]

Honestly, and I don't know how old you are Apocalypto, but the mame audio that is currently in place sounds pretty much like I remember it from back in the day. The one from mikes arcade sounds EQ'd, or of a poor recording quality, and rather shrill to me. I agree that the old DK SAMPLES were much 'brighter', but I never thought they were more 'accurate'. I always remember Donkey Kong sounding like it was inside a shoe box or something, very bass-ey, and very muffled. That machine really rumbled....

I was born in 81 so to be fair I never played Donkey Kong 'back in the day' and as such do not remember it sounding one way or the other. However I have listened to recordings from various sources of original DK machines and they all sounded a lot crisper and brighter than the current mame emulation.

Take this one for example, which is maybe one of the best as it is now part of the National Videogame Archive.

http://www.coinopvideogames.com/sounds.php

Donkey Kong is on page 1 on the left side menu.

To me the walk sound seems to be higher pitched than the current mame sound? It's definately clearer.

[BadMouth]

Don't own DK, but one of the things that always in stuck in my memory from childhood was loud bass(for the time) and muffled highs.
The first sample sounds too bright to me. 

Is it possible that the machine in the first sample has had the speakers replaced with ones with a greater range?
The original didn't have a tweeter or even a whizzer cone, so the highs couldn't have been too bright.

Another thing to consider is the microphone used. 
I'd guess the amateur recordings from 1982 used a microphone with a response curve meant for vocal recordings and a cassette recorder.
It wouldn't pick up much under 100hz, so don't hear any of the bass.

Of course, I'm making judgements based on childhood memories (don't know the last time I saw DK in the wild).
I'd be interested to hear what someone with an original machine has to say.

The question for the Devs would be - What did they use for reference, the signal from the board or the sound from the speaker?

[SavannahLion]

The question for the Devs would be - What did they use for reference, the signal from the board or the sound from the speaker?

In all likelihood, the signal from the board. Can't see much incentive for any MAMEDev to emulate the speaker itself.

[boardjunkie]

+1 on the suggestion to add some sort of audio EQ-ing....even if its just simple passive (cut only) bass and treble controls. Newer games don't generally need attention here, but the older games used cheap speakers with limited bandwidth, 'specially in the high end. This actually worked to an advantage....it rolled off high end not needed by the game sounds and the nasty audio artifacts that came from the rudimentary D/A converters. For an example, fire up Galaxian and route the audio thru a good home stereo.....ick. You don't hear the odd noises on the real game because the speaker can't reproduce it. Or its at least attenuated enough that you don't notice it.

So it would be great if the dev's would include audio eq adjustment overall, and on a game-by-game basis in the tab menu.

[newmanfamilyvlogs]

What we need is support for VST/DX effects 

If they're already using DirectX for sound output

There are tons of speaker cabinet modeling software that would probably liven up the sound of the emulation. A touch of reverb would probably add a good bit too.. just enough to simulate being in a large arcade with a few subtle reflections bouncing off the nearby cabs.

Here's a page with a simple host and some code:
http://www.hermannseib.com/english/vsthost.htm

[javeryh]

I honestly don't know the difference even though I played DK in the arcade a ton (25+ years ago!).  I use MAME .119u4 or something like that and I think I like the sound better than when you used to use samples - it's less annoying/softer, IMO but I honestly don't care at all - I can't tell and so far no one that's played on my cab can tell either.   

[Paul Olson]

I think there have been recent changes to the driver (since maybe .130 or so) that have made it sound worse. I saw a bug report about DKJr sounding muffled on Mametesters from .130.

these games all used the Sanyo EZ monitor which has an audio board on it. This could explain the difference if MAME is now emulating the DK PCB, but not the monitor board. I have a DK; I will try to record it this weekend.

[apocalypto]

I think there have been recent changes to the driver (since maybe .130 or so) that have made it sound worse. I saw a bug report about DKJr sounding muffled on Mametesters from .130.

these games all used the Sanyo EZ monitor which has an audio board on it. This could explain the difference if MAME is now emulating the DK PCB, but not the monitor board. I have a DK; I will try to record it this weekend.

You may be on to something there. Someone else mentioned about the new MAME not accounting for amplifiers on some boards/monitors. If only there was a way to artificially pre-amp specific games in MAME....

That would be great if you could record your DK!

I dont think DK Jr is muffled like DK is, it sounds pretty crisp. However, the jump and land sounds are really wacked out.... but thats another topic! 

[boardjunkie]

Reverb/efx on game sounds? Um....I don't think so. In a busy arcade there was such a din that any reflections would be drown out. And any spkr modeling software is going to be intended for guitar cabs....so useless here. The only thing we need is a way to get rid of the excess low and high end (on a per game basis) that was never there to begin with.

You *could* go as far as to emulate certain cabs that had a disctinct sound/resonance. Pacman cabs have a certain resonance thats pretty recognizable. I'm sure there's others. But....this would be a pretty huge hassle if actually attempted.

