VOGONS


Sound Blaster: From best to worst

Topic actions

Reply 100 of 156, by anakin94

User metadata
Rank Newbie
Rank
Newbie
Joseph_Joestar wrote on 2021-02-26, 15:05:

Doom, WarCraft 2 and Skyroads seem to be good test cases for the Vibra distortion bug.

Check here and also here for distortion examples documented by James-F.

I've got the Vibra 16XV CT4170.
It has DSP version 4.16.
For my test i used the 50 Hz WAVE file posted by @digistorm and i tested Doom and Skyroads.
When i maximize the mixer, i have hearable clipping.
At Doom it's very clear hearable.
It can also be seen, when loading my posted FLAC files in a audio editor like Audacity.
You can see that the waves are cutted and they looks like a brickwall.

EDIT: Added samples without music.

Attachments

  • Filename
    CT4170_Skyroads_No_Music.flac
    File size
    134.25 KiB
    Downloads
    32 downloads
    File license
    Fair use/fair dealing exception
  • Filename
    CT4170_Doom_No_Music.flac
    File size
    228.99 KiB
    Downloads
    31 downloads
    File license
    Fair use/fair dealing exception
  • Filename
    CT4170_Skyroads.flac
    File size
    192.94 KiB
    Downloads
    32 downloads
    File license
    Fair use/fair dealing exception
  • Filename
    CT4170_Doom.flac
    File size
    324.88 KiB
    Downloads
    30 downloads
    File license
    Fair use/fair dealing exception
  • Filename
    CT4170_50_Hz.flac
    File size
    858.4 KiB
    Downloads
    30 downloads
    File license
    Fair use/fair dealing exception
Last edited by anakin94 on 2021-03-07, 22:14. Edited 1 time in total.

Enhanced for Matrox Mystique

Reply 101 of 156, by Joseph_Joestar

User metadata
Rank l33t
Rank
l33t
anakin94 wrote on 2021-03-05, 23:27:
I've got the Vibra 16XV CT4170. It has DSP version 4.16. For my test i used the 50 Hz WAVE file posted by @digistorm and i teste […]
Show full quote

I've got the Vibra 16XV CT4170.
It has DSP version 4.16.
For my test i used the 50 Hz WAVE file posted by @digistorm and i tested Doom and Skyroads.
When i maximize the mixer, i have hearable clipping.
At Doom it's very clear hearable.

Yup, it is as we suspected. The clipping/distortion is not related to the DSP version, it's a shortcoming of the Vibra chip itself. On the positive side, with DSP 4.16, you will have fewer hanging note bugs. This matters if you decide to add an external MIDI device or a wavetable daughterboard at some point.

BTW, you might want to re-record those samples with the music turned off. That would make it easier to hear the clipping properly. You can either mute the music in the game's setup or using the Creative mixer.

PC#1: Pentium MMX 166 / Soyo SY-5BT / S3 Trio64V+ / Voodoo1 / OPTi 82C930 / AWE64 Gold / SC-155
PC#2: AthlonXP 1700+ / Abit KT7A / Voodoo3 / SBLive / Vortex2
PC#3: Athlon64 3000+ / Asus K8V-MX / GeForce4 / Audigy1

Reply 102 of 156, by digistorm

User metadata
Rank Member
Rank
Member

Yep, it is clearly visible that your sound is clipping. You might get away with setting up your DOS/Windows mixer with the wave / voice channel to a lower volume (50-60%) as that might prevent the clipping issue. It is possible that the digital sound part of the card comes in too loud in the analog or digital domain, because on my CT4520 card it is clearly out of balance and the digital sound is much louder than the FM / MIDI sound. Of course your sound setup must be compatible with a lower volume setting, in the case of the CT4520 card it is not a problem because most noise comes from the MIDI synthesizer so setting the volume levels lower does not introduce more noise than it already has in my case.

Reply 103 of 156, by Burrito78

User metadata
Rank Member
Rank
Member
anakin94 wrote on 2021-03-05, 23:27:
I've got the Vibra 16XV CT4170. It has DSP version 4.16. When i maximize the mixer, i have hearable clipping. At Doom it's very […]
Show full quote

I've got the Vibra 16XV CT4170.
It has DSP version 4.16.
When i maximize the mixer, i have hearable clipping.
At Doom it's very clear hearable.
It can also be seen, when loading my posted FLAC files in a audio editor like Audacity.
You can see that the waves are cutted and they looks like a brickwall.

Thanks! I will update the chart accordingly.

