Depends who you're asking. A random man off the street? A musician or audio producer? Your average Vogoner who posts in the Audio subforum? Personally I've been all 3 at different times, and all three would have a different perspective on all those questions.

I feel that I'm being somewhat led towards the sentiment that it's of limited value to compare recordings of very similar devices, because the differences are so minute that it doesn't really matter in practical terms anyway. An understandable sentiment and certainly true to a degree. I said as much about the Gravis Ultrasound in another thread this very day.

And in a way, that's actually why I asked for a comparison recording. Not because I doubt that this card is excellent, in the slightest. But because I believe it's scientifically important to quantify exactly how it's good - and yes, in niche quarters like these, it will be in fine detail!

Subjective claims such as "omg this is the greatest ISA soundcard ever!" need to be balanced in context with real world A/B comparisons, because if a random man off the street listens to clips A and B and concludes, "hmm these sound pretty much the same!" (or alternatively, "notably different, in ways X, Y and Z"), then they will know without any doubt where they stand on the matter. Without any need to rely upon, or be influenced by, Subjective Claim this or Subjective Claim that.

Shreddoc wrote on 2022-05-12, 04:54:

I feel that I'm being somewhat led towards the sentiment that it's of limited value to compare recordings of very similar devices, because the differences are so minute that it doesn't really matter in practical terms anyway.

No, I don't mean to lead anyone toward any particular conclusion. I am simply providing additional recordings, and asking the questions. Whether any potential buyer thinks that this card is so close to the genuine part that there is no point in paying for "the real thing" at current prices, or that it was a waste of silicon if there is even a single bit in a single sample that is different, real recordings that can support that opinion are essential.

I haven't shared any opinions, nor suggested any conclusions; I am merely providing more data, and encouraging you to form and/or discuss yours.

640K!enough wrote on 2022-05-12, 06:04:

Shreddoc wrote on 2022-05-12, 04:54:

I feel that I'm being somewhat led towards the sentiment that it's of limited value to compare recordings of very similar devices, because the differences are so minute that it doesn't really matter in practical terms anyway.

No, I don't mean to lead anyone toward any particular conclusion. I am simply providing additional recordings, and asking the questions. Whether any potential buyer thinks that this card is so close to the genuine part that there is no point in paying for "the real thing" at current prices, or that it was a waste of silicon if there is even a single bit in a single sample that is different, real recordings that can support that opinion are essential.

I haven't shared any opinions, nor suggested any conclusions; I am merely providing more data, and encouraging you to form and/or discuss yours.

That is good; thanks, it's interesting data. I must note that FM purism isn't my particular axe to grind, I'm of the opinion that several of the leading FM implementations all sound quite Fine & Good.

Your recording proves that there are differences in this case, and that those differences are audibly negligible. Which means different things to different people. To one, it may mean this is the perfect purchase, and to another, it may simply prove that it's essentially the same thing they already have several of.

Either way, data is appreciated, cheers.

From my recollection, at least with the CMI-8738, FM playback through the SPDIF output is an analog-to-digital signal. So, the recorded audio waveform will never look as good as pure digital like the YMF7x4 cards.

Beautiful achievement. This new sound card sounds perfect.
Forgive my late idea, but a PCMIDI MPU module in PC/104 format configurable via all available IRQs ports of the ISA Bus (3, 4, 5, 6, 7, 9, 10, 11, 12, 14, 15) could-be considered?
I have never seen one yet. And, what a 'utopia' to be able to mount several modules in a system.
Certainly for retrogaming, it is useless. But, for your next sound card design, you're going to have to create yet another new card format. That's a shame.
I wish him much success.

P.s.: It just lacks a digital audio out connector internally to be able to connect it to another sound card.

The MPU-401 module is already configurable for Ports 300h/330h and IRQ2/3/4 allowing for several MPU401 devices, though maybe not as many as you want.

http://pcmidi.eu/files/megacard/MPC010-jumpers.png

OK, so in line with the demand earlier, I've dug out the following YMF719 card, it is nearly identical to the Audician 32.

I have also recently acquired and restored this Terratec Gold 16/96 ES1868 which is one of the highest quality ISA cards in my collection. I will also record from that as well.

I should post recordings of Dune - Arrakeen Palace from both in a few days.

Cool card, terrible chip, sorry.

kitten.may.cry wrote on 2022-05-24, 19:53:

Cool card, terrible chip, sorry.

How so?

the terratec has a really clean output, but from my experience almost too clean x3 ... it sounded a bit "sterile" to me, I almost prefer the warm sound of my CT1740 although it's noisy.

but that's all very subjective

Is there a comparison of MK8330 and Orpheus that would highlight the differences between these two? I mean "usability" differences, not what hardware it uses.

appiah4 wrote on 2022-05-24, 22:37:

kitten.may.cry wrote on 2022-05-24, 19:53:

Cool card, terrible chip, sorry.