Donkey Kong, and most of the early Nintendo stuff was intentionally band limited in the high/high mid area. My guess is they wanted to avoid the screechy Galaxian type audio. They may have taken it to a bit of an extreme, but it made for a unique sound quality.

[boardjunkie]

I think there have been recent changes to the driver (since maybe .130 or so) that have made it sound worse. I saw a bug report about DKJr sounding muffled on Mametesters from .130.

these games all used the Sanyo EZ monitor which has an audio board on it. This could explain the difference if MAME is now emulating the DK PCB, but not the monitor board. I have a DK; I will try to record it this weekend.

Its just an audio amp....no filtering. No need to be concerned with it. Any filtering is done in software, possibly a bit of passive filtering after the D/A converter.

[BadMouth]

I have a DK; I will try to record it this weekend.

Looking at the manual, there is a board with a pot on it for "sound balance", but the game only has one speaker.
Any idea if "sound balance" is more like a bass/treble adjustment and not a left/right adjustment?
That could account for people having very different experiences.

The wording in these old manuals is funny sometimes.
A lot of them refer to any pot as "volume control", even if it's in a steering wheel or mounted gun.
"Theory of Operation" kinda cracks me up too, as if the game hasn't yet been proven to operate correctly.  

[boardjunkie]

Only thing I've ever seen is just a vol control thats grouped with the hold controls and such on the little board to the left of the main chassis.

The monitor's sound amp is interesting in that it uses an output transformer. This may eat up just a touch of high end, but I wouldn't count on alot.

[Jack Burton]

There's a DK cab in a bar near here and the sound on it definitely sounds more muffled and bass heavy than when I play the newest MAME through different kinds of speakers, with exception of Mario's walking sound, which sounds a lot sharper.  I have a big Mitsubishi TV that I've played it through and it sounds good, but different.

I'm thinking it's mostly just the acoustics of the cab and the type of speakers used in that era.  

There's a couple of guys on this forum with MAME'd DK cabs.  Might ask them. 

[Gray_Area]

While I am interested in the solution to this situation, I'm wondering whether this is going to turn into an equivalent of King of Kong. 'Did you use a real Kong board?!'

Of note are that Mario's walking sound does not fluxuate in the above-linked recording. In the retro, tape recording at Coin-Op Videogame.com, it does.

[CrazyKongFan]

There is a separate audio pot on the PCB. I have a cocktail, which has a 4-board stack, one labeled "audio". The pot is on that one, and adjusts either the digital or analog sounds (can't remember which). That being said, mine has a problem where sometimes, it sounds muffled, and sometimes (usually after being in the cold) sounds perfect. I don't think it's the audio board in the 4-board stack, because I have 2 of them and both act the same.

[RayB]

First as has been mentioned, I'm pretty sure the sound hardware is designed in a way that renders the sound best, and this just isn't emulated since it is all audio components as opposed to "computer logic" parts.

Second, you're unfairly comparing the sounds with your examples. a WAV file is uncompressed, while your MP3 is compressed down to a cruddy 64kbs bit rate. Crisp highs are the first thing to suffer with MP3 compression and the lower the quality the more muffled the sound. So... you should be posting an uncompromised WAV file instead.

[apocalypto]

you're unfairly comparing the sounds with your examples. a WAV file is uncompressed, while your MP3 is compressed down to a cruddy 64kbs bit rate. Crisp highs are the first thing to suffer with MP3 compression and the lower the quality the more muffled the sound. So... you should be posting an uncompromised WAV file instead.

you're right, I didnt realize it only converted the mp3 at 64kps. I have replaced it with a WAV file. However, it sounds pretty similar to the mp3 to be honest.


http://fuelhousedemo.com/dkong-mame.wav

[newmanfamilyvlogs]

you're unfairly comparing the sounds with your examples. a WAV file is uncompressed, while your MP3 is compressed down to a cruddy 64kbs bit rate. Crisp highs are the first thing to suffer with MP3 compression and the lower the quality the more muffled the sound. So... you should be posting an uncompromised WAV file instead.

you're right, I didnt realize it only converted the mp3 at 64kps. I have replaced it with a WAV file. However, it sounds pretty similar to the mp3 to be honest.


http://fuelhousedemo.com/dkong-mame.wav

Well the spectrum of that file falls off pretty dramatically after around 4.5kHz, so if the sampling rate was around 9kHz, 20kbps would probably be sufficient to capture the entire wave pretty accurately.

[boardjunkie]

There is a separate audio pot on the PCB. I have a cocktail, which has a 4-board stack, one labeled "audio". The pot is on that one, and adjusts either the digital or analog sounds (can't remember which). That being said, mine has a problem where sometimes, it sounds muffled, and sometimes (usually after being in the cold) sounds perfect. I don't think it's the audio board in the 4-board stack, because I have 2 of them and both act the same.

That sounds like e-lytics in the analog signal path have gone high ESR.....

