The Grand OPL3 Comparison Run!

[MaxWar]

OFFICIAL UPDATE #5

This is the Sound Blaster Pro 2 update (CT1600)

You can download the files in one go here : https://docs.google.com/open?id=0BzsnL20-4a37UFUxMGgzcFlONkk

Or you can go directly to the grand OPL3 Comparison run site:
https://sites.google.com/site/soundcardcomparison/

I felt many things should be said about this card and i fear my short description might be a bit lackluster. If you have some suggestion and relevant infos please share.
I really think i should be trying to turn this project into a kind of Open to all Wiki...

And sry for the long time since laste update, too many geeky projects at the same time .

The sound blaster Pro 2 is a well loved sound card, and for good reasons. It is a landmark card often used to compare other cards. ... ( add blah blah... )
I found the output quality to be quite good. Noise level is low, no audible distortion and it packs alot of power in the bass section. When connected to my main sound system this card could generate some quality wall rumbling, I remember thinking how this was a badass card. However when compared to some of my favourite sounding sound cards, I actually find it a bit too boomy. The higher frequencies are cut low and it lacks a bit in brightness for my taste. It is in no way a deal breaker as this card is renowned for its compatibility good quality. You can always EQ it to taste. The mixer settings has a Low pass filter that is on by default. You can deactivate it with the SBP-SET utils included in the drivers. It is supposed to only be applied to Digital audio and not FM, however i did not try it myself and the recordings were done with the default settings.

[Great Hierophant]

I think your issues with the ATI Stereo F/X and Mediavision Thunderboard and Monkey Island and Commanche : Maximum Overkill have nothing to do with the compatibility of the cards themselves. You are running these games and recording from them on an Intel Celeron 500MHz system. The system is way too fast for OPL2 chip, and the issues will show up on a true Sound Blaster or Ad Lib as well. Slow the machine down as much as possible and try running these games again.

[elianda]

I agree with Great Hierophant that all the games that were targeted for a 286 system like Dune, Monkey Island etc. should be recorded at thsi speed. The issues you describe are the same that appear if you run the game on a fast PC. You should consider that music playback routines are always timed routines and depending on implementation it can be critical that the timing routines mess up on fast CPUs.
On my K6 system MI messes up already at about 200 MHz and works good at lower speed. But going to 386 level performance is much more reliable with the timing of old music playback routines.

[MaxWar]

I think your issues with the ATI Stereo F/X and Mediavision Thunderboard and Monkey Island and Commanche : Maximum Overkill have nothing to do with the compatibility of the cards themselves. You are running these games and recording from them on an Intel Celeron 500MHz system. The system is way too fast for OPL2 chip, and the issues will show up on a true Sound Blaster or Ad Lib as well. Slow the machine down as much as possible and try running these games again.

The SFX and Thunderboard recordings were done on my 486.
Lately ive beeen using mostly my 486 instead of the celeron as i get stuttering audio with The Dig on the Celeron. The 486 is overdriven with an evergreen 5x86 upgrade though. However i tested my real SB 2.0 (CT1350b) on the celeron 500 and monkey island played fine.

I agree further testing could be done with many of these cards. I have about 50 cards in total i want to record so i was a bit hasty on side testings sometimes.

This makes me realize however that i should always indicate what system was used for each card, considering how i sometime use the 486 and sometimes the Celeron , sometimes both. When i get issues with a card on one system i usually test it in the other. From now on i will specify the system used on all my tests.

In any case my 386 has the SFX drivers already on it. I could install monkey island on it and give it a go. Would this rule out the CPU speed possibility? :p

Edit: typo

[raymangold22]

This comparison is quite interesting however, there are a few errors:
#1 It seems like a lot of people actually can't tell genuine OPL3 when they hear it. A lot of Vibra-16s have OPL3, even if there is not a YMF262 present (the FM is actually integrated like CT1747: which in some cases, sounds more pleasant than YMF262). In other words, it's essentially a YMF262 integrated inside a creative chip...
This is quite unfortunate because we have a lot of people who are throwing out or desregarding genuine OPL3 sound cards.
I believe the CT4180 as aforementioned actually does have genuine YMF262 integrated in the CT2505-TDQ2 chip (which is labeled Vibra16C, not "S" or "XV").

#2 ESFM (ESS's FM implementation) is hit or miss. Some MIDIs it can sound absolutely fantastic, other ones, it does not. Also behaving to certain commands, it can become distorted in higher frequencies.
--> if the audio sounds thin, that is due to the card's implementation and not ESFM itself.

#3 Crystal FM Synthesis! This is probably my favourite due to the distinctive "square hats" in the percussion, as well as that traveling hissy cymbal. There were different versions of this manufactured by crystal, some earlier versions having broken FM, but the good ones are quite nice.

