Discrete Vs. Integrated YM3438: Sega Genesis music blind test results.

Quote:

---

Originally Posted by TmEE

db don't range from 0 to 1, but infinity to 0, where 0 is loudest.

Half the buzzing comes from power rails themselves. Hook up a scope to some and look how much ripple there is.

---

Hmm, I see what you mean. I always watch my waveform with amplitude scale instead of DB scale. But here 1.0 means 0 db and all my stuff is normalized at 0 db. I think my graphs should give a pretty good idea of the relative amount of noise from one sample to the other. I cant make much sense of the DB view waveform in audacity, sometimes the signal goes completely flat.
Like in the "silent" section of the M2VA4 recording here: Maybe you can help me make sense of it?
http://i1204.photobucket.com/albums/...psb87ee8a6.jpg
Zoomed on
http://i1204.photobucket.com/albums/...psf3885a91.jpg

I would gladly check with an oscilloscope for ripple in the power rail but I think the one I acquired recently ( http://www.sega-16.com/forum/showthr...-Oscilloscope!) might be defective. I tried checking the output of a 4 mhz oscillator and some game gear data lines with it and I had to set it at 10 mv/division before I could see anything moving on the display.
I do not have a proper signal generator to test it... I bought a cheap adjustable oscillator off ebay and was hoping to try my luck with it... Didnt get around to do that yet.

Quote:

---

Originally Posted by sheath

I just added the CHC and D&D VA6 and TFIV and Sonic 2 VA3 files you posted above to the blind test files from before and cycled them through on the same playlist. With headphones I have to be paying attention to notice, with my PC speakers plus sub I only occasionally notice one being a little more filtered, I think it is the Sonic 2 VA3 one.

---

Wow! Im having a hard time believing this. To me the difference is very obvious. Listening to model 1 right after CCAM feels like putting your head underwater.

It seems especially obvious with thunderforce. I can blind test this one without efforts 100% of the times, in fact I just tried. Works both with my headphones and my audio monitors. I can tell instantly.
D&D on model 1 also has a shitton of DAC artifacts that gives it away, in fact there is so much digital noise that it sounds like a new instrument is playing in the song lol.

But anyway most of the differences is in the high frequencies. If you have very poor sound equipment, a noisy surrounding or, forgive me to say, poor audition, I suppose it is possible you would miss most of it.
See this site and check up to what frequency you can hear.
http://www.noiseaddicts.com/2009/03/...-hearing-test/

Quote:

---

Originally Posted by MaxWar

Hmm, I see what you mean. I always watch my waveform with amplitude scale instead of DB scale. But here 1.0 means 0 db and all my stuff is normalized at 0 db. I think my graphs should give a pretty good idea of the relative
But anyway most of the differences is in the high frequencies. If you have very poor sound equipment, a noisy surrounding or, forgive me to say, poor audition, I suppose it is possible you would miss most of it.
See this site and check up to what frequency you can hear.
http://www.noiseaddicts.com/2009/03/...-hearing-test/

---

I can hear all of them! :p

I work at a noisy job and have my hearing tested once a year, but always use hearing protection properly.

Quote:

---

Originally Posted by MaxWar

Wow! Im having a hard time believing this. To me the difference is very obvious. Listening to model 1 right after CCAM feels like putting your head underwater.

---

I should have been more clear. The differences are noticeable when scrutinizing. When I was just listening for something that sounded "bad" I would check winamp on the task bar and it was more frequently a high frequency FM instrument or a sample from one of the CCAM recordings than not. When I was just listening passively while working on graphics or typing something up my ears never distracted me from what I was doing. That is only when listening with my PC speakers in my normal office with PC fan noise and personal fan noise going. With headphones I can nitpick each sample no problem, though I'm not sure I'd say I prefer one over another still.

Quote:

---

Originally Posted by MaxWar

It seems especially obvious with thunderforce. I can blind test this one without efforts 100% of the times, in fact I just tried. Works both with my headphones and my audio monitors. I can tell instantly.
D&D on model 1 also has a shitton of DAC artifacts that gives it away, in fact there is so much digital noise that it sounds like a new instrument is playing in the song lol.

---

A couple of the recordings have artifacts that seem to cover the FM instruments as well. I was going to listen to them all again with fresh ears when the house is empty before trying to peg which one(s) seem to do this and under what circumstances.

