The "12-bit" Sound Blaster 16 Myth

Discussion about old sound cards, MIDI devices and sound related accessories.

Re: The "12-bit" Sound Blaster 16 Myth

Postby krcroft » 2019-5-10 @ 23:00

Scali wrote:Is that even a thing in the audio world?


It should or certainly could be; it's just deconvolution in 1D space. A pertinent example would be recordings captured from these early PC audio cards, which apparently were quite noisy. Maybe a progressive 90s wedding DJ recorded his PC-mixed gig from his SB16's line out :-)

If we assume that that the noise profile for a given model of audio card is roughly the same (as opposed to each card having a unique profile), then we can model or measure that noise using a working model of that card today, and then 1D-deconvolve out that noise in those old recordings. You can do one better if you still have access to the /exact/ card used in the first place, and thus can measure its exact noise profile (of course.. using equipment that is significantly less noisy and can sample at nyquist rates).

The cost and effort are non-trivial; but I can image there are PhD's floating around in National Archive houses that do this type of tinkering to squeeze out more fidelity from historical presidential speeches or concerts / musicals / etc... Also plenty of stories of 'master recordings' of TV shows being lost, and only having noisier derived copies remaining. As for if commercial music remasters use some of these techniques; I'll have to let those scientific audiophiles weigh in.
User avatar
krcroft
Member
 
Posts: 314
Joined: 2017-4-29 @ 15:07
Location: Ogden's Retreat

Re: The "12-bit" Sound Blaster 16 Myth

Postby yawetaG » 2019-5-11 @ 07:37

Scali wrote:
krcroft wrote:The same way Starbucks didn't break new ground with their urban franchised "fast-coffee" business model in the 70s and 80s, because early middle eastern shops predate them by 500 years - and their coffee is still vastly superior? I see both points.


That is not an analogy to what I was talking about.
PC sound cards generally used off-the-shelf components, which had also been used in various other computers already. Hardly groundbreaking in any way, more like catching up.
And as I said, selling them at higher prices to boot.


Are you suggesting they should have used professional-grade audio components? I sure hope not, because then the prices would have been even higher (up several thousands of dollars per card!), if you go by the prices most music modules sold for in that period.
Also, consider that back in the day PC gaming might have been established up to a degree, but most people would still use PCs for word processing and the like, so sound cards still were somewhat niche. The multimedia experience was just becoming a thing. If you wanted video games back in 1988-1992 you'd buy a game console...

In my experience (in Europe!), the real PC gaming boom only came in the mid to late 1990s, with games and other software becoming available outside specialist shops. Which also pretty much coincides with the greater availability of relatively cheap General MIDI sound modules...
yawetaG
Oldbie
 
Posts: 1451
Joined: 2016-6-14 @ 14:36
Location: Netherlands

Re: The "12-bit" Sound Blaster 16 Myth

Postby Scali » 2019-5-11 @ 08:04

krcroft wrote:If we assume that that the noise profile for a given model of audio card is roughly the same (as opposed to each card having a unique profile), then we can model or measure that noise using a working model of that card today, and then 1D-deconvolve out that noise in those old recordings.


I hate to disappoint you, but the noise is non-deterministic. It's what some people describe as 'computer thinking noises'.
Basically the signals that go over the ISA bus, bleed into the analog circuit of the sound card, so you actually hear your HDD spinning, and your CPU pumping data to your video card etc.
There's no real 'pattern' in there.
The sad part is that this is mainly a PC problem: cheap/poorly designed hardware circuits, poorly shielded from noise.
Well-designed cards, such as the PAS16, do not have this problem, which proves that it is not inherent to the PC with its ISA architecture, as some would claim.
Scali
l33t
 
Posts: 4351
Joined: 2014-12-13 @ 14:24

Re: The "12-bit" Sound Blaster 16 Myth

Postby Scali » 2019-5-11 @ 08:10

yawetaG wrote:Are you suggesting they should have used professional-grade audio components?


Not sure where you got that idea from, but no.
I mean that other computers used the same (or at least similar) chips, but offered their solution at lower prices, and often with higher quality. Which is just the result of skilled designers. Put expensive components in a poor design, and they will still be noisy (OPL3 being a case-in-point: it may be a low-end FM solution, but the OPL3 itself is entirely digital, so not susceptible to any noise whatsoever, and the DAC that Yamaha supplies is very good quality as well. On a good soundcard, OPL3 is very crisp and noise-free. On a bad soundcard, you get 'thinking noises').

yawetaG wrote:Which also pretty much coincides with the greater availability of relatively cheap General MIDI sound modules...


