VOGONS


Reply 320 of 486, by LSS10999

User metadata
Rank Oldbie
Rank
Oldbie
yawetaG wrote:

They just translate FM instructions to General MIDI PCM or other PCM patch set compatible instructions. Depending on the set of PCM patches used, some instruments will be more accurate than others. I personally cannot hear the difference between some of the instruments used when playing back FM music through my AudioPCI and the OPL of my Vibra 16 sound card, while other instruments are clearly different.
I expect that if a PCM patch set were to contain patches based on the sound of a real FM sound card/synthesizer (which can be identified fairly easy by their characteristic timbre), the emulation would be a bit more accurate.

That might be the case for cards that don't contain real FM logic, such as Creative PCI audio cards.

However, I don't think there's an accurate list on which audio chip really contained FM logic. AFAIK ES1938 (Solo-1), FM801 and YMF7x4 contained real FM logic, and Creative PCI (AudioPCI/Live!/Audigy), CMI8738* don't. Not sure about some other chipsets (such as Aureal Vortex and ESS Maestro/Allegro series**).

* I once read CMI8738's documents mentioned "FM emulator" so it's likely not real FM logic.
** This site rated both ESS Maestro and Aureal Vortex's FM as "Tinny" as they don't sound quite authentic. The reviews mentioned something about OPL2/OPL3 "clones" on those cards but I kind of question this as neither's documents explicitly mentioned FM synthesizers (ESS Maestro, AU8830 for example), but did mention something about Wavetable.

Reply 321 of 486, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++
LSS10999 wrote:

* I once read CMI8738's documents mentioned "FM emulator" so it's likely not real FM logic.

I've reviewed the same information and agree, that's what the datasheet suggests. However, user experience has been much more positive than wavetable based approaches from other manufacturers. If that's what they're using, they're clearly doing it better than pretty much everyone else.
I've got the OPL3 synth from a CMI8738 working in Windows 7 x64

All hail the Great Capacitor Brand Finder

Reply 322 of 486, by yawetaG

User metadata
Rank Oldbie
Rank
Oldbie
LSS10999 wrote:
That might be the case for cards that don't contain real FM logic, such as Creative PCI audio cards. […]
Show full quote
yawetaG wrote:

They just translate FM instructions to General MIDI PCM or other PCM patch set compatible instructions. Depending on the set of PCM patches used, some instruments will be more accurate than others. I personally cannot hear the difference between some of the instruments used when playing back FM music through my AudioPCI and the OPL of my Vibra 16 sound card, while other instruments are clearly different.
I expect that if a PCM patch set were to contain patches based on the sound of a real FM sound card/synthesizer (which can be identified fairly easy by their characteristic timbre), the emulation would be a bit more accurate.

That might be the case for cards that don't contain real FM logic, such as Creative PCI audio cards.

However, I don't think there's an accurate list on which audio chip really contained FM logic. AFAIK ES1938 (Solo-1), FM801 and YMF7x4 contained real FM logic, and Creative PCI (AudioPCI/Live!/Audigy), CMI8738* don't. Not sure about some other chipsets (such as Aureal Vortex and ESS Maestro/Allegro series**).

* I once read CMI8738's documents mentioned "FM emulator" so it's likely not real FM logic.
** This site rated both ESS Maestro and Aureal Vortex's FM as "Tinny" as they don't sound quite authentic. The reviews mentioned something about OPL2/OPL3 "clones" on those cards but I kind of question this as neither's documents explicitly mentioned FM synthesizers (ESS Maestro, AU8830 for example), but did mention something about Wavetable.

Yamaha owns the patents to their particular flavour of FM (see https://en.wikipedia.org/wiki/Frequency_modulation_synthesis). Other manufacturers therefore couldn't use exactly the same type of synthesis for their own FM synths - Casio got in hot waters with their Phase Distortion synths over this, as apparently it was too similar - and so they will sound slightly different. Maybe some cards contain a OPL clone, but the manufacturers didn't want to draw too much attention to it out of fear for a lawsuit from Yamaha...