I actually talked about the OPL3 variants here. As well as recordings of these awesome variants... typically I run dual OPL3 which is a Crystal FM and YMF262 in unison.

[Ace]

#1 It seems like a lot of people actually can't tell genuine OPL3 when they hear it. A lot of Vibra-16s have OPL3, even if there is not a YMF262 present (the FM is actually integrated like CT1747: which in some cases, sounds more pleasant than YMF262). In other words, it's essentially a YMF262 integrated inside a creative chip...
This is quite unfortunate because we have a lot of people who are throwing out or desregarding genuine OPL3 sound cards.
I believe the CT4180 as aforementioned actually does have genuine YMF262 integrated in the CT2505-TDQ2 chip (which is labeled Vibra16C, not "S" or "XV").

You, sir, are WRONG. If a SoundBlaster does not have a YMF262, CT1747, YMF289 or any other chip with an OPL logo on the chip itself, it DOES NOT have OPL3. Those have CQM, which is quite inaccurate to true OPL3.

[raymangold22]

You, sir, are WRONG. If a SoundBlaster does not have a YMF262, CT1747, YMF289 or any other chip with an OPL logo on the chip itself, it DOES NOT have OPL3. Those have CQM, which is quite inaccurate to true OPL3.

I do know what "CQM" is, and I've observed it extensively on the later AWE32 I have. The bass response is incorrect, and the noise generator is way off (much like DOSBox).

I'll start to do recording tests tomorrow. I believe the Vibra16C chip has the CT1747 integrated inside it (Vibra16XVs do not, purely CQM). This would make sense from what I heard after playing some MIDIs, but I'll do some recordings of it tomorrow once I get home.

For the record, if I'm not mistaken, CQM wasn't around in 1995 yet. IT just so happens that the CT2505-TDQ2 chip on the 4180 Vibra16s were made in 1995 (as per this photo).
So if this was the case, the chip would have no choice but to integrate CT1747, thus being genuine YMF262.

We'll soon know, tomorrow.

[MaxWar]

This comparison is quite interesting however, there are a few errors:
#1 It seems like a lot of people actually can't tell genuine OPL3 when they hear it. A lot of Vibra-16s have OPL3, even if there is not a YMF262 present (the FM is actually integrated like CT1747: which in some cases, sounds more pleasant than YMF262). In other words, it's essentially a YMF262 integrated inside a creative chip...
This is quite unfortunate because we have a lot of people who are throwing out or desregarding genuine OPL3 sound cards.
I believe the CT4180 as aforementioned actually does have genuine YMF262 integrated in the CT2505-TDQ2 chip (which is labeled Vibra16C, not "S" or "XV").

#2 ESFM (ESS's FM implementation) is hit or miss. Some MIDIs it can sound absolutely fantastic, other ones, it does not. Also behaving to certain commands, it can become distorted in higher frequencies.
--> if the audio sounds thin, that is due to the card's implementation and not ESFM itself.

#3 Crystal FM Synthesis! This is probably my favourite due to the distinctive "square hats" in the percussion, as well as that traveling hissy cymbal. There were different versions of this manufactured by crystal, some earlier versions having broken FM, but the good ones are quite nice.

I actually talked about the OPL3 variants here. As well as recordings of these awesome variants... typically I run dual OPL3 which is a Crystal FM and YMF262 in unison.

I only recorded 2 Vibra cards so far. The CT4180 and the CT2260 (genuine YMF262). Maybe the OPL3 is integrated on the CT1747 as you say and maybe it sounds like a real yamaha chip i just do not know. What i know is that the CT4180 does not sound like a real opl3 to me. In fact, it sounds almost identical to the SB32 (CT3600)

The best way to appreciate the magnitude of the difference with my recordings is with Descent intro.
Listen to Descent intro With CT4180(https://dl.dropbox.com/u/43851675/CT4180/Descent-Intro-CT4180%28Vibra16c%29.flac) and CT3600(http://dl.dropbox.com/u/43851675/CT3600/Descent-intro-CT3600.flac), pretty similar uh?
Then listen to it with a typical good quality opl3 chipped card such Rock16 (http://dl.dropbox.com/u/43851675/MAD16/descent-intro-MAD16.flac)or earlier sound blasters.

The bassy lead synth at the beginning is very different on the integrated creative chips, quite duller in fact. This is the most flagrant example i think but other differences are here also, such as more `chimey` or sort of `metallic`sounding hi hats. Basically i find the CT4180 to be very consistent with CT3600, a card which i was led to believe was a prime example of the CQM clones.