Sound Blaster: From best to worst
Member of DOSBox Staging

Reply 104 of 156, by anakin94

User metadata
Rank Newbie
Rank
Newbie
Joseph_Joestar wrote on 2021-03-06, 05:55:

Yup, it is as we suspected. The clipping/distortion is not related to the DSP version, it's a shortcoming of the Vibra chip itself. On the positive side, with DSP 4.16, you will have fewer hanging note bugs. This matters if you decide to add an external MIDI device or a wavetable daughterboard at some point.

BTW, you might want to re-record those samples with the music turned off. That would make it easier to hear the clipping properly. You can either mute the music in the game's setup or using the Creative mixer.

I have re-recorded them without music and added them to my previous post.

digistorm wrote on 2021-03-06, 09:52:

Yep, it is clearly visible that your sound is clipping. You might get away with setting up your DOS/Windows mixer with the wave / voice channel to a lower volume (50-60%) as that might prevent the clipping issue. It is possible that the digital sound part of the card comes in too loud in the analog or digital domain, because on my CT4520 card it is clearly out of balance and the digital sound is much louder than the FM / MIDI sound. Of course your sound setup must be compatible with a lower volume setting, in the case of the CT4520 card it is not a problem because most noise comes from the MIDI synthesizer so setting the volume levels lower does not introduce more noise than it already has in my case.

I'm very happy with the card.
It has no hanging note bug and not the dma clicking.
Btw. i use passive speakers and my mixer is mainly set up at 50%, exept the master volume that is ~10%.
So for me is the clipping not a problem. 😀
I like the CT4170 more than the CT2950.

Burrito78 wrote on 2021-03-07, 20:06:

Thanks! I will update the chart accordingly.

I have to thanks.
The list is very helpful, interesting and clear.

Enhanced for Matrox Mystique

Reply 105 of 156, by Burrito78

User metadata
Rank Member
Rank
Member

Added some "new" cards into the mix just for the fun on it!

Changelog:
List updated to v9 to include SB 1.0/1.5/Pro/Pro 2
Added information for CT4170 (thanks anakin94!)
Changed CT2760 DAC type to CT1701
Fixed typos
Updated title AGAIN - sorry!

Sound Blaster: From best to worst
Member of DOSBox Staging

Reply 106 of 156, by appiah4

User metadata
Rank l33t++
Rank
l33t++

I have an excel worksheet of this exact same chart where I keep ownership information for collecting purposes and I edited mine accordingly as well. Thank you for this, man. It's a great resource.

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

Reply 107 of 156, by pshipkov

User metadata
Rank Oldbie
Rank
Oldbie

Great info in this thread !

So it looks to me that from all sb16 versions the best are ct1780 and ct1790, assuming good dac and dsp versions.
We can also consider 1730 1740 1750 1770, if output volume is set low and passed to external amplifier, if one is bothered by self-noise.

From the awe32 ones ‐ ct3900 and ct3980 seem to have the least issues, again, considering good dac/dsp.

retro bits and bytes

Reply 108 of 156, by appiah4

User metadata
Rank l33t++
Rank
l33t++
pshipkov wrote on 2021-03-10, 08:01:
Great info in this thread ! […]
Show full quote

Great info in this thread !

So it looks to me that from all sb16 versions the best are ct1780 and ct1790, assuming good dac and dsp versions.
We can also consider 1730 1740 1750 1770, if output volume is set low and passed to external amplifier, if one is bothered by self-noise.

From the awe32 ones ‐ ct3900 and ct3980 seem to have the least issues, again, considering good dac/dsp.

Reducing the volume won't help with the noisein CT1730/40/50/70, you need to bypass the amplification circuit through modding. If you just lower the volume, then amplify it elsewhere then you will probably get even worse noise. I have a CT3980 and it is indeed an amazingly good card. I was lucky enough to also come across two sticks of 16MB 30-pin SIMMs for it that work. But finding larger than 1MB 30-pin stick is a big issue with this card.

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

Reply 109 of 156, by pshipkov

User metadata
Rank Oldbie
Rank
Oldbie

Agreed about tc3980. Great hardware.
As for the low volume + good external amplifier, seems to work really well actually.
Not just for these old audio cards only, but in general.

retro bits and bytes

Reply 110 of 156, by Cloudschatze

User metadata
Rank Oldbie
Rank
Oldbie
appiah4 wrote on 2021-03-10, 09:12:

Reducing the volume won't help with the noisein CT1730/40/50/70, you need to bypass the amplification circuit through modding.

