Genesis - SNES audio comparison

sheath · Post by **sheath** » Mon Jan 19, 2009 2:00 pm

tomaitheous,

Is it not possible for somebody to see error in statements like "there is no argument, the SNES just sounds *way* better" and want to set the record straight? I basically agree with your last paragraph, and I say just that on my "sega fan" website over and over. People who are biased for the popular console from each generation tend to think I'm just like them, only I prefer the opposite console. I actually prefer arcade action games over any other genre, and this has caused me to lean heavily on the libraries of certain systems.

People will prefer what they played first and have the fondest memories of, regardless of technical things. I just want to point out that the technical observations that people spout to support their bias are incorrect. Furthermore, I point out consistently that a more refined and objective view would appreciate both for different reasons.

Say what you want about me or this board, but I am not being dishonest and I don't believe this board has been either. Genesis and SNES sound are technically incomparable in any sort of overall sense. If the consensus had been in your favor, I would have reported that. If you or somebody else had provided some sort of objective test, I would have included that. Instead I will attempt to describe the unique qualities of Genesis *and* SNES sound, and I will state that one of them will only be preferred on a subjective basis.

Knurek · Post by **Knurek** » Mon Jan 19, 2009 4:53 pm

SmartOne wrote:Dragging MDX files onto Hoot shows a menu:
PLAY
STOP
They won't play...

Make sure you have mxdrv.bin in the same dir as hoot.exe. Then pressing the PLAY button will actually do something.

It's available on the main Hoot page:
http://dmpsoft.s17.xrea.com/hoot/index.html
Search for mxdrv(2001/03/30 and download the first link after that mxdrv20010330.cab

SmartOne · Post by **SmartOne** » Mon Jan 19, 2009 7:13 pm

Ah, so mxdrv.bin is compressed inside the CAB (saving, uh... 2 KB?) It works now. Thanks! You might want to add this information about MDX files to your Hoot site.

Why do MDX's sound different than the actual game drivers that you host on your site?

tomaitheous · Post by **tomaitheous** » Mon Jan 19, 2009 8:33 pm

People will prefer what they played first and have the fondest memories of, regardless of technical things. I just want to point out that the technical observations that people spout to support their bias are incorrect. Furthermore, I point out consistently that a more refined and objective view would appreciate both for different reasons.

What biased or incorrect technical observations would those be? I remember seeing any outrageous claims by snes fans before.

Say what you want about me or this board, but I am not being dishonest and I don't believe this board has been either.

I'm not sure what you mean. What have I've said about this board, other than there might be some bias-ism? It's a Sega board about programming for your old beloved consoles, how can you not expect some bias-ism?

If the consensus had been in your favor, I would have reported that.

Wait, so the popular opinion wins the vote? Isn't that what you're arguing against form the SNES fanboys?

If you or somebody else had provided some sort of objective test, I would have included that.

Objective test of what? What are you going to test for that isn't already known?

Genesis and SNES sound are technically incomparable in any sort of overall sense.

One, I disagree with that statement (I feel like a skipping record). In general, they are comparable. They both produce sound and both have specific strengths. It's when you get into specific details that they are less comparable. But regardless, you still needs those specific details to know the strengths and weaknesses of each, and to give weight to the strengths and weaknesses in their design.

Without good amount of audio theory under your belt, as well as some intermediate knowledge of both chips and designs, I don't see how you can really write an informative and objective comparison piece.

And saying that they are not really comparable is a convenient way of arguing one isn't more capable or superior to the other (or more precisely that the SNES audio isn't really superior). It's not like you're comparing a Jet to a car. That's the fallacy in this. I mean, you're out to prove people wrong or correct their views or whatever. This remains me of the 7800 VS the Coleco thread in another forum.

Enough my rambling. Anyway, good luck with the comparison piece and I do look forward to seeing your site updated with some more stuff. Alway love a good fan site

Knurek · Post by **Knurek** » Mon Jan 19, 2009 9:31 pm

SmartOne wrote:Why do MDX's sound different than the actual game drivers that you host on your site?

Because pretty much all of them are fanmade conversions, and as such they might differ from the original game data.

Snake · Post by **Snake** » Tue Jan 27, 2009 3:35 am

tomaitheous wrote:What biased or incorrect technical observations would those be?

How about...

Just play the opening to Star Fox sometime...

Starfox? You gotta be kidding me.

tomaitheous wrote:It's a Sega board about programming for your old beloved consoles, how can you not expect some bias-ism?

Well not from me, I owned a NES, a Gameboy, an Amiga, and two SNESes before I ever owned a single Sega console - but for me, the C64 and the Genesis produced far more interesting audio than all of them.