Reply 323 of 486, by LSS10999

User metadata
Rank Oldbie
Rank
Oldbie
yawetaG wrote:

Yamaha owns the patents to their particular flavour of FM (see https://en.wikipedia.org/wiki/Frequency_modulation_synthesis). Other manufacturers therefore couldn't use exactly the same type of synthesis for their own FM synths - Casio got in hot waters with their Phase Distortion synths over this, as apparently it was too similar - and so they will sound slightly different. Maybe some cards contain a OPL clone, but the manufacturers didn't want to draw too much attention to it out of fear for a lawsuit from Yamaha...

Yeah, I once read about this "patented" fact regarding Yamaha's OPL FM synths, and it appeared the reason why ESS stopped offering their ESFM in later models (Maestro/Allegro) is probably related to this. On the other hand, FM801 does have an authentic-sounding FM (its document also mentioned it).

Besides, it's not "slightly" different... the FM music from most PCI sound cards sounded *significantly* different from the original (as they're easily distinguishable), ranging from simple pitch differences (Aureal series) to playing FM through MIDI wavetables instead (like Creative/Ensoniq AudioPCI, which sounded completely different from original).

Last edited by LSS10999 on 2019-06-14, 05:39. Edited 1 time in total.

Reply 324 of 486, by yawetaG

User metadata
Rank Oldbie
Rank
Oldbie
LSS10999 wrote:
yawetaG wrote:

Yamaha owns the patents to their particular flavour of FM (see https://en.wikipedia.org/wiki/Frequency_modulation_synthesis). Other manufacturers therefore couldn't use exactly the same type of synthesis for their own FM synths - Casio got in hot waters with their Phase Distortion synths over this, as apparently it was too similar - and so they will sound slightly different. Maybe some cards contain a OPL clone, but the manufacturers didn't want to draw too much attention to it out of fear for a lawsuit from Yamaha...

Yeah, I once read about this "patented" fact regarding Yamaha's OPL FM synths, and it appeared the reason why ESS stopped offering their ESFM in later models (Maestro/Allegro) is probably related to this. On the other hand, FM801 does have an authentic-sounding FM (its document also mentioned it).

They might have licensed the FM equations from Yamaha for the FM801 (if made before 1995, see below). BTW, it isn't just the OPL synths (OPL is just a particular chip type implementing Yamaha FM), it applies to all of Yamaha's FM applications (but see the Wiki page). One thing mentioned on the Wiki page is that the patents lapsed in 1995, so anything made after 1995 can basically use Yamaha's patented method for free.

Besides, it's not "slightly" different... the FM music from most PCI sound cards sounded *significantly* different from the original (as they're easily distinguishable), ranging from simple pitch differences (Aureal series) to playing FM through MIDI wavetables instead (like Creative/Ensoniq AudioPCI, which sounded completely different from original).

"Slightly" was a figure of speech. All of the ones with real FM chips will feature certain characteristics associated with FM synthesis, even if the underlying logic has differences and therefore the sound output is different.

The latter, playing through MIDI wavetables/PCM samples, is easily explainable by the fact that they're not using the same synthesis method at all, so why people would expect them to sound exactly similar is beyond me. Ignoring marketeers claiming they would, of course...but we all (should) know that what marketeers say tends to be far from the truth up to some degree (quite some degree, in some cases, if not outright lying). Some people even say that using PCM samples is not synthesis because everything is pre-computed and therefore there's only a limited set of possibilities regarding how the sound can be manipulated...

Reply 325 of 486, by LSS10999

User metadata
Rank Oldbie
Rank
Oldbie
yawetaG wrote:

They might have licensed the FM equations from Yamaha for the FM801 (if made before 1995, see below). BTW, it isn't just the OPL synths (OPL is just a particular chip type implementing Yamaha FM), it applies to all of Yamaha's FM applications (but see the Wiki page). One thing mentioned on the Wiki page is that the patents lapsed in 1995, so anything made after 1995 can basically use Yamaha's patented method for free.