I don't think there's a direct relation between PC gaming and cheap General MIDI modules.
General MIDI was never a big thing in PC gaming, the only reasonably successful MIDI standards were MT-32 and Sound Canvas, which weren't in the cheap-category.
Cheap General MIDI devices were a result of advancements in technology, combined with the fact that they could be applied in many devices.
WaveBlasters and similar stuff were niche, never really caught on. CD-ROM technology and fast CPUs (software mixing/softsynth) made them obsolete.
Scali
l33t
 
Posts: 4351
Joined: 2014-12-13 @ 14:24

Re: The "12-bit" Sound Blaster 16 Myth

Postby NewRisingSun » 2019-5-11 @ 08:38

I have only two 16-bit sound cards from Creative: a early-ish (DSP version 4.05 instead of 4.04) Sound Blaster 16 "ASP" (CT1740), and a late-edition Sound Blaster AWE 64 Gold (CT4540).

I was always very dissatisfied with the CT1740's DOS game sound quality for two reasons:
  1. audible clicks during blocks of single-cycle DMA playback, much more noticeable than on earlier cards I had, such as the Sound Blaster Pro 2;
  2. a constant loud hiss at low sampling rates (such as the 11,025 Hz used by DOS games) whose frequency seems to be a function of the sampling rate, in both 8-bit and 16-bit modes.
Neither problem is mitigated by disabling the SB16's amplifier, and neither problem occurs on the CT4540, so they must have been problems that Creative was aware of and that they thus decided and managed to correct. The second problem is the one that must have been described by Creative in the SB16CQA.TXT file thusly:
Code: Select all
4. GENERAL

   4.1 ELIMINATING UNWANTED NOISES
   Q2. I just upgraded from SBPro to the SB16 and now some of my games
       sound hissy.  Do I need to return the SB16 for replacement?
                                       
   A2. The high quality 16-bit CODEC plays back every detail of the
       8-bit sample including the coarse resolution of the staircase
       waveform.  Hearing the "hissing sound" is like seeing all the
       sharp edges of a 320x240 picture on a very sharp 1280x1024
       monitor.  SB16 has a Treble control to suit your needs on
       different sound quality.  To eliminate the unwanted "hissing
       sound", you can set the Treble level to zero.
That answer is utter nonsense. Playing back an 8-bit sample on a 16-bit DAC does not add hiss. It does not even add quantization noise, as an 8-bit DAC cannot somehow smooth the "coarse resolution of the staircase waveform". And the hiss occurs during 16-bit playback as well. The "sharp edges" equivalent, when applied to sound, would be aliasing above the Nyquist rate of the original file, which is unrelated to the hissing sound.

I can record and post samples of the CT1740's output demonstrating the two problems I mentioned, if desired. I never had any complaints about the CT1740's performance at 44,100 Hz in 16-bit, though I have not taken any measurements.
NewRisingSun
Oldbie
 
Posts: 852
Joined: 2005-9-02 @ 02:26

Re: The "12-bit" Sound Blaster 16 Myth

Postby Scali » 2019-5-11 @ 08:48

NewRisingSun wrote:as an 8-bit DAC cannot somehow smooth the "coarse resolution of the staircase waveform".


Well, it can, actually. That is, the transition is not 'instant' in the analog domain, and can be filtered to make it smoother. You should try playing an 8-bit sample and recording it with a high-resolution ADC. You'll see that it's not pure stair-stepping. It's somewhat similar to the concept of PWM, where the signal itself is only 1-bit, but by introducing a bit of 'latency', you transform the time-component into a 'dynamic range' component. This can be done with a simple capacitor in the circuit, or even just by the inertia of the speaker cone itself.

But your point remains: if an 8-bit DAC can do that, then so can a 16-bit DAC (in fact, the better the DAC, the more time is probably spent on this, with oversampling and reconstruction filters, in the digital domain).
The whole assumption of analog-to-digital and digital-to-analog conversion is that both the input and the output are (or should be) continuous signals.
The analog sound is passed through a filter to remove any signals above the Nyquist frequency, so that any potential aliasing is removed before it could enter the ADC. As a result, there should be none of this aliasing in the output signal, so another filter will remove that after the DAC.
Scali
l33t
 
Posts: 4351
Joined: 2014-12-13 @ 14:24

Re: The "12-bit" Sound Blaster 16 Myth

Postby NewRisingSun » 2019-5-11 @ 09:07

More importantly, the hiss I described is unrelated to the actual signal, as it occurs even if the DSP is playing back a signal containing only $80 samples.
NewRisingSun
Oldbie
 
Posts: 852
Joined: 2005-9-02 @ 02:26

Re: The "12-bit" Sound Blaster 16 Myth

Postby yawetaG » 2019-5-11 @ 12:32

Scali wrote:
yawetaG wrote:Are you suggesting they should have used professional-grade audio components?


