Reply 20 of 44, by SaxxonPike
- Rank
- Member
Is this limited to just Creative cards? I have a CT3600, but I also have a later Aztech card and a Labway YMF724 card which have compatibility.
Sound device guides:
Sound Blaster
Aztech
OPL3-SA
Is this limited to just Creative cards? I have a CT3600, but I also have a later Aztech card and a Labway YMF724 card which have compatibility.
Sound device guides:
Sound Blaster
Aztech
OPL3-SA
Most older DOS games were mixed on a real SB 1.0 or 2.0 cards with no mixer, so the music and sound effects sound the most balanced with these cards as the developers intended.
Later SB cards or clones have a mixer but the balance between the FM and PCM might not be as accurate as the original SB 1/2 cards because of DSP and DAC changes, that is why I requested old authentic SB cards.
The goal was to make newer SB cards, clones, and dosbox to sound as balanced as the old SB cards on which the games were mixed.
With the test tools I attached to this thread and Jo22 samples, we discovered that on a SB2.0 the volume difference between the OPL and PCM is 8.5db with a 440Hz sine tone.
It may be slightly different on the SB 1.0 or 1.5 model, we don't know yet.
With this 8.5db value I can adjust the mixer of a SB16 for example to have the same OPL to PCM balance as the SB2.0 card.
This is great!
Have Creative SBPro's, SB16's or AWE32's actually been off the mark at full mixer volume? Any data on specific models?
Yeah why not, another reference is always good.
A few recordings of the SBPro will be very nice, it will let us know if the SBPro has the same balance as the SB2.0.
Newer cards like SB16 are considered as the culprit to be fixed, not as the source.
Meanwhile here are some results:
Name FM PCM Result
---- --- --- ------
CT1350B_1 21.7 13.3 = 8.5 db
CT1350B_2 14.5 6.0 = 8.5 db
QuickShot 11.6 3.1 = 8.5 db
CT2230 34.2 26.2 = 8.0 db
CT2890 28.3 16.6 = 11.7 db
CT2950(CQM) 26.9 19.7 = 7.2 db Distorted CQM signal.
YMF719 25.7 18.1 = 7.6 db
DOSBox 14.0 8.1 = 5.9 db
It's all over the place.
I take the SB 2.0 as the reference card with a delta of 8.5 db.
Now we need a SB1.0 or 1.5 and a SBPro 2 (SC1600).
Here you go. Feels like there are so many variables in recording the SB Pro. This thing popped at the start of 440_0DB.WAV playback every time.
Mine is one of the later CT1600's with new Creative logos.
Also here's the CT2940 with Vibra Pro CT2502 and YMF289 OPL3-L.
Thank you firage!
Results:
CT1600 7.7db
CT2940(OPL) 10.0db
As I thought, the SBPro DSP volume is different to the SB2.0 but very similar to the Yamaha YMF719 which emulates a SBPro.
Another SBPro and a SB1.0 should complete and solidify the research.
When the Aztech's are here I'll shall measure them too, both with a YMF262.
EDIT:
The smaller the delta db value, the louder the OPL is.
I'd say that 8.5db to 7.5db should do fine, as 1db considered a typical volume change that people can begin to hear.
Can you further define the difference between FM and PCM as well as which sound cards it applies to? I see that you equate PCM with SB, which is how my Audician seems to define it (since SETUPSA has both an SB and FM mixer setting). Is PCM/SB the same as Adlib? In other words, are we talking about a separate volume that is Adlib-specific?
For my CT2800, I don't see such a mixer setting. I have Voice, MIDI, CD, Line, Mic, PC Speaker, and Master. So Voice is digital effects, MIDI must be FM music and the rest are obvious.
Thanks.
The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks
Adlib was just an FM synth unit, the Yamaha OPL2. Sound Blaster brought digital sound along with OPL2 FM. CMS was another weird side step.
Then to achieve stereo FM SB Pro 1 had two OPL2 chips, and starting with SB Pro 2 they used OPL3 which could do most of what a dual OPL2 did. FM synth music wasn't a priority for them anymore by the time AWE64 and their Vibra chipsets came out and they dropped real Yamaha chips for their own CQM approximation.
Yeah, FM is often labeled MIDI.
@firage: You don't mention PCM/SB. I'm actually familiar with what you posted. I'm clear on what FM/MIDI is, just not clear on what James means by PCM/SB (and why my CT2800 mixer doesn't have a separate volume for PCM or SB).
The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks
Yeah, besides "digital sound effects" PCM digital sound is also often labeled Voice or speech, because those were some of the main applications that came to their minds back then for stuff FM/MIDI wasn't good for.
Yep, on the SBPro and SB16 WAV/PCM is called Voice in the mixer.