"Slightly" was a figure of speech. All of the ones with real FM chips will feature certain characteristics associated with FM synthesis, even if the underlying logic has differences and therefore the sound output is different.

The latter, playing through MIDI wavetables/PCM samples, is easily explainable by the fact that they're not using the same synthesis method at all, so why people would expect them to sound exactly similar is beyond me. Ignoring marketeers claiming they would, of course...but we all (should) know that what marketeers say tends to be far from the truth up to some degree (quite some degree, in some cases, if not outright lying). Some people even say that using PCM samples is not synthesis because everything is pre-computed and therefore there's only a limited set of possibilities regarding how the sound can be manipulated...

You're right... some might actually have licensed the FM for their audio chips to use... Actually, the "original" in that scope was mainly OPL3 (YMF262/YMF289), as that's what most ISA soundcards have, and that's how FM in DOS games usually sounded. Of course, OPL3 isn't the only FM synth, and each FM synth has its own traits that defines how instruments sound.

Even though ESFM is not OPL3, it sounded very close to it, with only some minor differences (on some specific soundtracks/instruments). However, under Windows the ESFM sounded a lot more different, and this difference is mainly considered as an extension to Yamaha's OPL3.

Yet I still consider it a bit strange that while some cards appear to play FM through wavetables, they don't appear to offer native General MIDI playback on port 330h (like post-ES1938 audio cards). AudioPCI/SBLive! is among the ones that do have support for MIDI on port 330h.

Although the FM synth is a bit off-topic, it is, however, one of the only things on those sound cards that could be made use of with little effort on modern chipsets, where DMA-based operations (such as digitized audio) no longer works.

Reply 326 of 486, by Kamerat

User metadata
Rank Oldbie
Rank
Oldbie
gdjacobs wrote:
LSS10999 wrote:

* I once read CMI8738's documents mentioned "FM emulator" so it's likely not real FM logic.

I've reviewed the same information and agree, that's what the datasheet suggests. However, user experience has been much more positive than wavetable based approaches from other manufacturers. If that's what they're using, they're clearly doing it better than pretty much everyone else.
I've got the OPL3 synth from a CMI8738 working in Windows 7 x64

CMI8738 passes nukeykt's OPL3 clone test while the FM801 someone talking about here doesn't.

DOS Sound Blaster compatibility: PCI sound cards vs. PCI chipsets
YouTube channel

Reply 327 of 486, by LSS10999

User metadata
Rank Oldbie
Rank
Oldbie
Kamerat wrote:

CMI8738 passes nukeykt's OPL3 clone test while the FM801 someone talking about here doesn't.

The FM801 document does mention about a FM synthesizer. However, I think that test mainly imply whether the FM core in a given sound card (real or not) is a logic-perfect replication of OPL3 (that it even replicated the original hardware bugs).

I'm not sure to how much extent the hardware bugs affect regular games' music (as I personally never really noticed), but from what I've read regarding the topic, the bugs do make a difference in some demos.

Reply 328 of 486, by appiah4

User metadata
Rank l33t++
Rank
l33t++
Kamerat wrote:
gdjacobs wrote:
LSS10999 wrote:

* I once read CMI8738's documents mentioned "FM emulator" so it's likely not real FM logic.

I've reviewed the same information and agree, that's what the datasheet suggests. However, user experience has been much more positive than wavetable based approaches from other manufacturers. If that's what they're using, they're clearly doing it better than pretty much everyone else.
I've got the OPL3 synth from a CMI8738 working in Windows 7 x64

CMI8738 passes nukeykt's OPL3 clone test while the FM801 someone talking about here doesn't.

I was talking about the FM801 and never said it's a 100% OPL3 clone, it's just a genuine FM synthesizer and is pretty accurate, though not perfect.