This sound similarity is the reason why I affirmed with alot of certitude that the CT4180 is a clone. I appreciate your input rayman and this is a nice example of how recordings can help solve questions like that. Im definitely looking forward to getting samples from other vibra card and see how some of em might be surprisingly interesting.
About the crystal FM, i sure do not like the card i have here, but curious as to hear other variants. Im not one of those who spits on anything that is not a tru branded yamaha chip, even though from all my experiment so far i definitely like the real chip bettter.

Oh And I like that Dual OPL3 setup that you use, ill have to try that one day

Edit: Added links for your convenience!

[Ace]

For the record, if I'm not mistaken, CQM wasn't around in 1995 yet. IT just so happens that the CT2505-TDQ2 chip on the 4180 Vibra16s were made in 1995 (as per this photo).
So if this was the case, the chip would have no choice but to integrate CT1747, thus being genuine YMF262.

I have two different Vibra16Cs, a CT2960 and CT4180. I KNOW the sound of CQM, and I can assure you the CT2505-TDQ2 has CQM just like my SoundBlaster 32 PnP CT3600, SoundBlaster 16 CT2940, SoundBlaster 16 WavEffects CT4170 and my two SoundBlaster AWE64 CT4520s.

[Eep386]

Now now, it's possible that Creative DID make a non-CQM sub-variant, seeing how Creative sure was fond of them. Let's hear his recording and judge from that.

[MaxWar]

Now now, it's possible that Creative DID make a non-CQM sub-variant, seeing how Creative sure was fond of them. Let's hear his recording and judge from that.

I doubt they would have given it the same CT number.

[Eep386]

Again, we'll have to hear his recording to find out for sure.

In any case we can put this question to rest once and for all.

[Eep386]

It's also possible that his CQM equipped card may sound better than most CQM cards (better output filtering etc.)

[Cloudschatze]

In any case we can put this question to rest once and for all.

Was there a big debate about this somewhere?

There's basically a zero-percent chance that the CT2505 contains Yamaha's OPL3 core, for a number of reasons, including the mentioned trademark issue. There's also the fact that the chips would need to have been manufactured in Japan (by Yamaha), and would have been marked as such.

Let's try some deductive reasoning though. Creative's debut CQM offering was the standalone CT1978 chip, often found paired with the CT2502 ASIC. Does it make sense that the CT2505, a later chip, would contain anything other than Creative's own CQM derivative?

[Eep386]

I heard that CT4180 clip posted a few posts back... CQM is so abominably chirpy. And the bass is just totally killed. Just like I remember CQM...

Raymangold22 tells me his sounded different... that's why I am so interested.

[raymangold22]

In any case we can put this question to rest once and for all.

Was there a big debate about this somewhere?

There's basically a zero-percent chance that the CT2505 contains Yamaha's OPL3 core, for a number of reasons, including the mentioned trademark issue. There's also the fact that the chips would need to have been manufactured in Japan (by Yamaha), and would have been marked as such.

Let's try some deductive reasoning though. Creative's debut CQM offering was the standalone CT1978 chip, often found paired with the CT2502 ASIC. Does it make sense that the CT2505, a later chip, would contain anything other than Creative's own CQM derivative?

I actually think CT1749 was the first to implement CQM (as per that SB32).

[Cloudschatze]

I actually think CT1749 was the first to implement CQM (as per that SB32).

The CT1749 is simply a plug-and-play bus-interface chip, and lacks any sound generation capabilities. You seem to have overlooked the CT1978 chip in that photo...

[raymangold22]

Okay so I finally did runs of that 4180. And wow, that has to be the best sounding CQM -- EVER.

It *is* CQM, but its response is so well articulated with the nichicon caps and philips amp.
(I'm guessing MaxWar has the non-philips amp varation of the card, as those recordings were not pleasant in the least, probably gross caps like "wincaps" or "jackcon" as well).

Here's a quick youtube sample I uploaded.
http://www.youtube.com/watch?v=MGoV-OPdbRE

I'll do some more recordings later. But as you can see, the bass (while off), has good balance, and the highs aren't TOO distorted.

Now if you listen to this without headphones on average speakers (and to tracks you're not familiar with in YMF262), it can sound fairly believable at first.

RE: Cloudschatze, yes I probably did overlook a lot.

[Teppic]

Here's a YMF262 version if you want to compare:
http://www.youtube.com/watch?v=8xwybXhfwzw

[Ace]

It *is* CQM, but its response is so well articulated with the nichicon caps and philips amp.

Seriously? Think I need to check my Vibra16C CT4180 and see what caps and amp my card has.

By the way, you were using Speaker Out on this Vibra16C, right? Or were you using Line Out?

Oh, and one last thing: PLEASE don't say "emulation" unless you're referring to software emulation. From my experience, people seem to immediately think "software emulation" when anyone brings that up.