[Haze]

Honestly, and I don't know how old you are Apocalypto, but the mame audio that is currently in place sounds pretty much like I remember it from back in the day. The one from mikes arcade sounds EQ'd, or of a poor recording quality, and rather shrill to me. I agree that the old DK SAMPLES were much 'brighter', but I never thought they were more 'accurate'. I always remember Donkey Kong sounding like it was inside a shoe box or something, very bass-ey, and very muffled. That machine really rumbled....

Yeah the Mamedevs say its 300% accurate.  Better than sliced bread.

I like the older versions better with samples, and I do not need a cray to play it.

I'm sure Haze will pop up and say its wonderful and perfect, etc.

But I have played it enough times in the arcade, to tell the difference.  

You called? All this trolling is unnecessary.

Derrick has said, if somebody provides direct from PCB recordings, and measurements / readings of all the components on that PCB he'll review it.  Until then the emulation matches the details / information available.

It's acknowledged that no 2 machines will be the same, some have undergone poor repair jobs, or have the wrong components on from the factory, and the schematics are more often than not _wrong_ but until there is more solid information he can't do anything without simply making potentially widely inaccurate guesses.

So far nobody has come to him with that actual information, so things remain the same.  All the people complaining seem to be more hot air than action.

Trolling and just plain insulting the devs for their efforts (which in this area is a considerable effort) is probably only going to make them think the whole thing isn't worthwhile and go and do something else instead, at which point NOBODY will be there to review it even if new information does come to light.

Derrick is doing his job, his further requirements from the community have been laid out plain and clear.  Can't say / ask much fairer than that.

[CapriRS302]

From what I remember DK always sounded a bit muffled in the arcade, for sure.

[Paul Olson]

I sent Derrick a PM on this. Hopefully he still checks the boards after the last fiasco here. I am tied up with school until Mid May, but I will try to get the measurements then. I would like to see this issue put to bed. I have DK, DkJr, and DK3 cabs, so I can check them all.

Paul

[Malenko]

it sounds fine to me  

as for older MAMEs using samples...... set up your front end to use an older version of MAME for that game?

[nordemoniac]

I know this is a long time ago, but I'm pretty sure it's still an issue.

Mario Bros. and Donkey Kong sounds awful, and it's too bad - They're the classics I remember! They sound really muffled and noisy. Not the clear sounds from real cabinets.

Is there any way to inject wav-files into the ROM or something else that might fix the audio?

If it's perfect emulation mame is after, this should be one of their concerns right?

[paigeoliver]

I know this is a long time ago, but I'm pretty sure it's still an issue.

Mario Bros. and Donkey Kong sounds awful, and it's too bad - They're the classics I remember! They sound really muffled and noisy. Not the clear sounds from real cabinets.

Is there any way to inject wav-files into the ROM or something else that might fix the audio?

If it's perfect emulation mame is after, this should be one of their concerns right?

Set up your front end to use one of the older Mame versions that use samples for those titles. I have several mame versions on my cabinet and about a dozen on my desktop for reasons just like that.

Mario Bros and Donkey Kong both have an amplifier on the monitor. I don't know if that is emulated or if it is even doing anything that should be emulated (it is a pretty simple amp).

[nordemoniac]

Is it possible to have multiple MAME versions, so that I can have an older version for these games only? (Sorry, new to this).

[ppv]

Paige, what version of Mame are you using for those two games?

[paigeoliver]

Paige, what version of Mame are you using for those two games?

Both titles were well emulated pretty early on in mame's history from a player's point of view (I don't know how "accurate" everything was behind the scenes but I sure couldn't tell the difference). I am not sure exactly when they ditched the samples, but they seem to play perfectly in .55, .60 and .78.

[dgame]

From http://samples.mameworld.info/

January 15, 2008

As of MAME 0.122u5, the Mario Bros. sample set is no longer required and thus I've moved it to the 'Older Samples' page. Once again, couriersud is responsible for emulating its sound. ~Twisty~

June 15, 2007

As of MAME 0.116, the Donkey Kong sample set is no longer required and thus I've moved it to the 'Older Samples' page ~Twisty~

So it looks like MAME 0.115 is the last version to use the Donkey Kong sample set.

[nordemoniac]

Thanks guys!

Will test with the older versions.

Is it simple to use a different MAME version for these ROMS, but not the rest?

[ahofle]

To me, Donkey Kong sounds pretty damn good and is close to the original, but the running sound in Mario Bros is not even close (last I checked anyway which was around 0.143 or so).  It's way too loud and is distorted -- the samples from the older Mame version sound much closer to the real thing.  I ended up doing what paigeoliver suggested which is to setup an older version in my frontend just for Mario Bros.

[Haze]

To me, Donkey Kong sounds pretty damn good and is close to the original, but the running sound in Mario Bros is not even close (last I checked anyway which was around 0.143 or so).  It's way too loud and is distorted -- the samples from the older Mame version sound much closer to the real thing.  I ended up doing what paigeoliver suggested which is to setup an older version in my frontend just for Mario Bros.

yeah Mario is an acknowledged bug, no idea when it will be fixed.

afaik the devs involved with Donkey Kong consider it perfect at this point, so I can't see it changing.