Quote:

---

Originally Posted by MaxWar

But anyway most of the differences is in the high frequencies. If you have very poor sound equipment, a noisy surrounding or, forgive me to say, poor audition, I suppose it is possible you would miss most of it.
See this site and check up to what frequency you can hear.
http://www.noiseaddicts.com/2009/03/...-hearing-test/

---

The highest frequencies in FM and PSG usually annoy me more than filtering does. That is, outside of SNES SPC filtering which annoys me the instant the sound starts. I'll reply in a minute about the hearing test over my ancient JVC headphones. Tomorrow I will hopefully have a chance with an empty house to try it with the Logitech PC setup. That's not even getting into the Yamaha Dolby Prologic Tuner plus RCA speakers and Yamaha sub that I use for my actual gaming. ;)
-edit-
JVC Headphones for the hearing test, 8Khz -12Khz is ear bleed territory, way too sharp. 14Khz-16khz are more soft, with the later most being hard to hear except for the pop at the beginning and end. With 17Khz-22Khz all I hear is the pop at the beginning and end. I do have the TV on right now, or rather the wife does, so I'll test this again in a quiet house.
-edit-edit-
Yamaha Dolby Prologic with RCA Speakers and Yamaha Sub, no sound whatsoever after 12Khz for myself or mine Frau. I don't even remember what the frequency cut off was for the RCA speakers in particular, I just remember that after I bought them I could hear line noise that wasn't present before.

My right ear can hear very clearly up to 18k. I also hear 19k but its not as loud.
If I crank the volume I can make out 20 and 21k but 22k is just hopeless.

Strangely my left ear is not as good. It can hear up to 17k clearly but can barely make anything at 18k, even with high volume. 19k and on is dead silence on the left ear.

Some people in the comments complain that what some people hear at the higher frequencies are harmonics from poor sound samples and not the actual pitch. I don't know about that, certainly is not a professional test in controlled conditions but should give a good idea.
Im 31, one of my friend is about 25 and he could not hear anything past 14 or 15k. And if you check at the comments bellow, sounds like quite a few people have pretty dull hearing.

Edit: Sheath, sounds like that yamaha setup sucks, but then that might explain how some people get such poor results.

Logitech PC speakers transmit a clear buzzing sound from 18khz-20khz. I guess 18Khz has a little bit of the high pitch that I was expecting, but nothing above that does. I can only hear a pop and brief hum and then silence beyond that. Below that I had to turn the volume down or it was ear splitting. I guess my logitech speakers are pretty high frequency. I'm going to test my Yamaha Dolby Prologic again in a quiet house, but yeah I wouldn't suspect that it has the high frequencies. In fact that has caused me to hesitate to upgrade for years, I'm worried about all of the hum and signal noise I might start hearing. ;)

Quote:

---

Originally Posted by sheath

In fact that has caused me to hesitate to upgrade for years, I'm worried about all of the hum and signal noise I might start hearing. ;)

---

Just get some isolation transformers. I use behringer HD400 on all my audio setups, it is close to miraculous.

Quote:

---

Originally Posted by MaxWar

Hmm, I see what you mean. I always watch my waveform with amplitude scale instead of DB scale. But here 1.0 means 0 db and all my stuff is normalized at 0 db. I think my graphs should give a pretty good idea of the relative amount of noise from one sample to the other. I cant make much sense of the DB view waveform in audacity, sometimes the signal goes completely flat.
Like in the "silent" section of the M2VA4 recording here: Maybe you can help me make sense of it?

---

When you see the flat then just zoom it more, and when you cannot you have basically hit the precision of your recording. One more thing that is useful when testing are spectrograms, they are much better at analyzing relative silence, but you have to toy around with the settings to get best results.

I find Audacity pretty poor for this kind of stuff, but I am used to Cool Edit Pro (aka Adobe Audition these days), that thing is serious business.

Also on modern sound cards do not record with any sample rate that is not a multiple of 48KHz. All other rates are digitally scaled from 48/96/144/192 to what you chose in 99% of devices. Anything AC97 or HD Audio suffers from this, they are only supposed to use 48KHz or a multiple, the standard does not support stuff like 44.1KHz.