Maybe some of these 'SNES fanboys' haven't actually heard much on the Genesis.

tomaitheous wrote:They both produce sound

But in an entirely different way, which means... they can't be compared.

tomaitheous wrote:And saying that they are not really comparable is a convenient way of arguing one isn't more capable or superior to the other (or more precisely that the SNES audio isn't really superior).

Well in many ways, it isn't. Example:

I think the SNES sound was far superior... especially played through a sound system with some big speakers...

SNES sound will always sound muddy through 'a sound system with some big speakers', because its below 40KHz. The Genesis sounds far cleaner, even if its only playing an 8KHz sample, because the hardware is still writing to the DAC higher than 40KHz.

Given that the YM2612 is entirely digital - each operator (there are 24 of them) is playing a sample from an internal ROM. (Incidentally, you have complete control over all of them, so you could use the thing as a 24 channel chip if you so desired.) Ok, so given that this ROM is at a fixed address and size, that simplifies things somewhat. However, the bottom line is that the chip is a lot more complicated than a simple sample playing device, is doing all of the same things, and a lot more. But it doesn't play user-supplied samples from RAM, because that's not what it was designed for. I'm not really seeing how that makes this chip 'inferior'. Quite the opposite.

Christuserloeser · Post by **Christuserloeser** » Wed Jan 28, 2009 7:43 am

Snake wrote:
tomaitheous wrote:And saying that they are not really comparable is a convenient way of arguing one isn't more capable or superior to the other (or more precisely that the SNES audio isn't really superior).
Well in many ways, it isn't. Example:
I think the SNES sound was far superior... especially played through a sound system with some big speakers...
SNES sound will always sound muddy through 'a sound system with some big speakers', because its below 40KHz. The Genesis sounds far cleaner, even if its only playing an 8KHz sample, because the hardware is still writing to the DAC higher than 40KHz.

That is true for the model #1 Genesis / Mega Drive, at least those with "High Definition" logo. But yeah, these do blow away practically everything on SNES.

- The model #2 units do sound more like 11khz though.

sheath · Post by **sheath** » Fri Jan 30, 2009 8:14 pm

Okay, so I've started working on a new comparison section. I've recorded fifty or so Genesis and SNES games in their demonstration modes using an older Avermedia video capture card that does native MPEG-2 compression. I've already converted the videos to DIVX 6.8 using Virtual Dub, but the result is not only lossy but just plain distorted at points.

The good news is the Audio is uncompressed 16-bit 48Khz PCM that sounds great. The bad news is that I can't honestly claim the video quality is authentic to anything, not Standard Definition, not Composite, not even Coaxial signal quality. I'd say that at least 85% of the frames are good. But the things that go wrong in the backgrounds in other scenes are so glaring that I'm just torn about even using the videos. I am not in the position currently to buy another card, does anybody have any codec related suggestions that might have better results? I am willing to post a few of the offending videos if anybody wants to critique them.

Snake · Post by **Snake** » Fri Jan 30, 2009 10:57 pm

Try Xvid. DivX has sucked majorly for a long time.

sheath · Post by **sheath** » Fri Jan 30, 2009 11:08 pm

I tried that once but it was creating the same artifacts. I'll make sure I've got the most up to date codec and give it another go.

-edit

Yeah, Divx 6.8 can get to where most scenes have no artifacts anywhere on the screen if I encode at 1500 (edit: should have been 4900) kbps. Xvid, even with the target quantizer at maximum, just seems to cause more of the same problem I'm having with Divx.

It's probably not that big of a deal, I just really wanted to present the videos as an accurate representation of the actual consoles at full screen. That's probably a bit silly anyway, as everybody is going to be viewing them on different types of monitors under different lighting conditions anyway. We'll see how it goes.

Chilly Willy · Post by **Chilly Willy** » Fri Jan 30, 2009 11:45 pm

What I normally do is record live video using a simple 1-pass constant quality setting with a very high bitrate, then go back once it's done and re-encode the video with 2-pass vbr with all the bells and whistles (b-frames, hi-q search, etc). That's the only way to make a video of a live event that looks good at a small size.

sheath · Post by **sheath** » Sat Jan 31, 2009 12:09 am

Oh, yeah, I'm only at phase one of your approach technically. I've encoded at the highest (least compressed) quality setting in the capture card's native MPEG-2 format, and I've tried to convert it over to MPEG-4 format using Dr. Divx 1 & 2, Divx 7.0 Converter, and Virtual Dub plus Divx 6.8 or Xvid. I've posted the most offensive of the bunch below. I'm actually now wondering if the problem is Media Player Classic with the latest K-Lite codecs actually outputting the video ugly. These are 34MB and 46.6MB files respectively, feedback is very welcomed.