How so?

I mean, I've seen it on every cheap motherboard, didn't sound great, they're even below CT4810 imo.

what about giving a chance to people that have proven they know what they are doing, making a proper card with proper audio grade components, before judging stuff as "shit" because it was used on cheap hardware?

"it sounded like shit on shitty hardware". Uhhhh well, you don't say?

Boohyaka wrote on 2022-05-25, 09:47:

what about giving a chance to people that have proven they know what they are doing, making a proper card with proper audio grade components, before judging stuff as "shit" because it was used on cheap hardware?

"it sounded like shit on shitty hardware". Uhhhh well, you don't say?

I didn't? Where did you get this idea from? Re: MK8330 : new soundcard based on the CMI 8330 chipset

kitten.may.cry wrote on 2022-05-25, 09:35:

appiah4 wrote on 2022-05-24, 22:37:

kitten.may.cry wrote on 2022-05-24, 19:53:

Cool card, terrible chip, sorry.

How so?

I mean, I've seen it on every cheap motherboard, didn't sound great, they're even below CT4810 imo.

"sounds great", "didn't sound great" - meaningless terms. If such a claim can't be quantified with real recordings, comparisons, or other actual data, then it's about as valuable as any marketing claim.

You like the CT4810 card's FM implementation compared to the CMI8330 chip's one, do you? Or which other properties were you referring to, in particular?

the CMI8330 is just fine, I think I mentioned it previously but the reason people often didn't like it is because it usually was implemented on super cheap cards or motherboards with lackluster audio circuits, same goes for the ALS100 and similar for example.

Shreddoc wrote on 2022-05-25, 11:51:

kitten.may.cry wrote on 2022-05-25, 09:35:

appiah4 wrote on 2022-05-24, 22:37:

How so?

I mean, I've seen it on every cheap motherboard, didn't sound great, they're even below CT4810 imo.

"sounds great", "didn't sound great" - meaningless terms. If such a claim can't be quantified with real recordings, comparisons, or other actual data, then it's about as valuable as any marketing claim.

You like the CT4810 card's FM implementation compared to the CMI8330 chip's one, do you? Or which other properties were you referring to, in particular?

I hate both, not many seem to agree, that's cool, to each their own.

kitten.may.cry wrote on 2022-05-25, 12:04:

Shreddoc wrote on 2022-05-25, 11:51:

kitten.may.cry wrote on 2022-05-25, 09:35:

I mean, I've seen it on every cheap motherboard, didn't sound great, they're even below CT4810 imo.

"sounds great", "didn't sound great" - meaningless terms. If such a claim can't be quantified with real recordings, comparisons, or other actual data, then it's about as valuable as any marketing claim.

You like the CT4810 card's FM implementation compared to the CMI8330 chip's one, do you? Or which other properties were you referring to, in particular?

I hate both, not many seem to agree, that's cool, to each their own.

A recording's already been posted in the thread proving how close to original the 8330's FM implementation is. We had to inspect the waveforms very closely indeed to catch the differences.

The CT4810, on the other hand, has (iirc) that abomination of a "wavetable FM" style implementation.

You're welcome to hate them both, but they're very different.

kitten.may.cry wrote on 2022-05-25, 12:04:

I hate both, not many seem to agree, that's cool, to each their own.

If we're talking about FM synth implementations between the CMI8330 and Creative's CT4810, they are fundamentally different.

The CMI chip attempts to mimic the original OPL3 sound as closely as possible, and uses actual FM synthesis to do so. On the other hand Creative's CT4810 (which is a rebadged Ensoniq Audio PCI) emulates OPL3 using its wavetable. This is objectively a far worse implementation. Examples of the CT4810's OPL3 emulation can be heard in this video.

Joseph_Joestar wrote on 2022-05-25, 12:12:

kitten.may.cry wrote on 2022-05-25, 12:04:

I hate both, not many seem to agree, that's cool, to each their own.

If we're talking about FM synth implementations between the CMI8330 and Creative's CT4810, they are fundamentally different.

The CMI chip attempts to mimic the original OPL3 sound as closely as possible, and uses actual FM synthesis to do so. On the other hand Creative's CT4810 (which is a rebadged Ensoniq Audio PCI) emulates OPL3 using its wavetable. This is objectively a far worse implementation. Examples of the CT4810's OPL3 emulation can be heard in this video.

Even CQM looks better in comparison to the CT4810's OPL3 emulation. 🤣