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 329 of 486, by MKT_Gundam

User metadata
Rank Member
Rank
Member

So CMI8738 has some software GM like the Audician 32? My Mobo has the soft GM synthesizer option in the MIDI tab. Only sfx works fine, like doom 1 and 2

Retro rig 1: Asus CUV4X, VIA c3 800, Voodoo Banshee (Diamond fusion) and SB32 ct3670.
Retro rig 2: Intel DX2 66, SB16 Ct1740 and Cirrus Logic VLB.

Reply 330 of 486, by Srandista

User metadata
Rank Oldbie
Rank
Oldbie
Srandista wrote:

Hopefully I will find 4CoreDual-VSTA in future, and for sure will test if sound works on Windows 98 on it as well. SATA2 isn't important for me in the slightest, HDDs from that time couldn't even saturate SATA1, but missing sound in 98 was real showstopper for me.

So now, after I put 4CoreDual-VSTA in my case, and test it, I can confirm this. And indeed, ESS Solo-1 works on it flawlessly, in DOS, Win XP and even in Win 98! So yeah, it really was a matter of chipset. And even though, that with newest VIA VT8237S you can use ESS Solo-1 for DOS gaming, it wouldn't work with 98 for some inexplicable reason. With VIA VT8237A, it works just fine in 98 as well.

Also, maybe a little bit off-topic question, but I'm curious. Is there option to get digital signal through S/PDIF from ESS Solo-1 (with some external bracket)?

Socket 775 - ASRock 4CoreDual-VSTA, Pentium E6500K, 4GB RAM, Radeon 9800XT, ESS Solo-1, Win 98/XP
Socket A - Chaintech CT-7AIA, AMD Athlon XP 2400+, 1GB RAM, Radeon 9600XT, ESS ES1869F, Win 98

Reply 331 of 486, by appiah4

User metadata
Rank l33t++
Rank
l33t++

Solo-1 datasheet has no mention of digital output and I've never seen a Solo-1 card with it.

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 332 of 486, by Kamerat

User metadata
Rank Oldbie
Rank
Oldbie
Srandista wrote:

So now, after I put 4CoreDual-VSTA in my case, and test it, I can confirm this. And indeed, ESS Solo-1 works on it flawlessly, in DOS, Win XP and even in Win 98! So yeah, it really was a matter of chipset. And even though, that with newest VIA VT8237S you can use ESS Solo-1 for DOS gaming, it wouldn't work with 98 for some inexplicable reason. With VIA VT8237A, it works just fine in 98 as well.

Also, maybe a little bit off-topic question, but I'm curious. Is there option to get digital signal through S/PDIF from ESS Solo-1 (with some external bracket)?

I were unable to get the Solo-1 to work in Windows 98SE on my Asus P5VD2-VM SE that also utilizes the VIA VT8237S southbridge.

DOS Sound Blaster compatibility: PCI sound cards vs. PCI chipsets
YouTube channel

Reply 333 of 486, by Srandista

User metadata
Rank Oldbie
Rank
Oldbie

Yeah, from my testing (and now your confirmation), I'm pretty sure it's chipset issue. Fortunately for me, between 4CoreDual-VSTA and SATA2, there's no difference that would make me go back to SATA2. Especially, when for 98 I'm running IDE drives.

Socket 775 - ASRock 4CoreDual-VSTA, Pentium E6500K, 4GB RAM, Radeon 9800XT, ESS Solo-1, Win 98/XP
Socket A - Chaintech CT-7AIA, AMD Athlon XP 2400+, 1GB RAM, Radeon 9600XT, ESS ES1869F, Win 98

Reply 334 of 486, by Kamerat

User metadata
Rank Oldbie
Rank
Oldbie

I have to correct myself. The VxD driver doesn't work while the WDM driver does. The SATA controller works in Windows 98SE but only at SATA I speeds.

DOS Sound Blaster compatibility: PCI sound cards vs. PCI chipsets
YouTube channel

Reply 335 of 486, by Srandista

User metadata
Rank Oldbie
Rank
Oldbie

On my 4CoreDual-SATA2, neither VxD, nor WDM drivers works. I don't remember now, if there was some exclamation mark in Device manager, but everything in Sounds options in Control panel was grayed out and I couldn't get any sound to be played in 98. And also, I want to use Dreamblaster on my Solo, so I do want to use VxD drivers.

