VOGONS


Reply 40 of 48, by MattRocks

User metadata
Rank Member
Rank
Member
leileilol wrote on 2025-11-26, 00:52:

Both of those games mixed at 22khz. It was the Windows Sound System and AC'97 pushing for 48khz, and ES1371 was focused on AC'97 compliance.

Partly right. You may be right about the default mixer configuration in those game engines - I thought they were different, but I might be remembering changes that came with various mods.

Doesn't matter because 22kHz output from both games is even better for ES1370. Below is a screenshot is from Ensoniq's technical data (before Creative) and it shows ES1370 will clock to 22kHz natively. If the ES1370 driver finds a match between 22kHz PCM output from the game, and 22kHz PCM input to the hardware, then the driver will not resample the stream. What really matters is that the software driver tells Microsoft DirectSound that the sound card will ingest 22kHz* PCM audio.

Screenshot-2025-11-26-at-10-14-22.png

AC'97 was an Intel standard, not a Microsoft standard. AC'97 aligns to the DVD audio standard, not something Intel made up. Intel pushed AC'97 to simplify its motherboards. ES1370 is a first generation PCI sound card that predates AC'97. Furthermore, ES1370 flatly rejects 48kHz - its hardware simply cannot clock to that rate. That constraint is in hardware, aligned to the CD audio standard (CD audio and DVD audio operate at different rates), and all of this has nothing to do with Windows Sound System pushing anything.

Summary: Q3/UT produce 22kHz output, Windows asks the sound card if it can open a 22kHz PCM stream, sound card says yes and the FIFO pump does the rest with zero CPU overhead.

You mentioned ES1371, which is a Creative Labs variant created after Ensoniq was destroyed. ES1371 did support 48kHz (probably it was AC'97 compliant and probably works nicely with DVD audio) - but it flatly rejected 22kHz PCM streams: UT/Q3 produce 22kHz output, the ES1371 says "no", software resamples it to 48kHz, ES1371 says "yes" but by then you're down ~5FPS. ES1371 was not tuned for those games. ES1371 was, in my opinion, tuned to sell more SB Live cards!

Conclusion: ES1370 is not ES1371 - but to doubly check you should read the frequency of the oscillating crystal because that is the component that is dictating the physical hardware constraint.

*22kHz needs to be specifically 22.05kHz. 22.05kHz is exactly half of 44.1kHz. 44.1kHz is the standard for CD ROM music / redbook audio.

Reply 41 of 48, by MattRocks

User metadata
Rank Member
Rank
Member
NeoG_ wrote on 2025-11-26, 03:16:

I'm using a CT4830 in a super 7 system and it sounds and runs fine, it was easy to install the Audigy drivers using documented methods and easy driver packs. Plus it doubles as a useful stereo mixer in DOS using the ek1m tool.

I'm not experiencing any crackling or strange sound issues, EAX and 4PS works whenever I turn it on in games. I think the sample rate conversion issue is overblown, it's not optimal but since it's used as a playback device for the vast majority of use cases you can't even tell.

It's a busy system too, 4 IDE devices, AGP Graphics, 3 PCI cards, 3 ISA cards, SCSI and USB

What is the motherboard chipset?

Reply 42 of 48, by NeoG_

User metadata
Rank Oldbie
Rank
Oldbie
MattRocks wrote on 2025-11-26, 12:13:

What is the motherboard chipset?

ALi ALADDiNV (M1541/M1543) on an Acorp 5ALI61 board.

The board has damaged capacitors, a broken CPU socket clip and a burned out USB header (replaced with an Opti PCI card) but it’s a survivor.

98/DOS Rig: BabyAT AladdinV, K6-2+/550, V3 2000, 128MB PC100, 20GB HDD, 128GB SD2IDE, SB Live!, SB16-SCSI, PicoGUS, WP32 McCake, iNFRA CD, ZIP100
XP Rig: Lian Li PC-10 ATX, Gigabyte X38-DQ6, Core2Duo E6850, ATi HD5870, 2GB DDR2, 2TB HDD, X-Fi XtremeGamer

Reply 43 of 48, by MattRocks

User metadata
Rank Member
Rank
Member
NeoG_ wrote on 2025-11-26, 12:42:

ALi ALADDiNV (M1541/M1543) on an Acorp 5ALI61 board.

The board has damaged capacitors, a broken CPU socket clip and a burned out USB header (replaced with an Opti PCI card) but it’s a survivor.

At this point the only upgrade you’re missing is an off-cut of MacGyver’s second-hand duck tape! Now I'm even more tormented by knowing I have an ALi in a PC just a few hundred miles away.. probably in mint condition, if the caps haven't dried out.

Reply 44 of 48, by swaaye

User metadata
Rank l33t++
Rank
l33t++

ES1370 will play non-native sample rates with varying degrees of success. It should be out of pitch if Windows kmixer doesn't resample before sending to the card. VXD and WDM driver behavior varies. I'm not sure what the DOS driver does.

Reply 45 of 48, by MattRocks

User metadata
Rank Member
Rank
Member
swaaye wrote on 2025-11-26, 18:10:

ES1370 will play non-native sample rates with varying degrees of success. It should be out of pitch if Windows kmixer doesn't resample before sending to the card. VXD and WDM driver behavior varies. I'm not sure what the DOS driver does.

It will, but that's not the point.

A3D, EAX, SRS, EQ, etc. - all of that can all be done in software, and it can be done "better" in software because the algorithms can be fine tuned, run on multi-core multi-GHz CPUS, and even run as a pre-process where rendering time is immaterial.

And if you find software resampling and mixing acceptable, which has been sold mainstream since 2006, then there is literally no point in having a hardware sound processor in your PC because whatever that processor is - it simply doesn't do anything that you value.

Last edited by MattRocks on 2025-11-26, 21:36. Edited 1 time in total.

Reply 46 of 48, by swaaye

User metadata
Rank l33t++
Rank
l33t++

I did not voice an opinion about software or hardware processing in that post. But yes I agree that software processing is better. PCI audio cards with hardware acceleration have forever been fraught with problems related to bus transfers to and from the audio ASIC, and with things like reduced complexity resampling algorithms.

Last edited by swaaye on 2025-11-26, 21:37. Edited 1 time in total.

Reply 47 of 48, by MattRocks

User metadata
Rank Member
Rank
Member
swaaye wrote on 2025-11-26, 21:35:

I did not voice an opinion about software or hardware in that post.

You didn't score it but you did present Windows kmixer as a solution 😉

Reply 48 of 48, by MattRocks

User metadata
Rank Member
Rank
Member
swaaye wrote on 2025-11-26, 21:35:

PCI audio cards with hardware acceleration have forever been fraught with problems related to bus transfers to and from the audio ASIC, and with things like reduced complexity resampling algorithms.

That bus transfer is the one thing Ensoniq got right. Ensoniq had a FIFO for audio that you could directly compare to the 3Dfx FIFO for video. My opinion is that each time a sound card company got something right, Creative acquired and destroyed it - whether that was a recurring coincidence or an intentional strategy is for someone else to conclude 😉

3Dfx also provided reduced complexity algorithms - not a good choice for rendering scientific 3D models, but perfectly okay for games. This is actually very similar to the divide we can prove existed between professional audio cards for musicians, and the sound cards that deliver an ultra-low CPU hit in games.

Another way video and audio are similar: Some games had detailed 3D software rendering that looked better than 3Dfx’s shortcut render paths - Motorhead is one.