DIVX
http://www.gamepilgrimage.com/16-bitCom ... NES001.avi

XVID
http://www.gamepilgrimage.com/16-bitCom ... VID001.avi

tomaitheous · Post by **tomaitheous** » Sat Jan 31, 2009 1:16 am

SNES sound will always sound muddy through 'a sound system with some big speakers', because its below 40KHz.

The SNES sounds muddy because the sampling output of 32khz? I don't think 32khz sample output sounds muddy to me. I don't have my A500 anymore, but I don't remember it sounding muddy either. Nor my 32khz oki ADPCM software decoder demo and it sounds pretty damn clear to me. I would attribute the muddy sound of most SNES games due to the developer using smaller sample sizes (and sample point-to-point filter enabled to hide the aliasing). And as far as I know, it's not a period based divider either.

However, the bottom line is that the chip is a lot more complicated than a simple sample playing device

A simple sample playing device would be the A500/1000 or a GUS. I would say the SPC700 is a little more than a simple sample playing device. Sample point-to-point filter enable per SS index block group, loop enable bit per SS index block group, multilevel reverb per channel, left/right panning per channel (real stereo), not to mention hardware envelopes. It does more than just pitch bending. You've also got a high res timer and a cpu to create new samples on the fly. There's a SID and NES emulator for it. Granted, is not as intricate as FM - but that's just thing, it doesn't need to be.

I'm not really seeing how that makes this chip 'inferior'. Quite the opposite.

So having a limited number of operators, algo's, and only a single waveform is not a limiting factor for sound generation? Like I said previously, it's a means to an end. The 2612 can only replicate so many types of instrument sounds. The SPC700 unit can replicate many more types of instrument sounds and more realistically, or other type of synth sound sounds the 2612 cannot. X-Men Spider-man Arcade revenge is pretty good example of this.

To be honest. I rather not even argue the SPC700 side. I seemed to have gotten stuck doing it though. I'm really not interested in the SPC700, save for if I did something like Memblers did with his NES sound emulator or that SID player, but for some other system. I'd rather see older chips pushed to make sounds that people never expected from them. For me, that's a big part of the appeal to console dev.

Snake · Post by **Snake** » Sat Jan 31, 2009 9:38 pm

tomaitheous wrote:The SNES sounds muddy because the sampling output of 32khz?

Yes. Maybe you can't hear it. Most people can, which is why people use higher sample rates these days.

tomaitheous wrote:I would say the SPC700 is a little more than a simple sample playing device.

Not really.

tomaitheous wrote:a cpu to create new samples on the fly.

Good luck with that

tomaitheous wrote:There's a SID and NES emulator for it.

Well you could do that with anything that can play a sample. But it's missing the point anyway, since the people arguing this stuff are talking about games. Unfortunately no games did anything like this, at least none that I've heard of.

tomaitheous wrote:So having a limited number of operators, algo's, and only a single waveform is not a limiting factor for sound generation?

Missing the point of what I was saying again - which is, it would be far easier to have made the chip just play samples. It's far more complex than that. Yamaha didn't just do this 'for a laugh', rather, they believed it was more powerful. Which it is.

tomaitheous wrote:Like I said previously, it's a means to an end. The 2612 can only replicate so many types of instrument sounds.

But it's not about 'replicating' sounds, its about creating new ones. The SPC700 can't do that, it can only replicate. Most SNES music uses samples taken from other synths, 90% of which could have been generated on the YM2612 in the first place - minus the obvious loop points, lower sample frequency, and the fact that they sound odd when played at different frequencies.

It's all very well to argue that a sample playing device can recreate the sound of anything, and its true to a certain extent (i.e. if you have enough memory to make lots of different high quality samples at different pitches - which is not the case here). But first of all you need something to sample. You can't start from nothing.

But this is why the two chips can't be compared. They're very different. One is about sound generation, the other sound replication.

tomaitheous wrote:I'd rather see older chips pushed to make sounds that people never expected from them. For me, that's a big part of the appeal to console dev.

Absolutely.

Chilly Willy · Post by **Chilly Willy** » Sat Jan 31, 2009 11:04 pm

Snake wrote:
tomaitheous wrote:The SNES sounds muddy because the sampling output of 32khz?
Yes. Maybe you can't hear it. Most people can, which is why people use higher sample rates these days.

I think it's more a matter of what you're used to. Today, we're used to 48 or 96 kHz sample rates. Back then, 8kHz was something to rant about.