Socket 775 - ASRock 4CoreDual-VSTA, Pentium E6500K, 4GB RAM, Radeon 9800XT, ESS Solo-1, Win 98/XP
Socket A - Chaintech CT-7AIA, AMD Athlon XP 2400+, 1GB RAM, Radeon 9600XT, ESS ES1869F, Win 98

Reply 336 of 486, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

I finally tested GA-P75-D3P and with Yamaha 724 its working so B75 chipset and with high probability all its derivatives (X7X) are Dos sound capable.. same as X6X which already made Rayer working - difference is that he has some addition PCI chip for because his chipset has not native intel PCI and B75 has native intel PCI, i dunno so far if it better or now.. but i a bit hoped that it would..
Here is table, which 1155 chipset has native PCI and which doesnt - PCI column:
https://en.wikipedia.org/wiki/List_of_Intel_c … ipsets#LGA_1155

I once again had issues with that cheap Yamaha 744 Aopen Cobra, it not worked in DOS, in Windows / Linux it was fine, but i dos i god DSDMA freeze and problem with IRQ utility.. With 722 its working fine, i guess that with my Labtec 744 it would be fine too. I dunno if its something but with this card design or its simply half broken - i tired multiple slots, i now really cared about very good slots contact, but its not working in Dos. I have other some Aureal V1 for testing on the way, because i had it only in other location and transport it its annoying, i want to use it in other machines.

Good is that MBs with it as X79, have usually UEFI, it means again that UEFI is not problem, i dont have now time for that but maybe these MBs could be a bit more compatible for DOS (for PCI slots are nice too, there is LPT, COM Realtek NIC and HD sound with digital audio..) than X58 (2,3 PCI slots) and X79 (only 1 PCI) and in my case 1 PS/2 (other not gigabyte vendors maybe have 2 x PS/2, but not primary pci-e gpu slot selection, so no go for me).
Other advantage is native USB3 ports for Win7+, disadavantage is not primary PCI-E slots selection, if you have fiddle with dual gpu stunt.. what is annoying but i made it working in Windows and XP/7.. so i have have some other PCI-E gpu for Win98.. and if want to i could have PCI one for DOS.. Slots design is nice too, i can have 2 full - 2 slots PCI-E gpus with blocking only 1 pci slot and there is still free 1x PCI-E slot for.. for NIC or Sata controller (on board has only 1x Sata 6Gb/s + 1x Msata, rest of ports are 3 Gb/s)

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 337 of 486, by LSS10999

User metadata
Rank Oldbie
Rank
Oldbie
ruthan wrote:
I finally tested GA-P75-D3P and with Yamaha 724 its working so B75 chipset and with high probability all its derivatives (X7X) a […]
Show full quote

I finally tested GA-P75-D3P and with Yamaha 724 its working so B75 chipset and with high probability all its derivatives (X7X) are Dos sound capable.. same as X6X which already made Rayer working - difference is that he has some addition PCI chip for because his chipset has not native intel PCI and B75 has native intel PCI, i dunno so far if it better or now.. but i a bit hoped that it would..
Here is table, which 1155 chipset has native PCI and which doesnt - PCI column:
https://en.wikipedia.org/wiki/List_of_Intel_c … ipsets#LGA_1155

I once again had issues with that cheap Yamaha 744 Aopen Cobra, it not worked in DOS, in Windows / Linux it was fine, but i dos i god DSDMA freeze and problem with IRQ utility.. With 722 its working fine, i guess that with my Labtec 744 it would be fine too. I dunno if its something but with this card design or its simply half broken - i tired multiple slots, i now really cared about very good slots contact, but its not working in Dos. I have other some Aureal V1 for testing on the way, because i had it only in other location and transport it its annoying, i want to use it in other machines.