Man now I want to finally try this mod sounds promising and I've always liked the look of the model 1 genesis/mega drive

Quote:

---

Originally Posted by CathodeRayTubes

Man now I want to finally try this mod sounds promising and I've always liked the look of the model 1 genesis/mega drive

---

Yes its very cool, Im now happily enjoying CCAM on my model 1 :)
Note that there are some compatibility issues with discrete ym3438 though. At least most of them can be fixed by piggybacking it to the original Ym2612 in the way that Evildragon devised, but there is a possibility of more, such as reported problems with Earth worm Jim. One of the cool aspect of doing this mod on a model 1 however is that since you keep the ym2612 anyway you can easily switch to stock sound and even have both chips play at the same time and mix them together.
http://www.sega-16.com/forum/showthr...ome-Back/page2
http://img.photobucket.com/albums/v3...ps6e4e3659.png

Quote:

---

Originally Posted by MaxWar

Yes its very cool, Im now happily enjoying CCAM on my model 1 :)
Note that there are some compatibility issues with discrete ym3438 though. At least most of them can be fixed by piggybacking it to the original Ym2612 in the way that Evildragon devised, but there is a possibility of more, such as reported problems with Earth worm Jim. One of the cool aspect of doing this mod on a model 1 however is that since you keep the ym2612 anyway you can easily switch to stock sound and even have both chips play at the same time and mix them together.
http://www.sega-16.com/forum/showthr...ome-Back/page2
http://img.photobucket.com/albums/v3...ps6e4e3659.png

---

Actually those timing issues are more due to differences in the /SOUND line. Also, would you please try to wire the YM3438 directly to the Z80 lines? That should eliminate the timing problems as well.

Interesting.
Im working on different projects atm though, but eventually I will put more time on this console and probably will try what you say.

So what makes the model 1's superior to model 2's in sound quality? Is it that it's in stereo? From what I gather from this, using CCAM with the ym3438 on a model 1 is in theory the best sound we can get from a Sega?

Quote:

---

Originally Posted by ironnick23

So what makes the model 1's superior to model 2's in sound quality? Is it that it's in stereo? From what I gather from this, using CCAM with the ym3438 on a model 1 is in theory the best sound we can get from a Sega?

---

Stereo has nothing to do with the sound quality difference of both an average model 1 & 2 Genesis. Both can output stereo sound out of the box with no modification required.

A lot of model 2's use bad opamps and bad mixing circuits Some of the better sounding ones still use a second order lowpass filter (steeper high frequency roll off) just like the bad sounding ones do and typically use an ASIC YM3438 with one revision using a GOAC YM3438. 2 revisions have a messed up mixing circuit but actually use a discrete YM2612.

The best sound is a subjective preference but the intended sound should come from a Model 1 with a motherboard revision of VA3 through VA6.8 using the original stock YM2612 as this is what SEGA seems to use in a lot of commercialized recordings or soundtracks. No modifications required to use.

Just a warning: Early model 1s have a preamp problem resulting in premature clipping when things get loud (VA0 - VA2) but can be fixed with a modification. They also have a higher cutoff point in the lowpass filter than any of the other models 1 (more high frequency a.k.a. treble content). The last revision the VA7 sounds as bad as the average model 2.

The model 3 despite outputting mono has surprisingly good sound and can be modified for stereo output. All model 3s use a GOAC YM3438.

If you want a different sound I suggest you look at the Mega Amp as it's newer than the Crystal Clear Audio Mod (CCAM) and supports more hardware revisions. The CCAM only supports ASIC YM3438 versions of the Mega Drive / Genesis, other versions require you to use a discrete YM3438 as part of the modification process.

I have a va3 with the ym2612 and I am very happy with the sound. Your complete breakdown makes perfect sense, thank you! If I come into possession of another revision I am definitely interested in hearing the differences the ym3438 can make.

I know this thread is pretty old and that Evildragon hasn't been here for quite some time. But using his schematic, I did this mod about a month ago and have made quite a few recordings directly from my Genesis to demonstrate what it sounds like. My Genesis is a va3 and was just recapped all the way around by me about 2 years ago. I have RCA L/R jacks added to the back tied directly off the sony 1034 amp chip and not through the headphone section. There is still ground noise in some games that can be heard, but interesting, some games generate this and many do not. So something in the cartridge construction within the games makes this more noticeable. To hear a few of the songs just check here:

https://www.youtube.com/watch?v=DHoE...BYmcPaoOdJ8pZZ