On the Yamaha YMF719 or Audician32 Setupsa.exe it is called SB.
EDIT: I was wrong, read following posts.
Any digital sound that does not come from the FM chip, comes from the DSP/PCM.
With the old SB1/2 the balance between these two was fixed.
So in comparing the Audician to the CT2800 mixer settings,
CT2800(left)/Audician(right)
Voice=SB
MIDI=FM
CD=CD
??=WAVE
@James-F: there doesn't seem to be a mixer setting on the CT2800 that equates to the WAVE mixer setting on the Audician. So on the Audician, if SB=WAV/PCM, then what is the WAVE setting?
The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks
Probably short for Wavetable? Those Yamaha cards came with provisions for an extra DSP and onboard ROM.
Edit: Apparently not! A whole different kettle of fish.
Hello everyone, I've done some more recordings.
For the game GODS, I used the SoundMachine, SoundBlaster 2.0 and the Protac card.
That ThunderBoard clone was already in the machine and so I thought it would be a good idea to do
a recording for comparison.
I also turned the volume knob a bit up for a better SNR, I hope I didn't overdrive the cards' amp. 😅
At least Audacity's input levels were on the safe side..
I also did another recording of the Protac card.
Previously, I had to use another test because MPXPlay thought it was a SB Pro..
This time I converted the wave and played it back using playfile.
"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel
//My video channel//
Thanks Jo22!
This should give another excellent reference.
Good news!
I got the CMS upgrade working! 😁
I was very nervous to not to ruin my REV 4 card..:
It is the oldest SB2 I've got and I did a recap just recently.
Before the first try I even removed the YM3812 (FM1312) as a safety measure.
Besides, it was part of the bequest of one our families friend who passed away a few years ago.
A few weeks before he went, he also gave me his beloved C64 as a gift. He was such a nice person.
So when I got that card from his wife (among other geeky things) and repaired it (new voltage regulator),
I wanted to keep this one as a memento of him. Luckily, everything went fine this time.
I think he would be happy to see that his old "junk" wasn't just thrown away, but was helpful and brought
joy to other people. 😀
"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel
//My video channel//
Excellent!
Now we have a complete reference SB 2.0 card.
Thank you again Jo22.
wrote:Excellent!
Now we have a complete reference SB 2.0 card.
Thank you again Jo22.
Glad I could help!
And thank you very much for this thread!
I've always felt that there was a difference between SB2 and SB16 when playing old games,
but I've never found out that they had different balance.
Also greetings to the other people in this thread who helped. You're awesome. 😎
And last but not least, thanks to Jepael for his cms440 utility!
"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel
//My video channel//
wrote:EDIT:
The smaller the delta db value, the louder the OPL is.
I'd say that 8.5db to 7.5db should do fine, as 1db considered a typical volume change that people can begin to hear.
To what extent would the frequency response affect this measurement?
I mean, early SB cards had quite a 'bassy' sound to them, because they had heavy low-pass filtering of the signal to limit noise and aliasing from the low samplerate and 8-bit sampling.
Later SBs and clones tend to sound 'brighter' and more 'hifi', because they have CD-quality DACs on board, and they don't low-pass filter things as aggressively as the early ones did.
I can imagine that this might affect the comparison here.
I know it affects gaming at least, because if I play an old game like DOOM, the sounds seem to be tailored to these old 'bassy' SBs, where it sounds really fat and gritty. Play it on a modern SB or clone, and it sounds very bright and thin, and takes away from the overall experience.
The OPL chip has no lowpass filter on any of the SB cards.
A typical early SB card (1, 2, pro) filter starts at 3.2kHz so the Wave/Voice 440Hz test frequency is low enough that it isn't affected by the low pass filter, plus MPXPlay player disables the filter if it is run on a SBPro card.
So both 440Hz test tones from the OPL and from the DSP are not affected by any lowpass filters.
The lowpass filter on the SB2.0, SBPro and SB16 differ from one another.
The most filtered is the SB16 where it has absolutely nothing above half the sampling rate, but the SB2.0 and SBPro low pass filters is much less abrupt and very similar in character.
That means the quality of the lowpass filter does indeed have an audible affect on how the balance between OPL and WAV sounds, but since most SBPro cards (and their clones) have the same Lowpass filter or at least SHOULD have the same filtering, we can adjust the difference by db values of the 440Hz tones and say it is pretty close.
Sound cards that radically differ in their Lowpass filter design or completely lack one (Yamaha YMF71x) will sound imbalanced even after matching the 440Hz tones to the samples on this thread.
That is why I asked for a recording of the game GODS where the music consists of the WAV and OPL combined so we can tweak the balance by ear even though the subject sound card has improper lowpass filter.