Good is that MBs with it as X79, have usually UEFI, it means again that UEFI is not problem, i dont have now time for that but maybe these MBs could be a bit more compatible for DOS (for PCI slots are nice too, there is LPT, COM Realtek NIC and HD sound with digital audio..) than X58 (2,3 PCI slots) and X79 (only 1 PCI) and in my case 1 PS/2 (other not gigabyte vendors maybe have 2 x PS/2, but not primary pci-e gpu slot selection, so no go for me).
Other advantage is native USB3 ports for Win7+, disadavantage is not primary PCI-E slots selection, if you have fiddle with dual gpu stunt.. what is annoying but i made it working in Windows and XP/7.. so i have have some other PCI-E gpu for Win98.. and if want to i could have PCI one for DOS.. Slots design is nice too, i can have 2 full - 2 slots PCI-E gpus with blocking only 1 pci slot and there is still free 1x PCI-E slot for.. for NIC or Sata controller (on board has only 1x Sata 6Gb/s + 1x Msata, rest of ports are 3 Gb/s)

Not sure how much free upper memory area you'd have with this setup. AHCI or bootable external hard disk controllers (or internal ones not directly provided by the southbridge) are known to take up a good amount of precious upper memory blocks for its BIOS/option ROM to function. Some modern video cards with larger BIOS images might also reduce usable upper memory but not as significant compared to the former (depending on the card itself).

Having less usable upper memory area mean you can only load fewer TSRs and, if you don't have a contiguous area large enough for the page frame, can even prevent you from using EMS altogether.

Aside from audio (which is the main topic here), how much usable upper memory area is also an important aspect which I'm struggling to make a compromise, and this is increasingly becoming a problem with modern gaming-class motherboards which integrate a lot of 3rd-party controllers to enrich the functionalities beyond what the chipset itself could offer. If you don't really utilize the external controllers' option ROMs (such as booting from a disk attached to an onboard/offboard SATA port not provided by the chipset), you might be able to disable them in BIOS. This prevents them from being seen under DOS, but other operating systems (Windows, Linux, etc.) that have drivers for them may still utilize the controllers there, provided that you did not disable the controller entirely.

As for PS/2 ports... recent boards either provide one, or none. For those with 1 PS/2, it's often advised to use it for keyboard (which permits NKRO if your keyboard supports such). PS/2 port splitters exist, but these motherboards are unlikely to support them properly. USB legacy mode usually still works, but modern keyboards and mice are very likely to break support against it in order to enable more capabilities.

By the way, vanilla Win7 install medium doesn't always have native USB3 support. In the past I often find myself having to plug the installation medium to a USB2 port in order to proceed installing Win7, and install USB3-related drivers later on.

Reply 338 of 486, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

I dunno details about B75 and X79, but with X58 i made all tested 160 of 160 dos games to start, so all mem issues are solvable. With B75 it would be easier because UMBPCI would work for realmode, there are Bios profiles too, so you can disable unneeded devices for Dos.. its harder, but its possible free enough mem with my ultimate boot config..

PS/2 ports - Rayer modified his H67 - 1x PS/2 MB to 2x PS/2 and i he will try modify mine too, some tutorial is on his page.

Details could be find here:
X58/i865/V880 - Yamaha7x4/AurealV1/2 pure Dos7.1- compatibility list/research/ultim. drivers configs, WIP- gurus needed

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 339 of 486, by Srandista

User metadata
Rank Oldbie
Rank
Oldbie
ruthan wrote:

PS/2 ports - Rayer modified his H67 - 1x PS/2 MB to 2x PS/2 and i he will try modify mine too, some tutorial is on his page.

Why even care with HW modifying of the board, when you can just use PS/2 splitter from eBay for 1$? With it, I can use PS/2 keyboard and mouse even on my really modern Z270 board, no hard modification necessary.

Socket 775 - ASRock 4CoreDual-VSTA, Pentium E6500K, 4GB RAM, Radeon 9800XT, ESS Solo-1, Win 98/XP
Socket A - Chaintech CT-7AIA, AMD Athlon XP 2400+, 1GB RAM, Radeon 9600XT, ESS ES1869F, Win 98