There's no modding necessary - the OPSL and OPSR jumpers are used to enable/disable internal amplification on the earlier SB16s. With the amplifier bypassed, and using the following mixer settings as a starting point, you'll have a respectable and relatively noise-free card:

/MA:240;240;50
/VO:240;240;50
/MI:240;240;50
/CD:0;0;50
/LI:0;0;50
/MIC:0
/SP:0
/TR:128;128;50
/BA:128;128;50
/IPL:MIC- CDR- LIR- LIL- MIR- MIL-
/IPR:MIC- CDR- LIR- LIL- MIR- MIL-
/OPS:MIC- CDR- CDL- LIR-
/AGC:-
/IPG:1,1
/OPG:1,1
/SE:- (where applicable)

The following position of the OPSL/OPSR jumpers reflects a bypassed internal amplifier (line-level output):

sb16_ops.jpg

As an easy test means, if the physical volume wheel affects the output volume, the internal amplifier is *not* being bypassed.

Last edited by Cloudschatze on 2021-03-10, 17:10. Edited 1 time in total.

Reply 112 of 156, by carlostex

User metadata
Rank l33t
Rank
l33t

@Cloudschatze

That's the most beautiful sound card close up picture that i've ever seen! The card looks absolutely pristine, and it seems it just came out of the production line. 😀

Reply 113 of 156, by pshipkov

User metadata
Rank Oldbie
Rank
Oldbie

So, based on simple solution that @Cloudschatze suggested, does it make sense to change the status in the "self-noise" field for the CT1730/40/50/70 ?
With an added node perhaps ?

retro bits and bytes

Reply 114 of 156, by Burrito78

User metadata
Rank Member
Rank
Member
pshipkov wrote on 2021-03-11, 20:58:

So, based on simple solution that @Cloudschatze suggested, does it make sense to change the status in the "self-noise" field for the CT1730/40/50/70 ?
With an added node perhaps ?

Big thanks to Cloudschatze for posting these tips for noise optimization. While i personally trust Cloudschatzes findings i'd feel better to include this information if i had a second source that disabling the Amplifier significantly reduces noise.

Can someone else check this with their card and report back? Or post a link to a resource online? I couldn't find anything substantial myself.

Maybe there is more info regarding noise in Philscomputerlab's video about the CT1740? Can't check it right now.

Sound Blaster: From best to worst
Member of DOSBox Staging

Reply 115 of 156, by pshipkov

User metadata
Rank Oldbie
Rank
Oldbie

I can confirm that his suggestion works. Tried it on 2 "noisy" CT17##'ies.

My original point about lowering output volume and passing to any decent amplifier works just fine as well.

retro bits and bytes

Reply 117 of 156, by Burrito78

User metadata
Rank Member
Rank
Member
pshipkov wrote on 2021-03-12, 19:31:

Tested a bit more.
One of the cases - 1750 with turned-off amplifier has the same noise levels as 2230, which passes as "low-noise" implementation.

Thanks! It's on my list for the next revision! Good work!

Sound Blaster: From best to worst
Member of DOSBox Staging

Reply 118 of 156, by zoinknoise

User metadata
Rank Newbie
Rank
Newbie

nice work on the latest update. i really have to say though, i am uneasy with the "Self Noise" column and how it's purely based on the DAC. it's uncomfortably close to those endless arguments in audiophile forums comparing two CD players, one with 64x oversampling DACs, and the other with 128x oversampling DACs, and how the latter has to sound far better. (in reality of course, nobody can actually tell the difference under blind, controlled ABX testing.)

i own many Sound Blasters, including a CT2760, CT3900 and CT3980. i cannot honestly say that the CT2760 has "high" self noise, compared to a CT3900 or CT3980, just because it has a slightly older DAC. i just think it's a misleading way to put it. the DAC is one of the least important factors for determining noise on an old sound card, just look at this thread. these cards are full of strange DSP issues, mixer issues, clipping, etc... all of these cause much more noise and distortion than the DAC. the PCB design of a sound card has far more impact on noise than the DAC. noise can leech into the card from a busy ISA bus with lots of cards, and the DAC won't help you at all there.

quite frankly, i would argue the #1 source of self-noise on an old sound card is the high number of old, dried-out electrolytic capacitors. i learned that lesson when i replaced all the electrolytic capacitors on my CT1350B. it instantly went from being borderline unusable, to being a decent sounding card. that showed me where the noise is really coming from on these old cards, not the DAC.

i would like to suggest that the "Self Noise" column be retitled "DAC," and it should simply list the model of the DAC. that way, if someone believes that the DAC on their card affects the noise floor, they will already know which part numbers to look for, and which to avoid. plus, the chart will be providing more good information, because right now it doesn't actually list the DACs being used on each card, besides CT1703.

Reply 119 of 156, by perhenden

User metadata
Rank Member
Rank
Member

I found a model not mentioned in the list, and thought I would contribute.
The CT3630. I've just seen a picture of it (attached to this post). Hopefully chips and versions can be identified from this single image.

Thanks! Great list.

Attachments

Last edited by perhenden on 2021-03-22, 09:25. Edited 1 time in total.