Not sure where you got that idea from, but no.
I mean that other computers used the same (or at least similar) chips, but offered their solution at lower prices, and often with higher quality. Which is just the result of skilled designers. Put expensive components in a poor design, and they will still be noisy (OPL3 being a case-in-point: it may be a low-end FM solution, but the OPL3 itself is entirely digital, so not susceptible to any noise whatsoever, and the DAC that Yamaha supplies is very good quality as well. On a good soundcard, OPL3 is very crisp and noise-free. On a bad soundcard, you get 'thinking noises').


On Yamaha's own PSR keyboards you also get more noise than on their professional products. I wouldn't call them "bad", the target public is just different and likely cares less about noise, so Yamaha didn't bother with using the same high standards as they use for their professional electronic instruments. It's all economics.
Creative marketed their products towards the normal PC owner*, while other manufacturers aimed at a more limited public with higher expectations. The latter translated to things like smaller production runs, confidential distribution (only available at PC enthusiast shops), and higher prices.
Also, which other computers do you mean? Macs were still quite expensive back then, Amigas and Ataris weren't very popular, NEC PC-98 etc. had a market mostly limited to Japan (and weren't cheap at all), Amstrads were becoming ancient (and non-compatible) by then...

* Their attempts at getting into the professional market mostly resulted in killing off some interesting pro-audio manufacturers after Creative bought them. :depressed:

yawetaG wrote:Which also pretty much coincides with the greater availability of relatively cheap General MIDI sound modules...


I don't think there's a direct relation between PC gaming and cheap General MIDI modules.
General MIDI was never a big thing in PC gaming, the only reasonably successful MIDI standards were MT-32 and Sound Canvas, which weren't in the cheap-category.[/quote]

Sound Canvas is General MIDI. Sound cards with wavetables also use General MIDI. :confused: If it wasn't a big thing in PC gaming, why are people on this forum and elsewhere raving about wavetables?

And by "cheap" I did not mean "50 dollar"-cheap, I meant "cheap compared to a professional-grade sound module".

Cheap General MIDI devices were a result of advancements in technology, combined with the fact that they could be applied in many devices.
WaveBlasters and similar stuff were niche, never really caught on. CD-ROM technology and fast CPUs (software mixing/softsynth) made them obsolete.


I think this really depends on who and where you ask. At least in my corner of the woods the reason sound cards like the AWE32 (and even the Live, later on) were really hyped was that you didn't need a daughter board to be able to get good sound.

Most early softsynths really suck when it comes to features, compared to a real module**. The ones that didn't back in the day required massively powerful systems to run, e.g. Creative/Emu's Proteus softsynth was too demanding for all but the most high-end systems back when it was first released, making playing more than a single audio track at a time impossible.

** By which I don't mean a Sound Canvas. They are truly very limited as a synthesizer.
yawetaG
Oldbie
 
Posts: 1451
Joined: 2016-6-14 @ 14:36
Location: Netherlands

Re: The "12-bit" Sound Blaster 16 Myth

Postby rasz_pl » 2019-5-11 @ 12:47

NewRisingSun wrote:[code]
A2. The high quality 16-bit CODEC plays back every detail of the
8-bit sample including the coarse resolution of the staircase
waveform.


used car salesman snakeoil. Neo, there are no stair steps.

https://www.youtube.com/watch?v=cIQ9IXSUzuM
https://www.youtube.com/watch?v=pWjdWCePgvA
rasz_pl
Member
 
Posts: 412
Joined: 2017-6-04 @ 00:57

Re: The "12-bit" Sound Blaster 16 Myth

Postby Scali » 2019-5-11 @ 14:47

yawetaG wrote:Creative marketed their products towards the normal PC owner*, while other manufacturers aimed at a more limited public with higher expectations. The latter translated to things like smaller production runs, confidential distribution (only available at PC enthusiast shops), and higher prices.


The flaw in this reasoning is that other brands used basically the same components, at basically the same price, yet got considerably better results.
The most obvious example is the PAS16 vs the SB16.

yawetaG wrote:Also, which other computers do you mean? Macs were still quite expensive back then, Amigas and Ataris weren't very popular, NEC PC-98 etc. had a market mostly limited to Japan (and weren't cheap at all), Amstrads were becoming ancient (and non-compatible) by then...


You are confusing things here, moving goalposts etc (aside from the fact that it's highly debatable whether or not Atari STs and Amigas were popular. I'd say they were, at least in the gaming market, seeing as most games were developed for these machines, and PCs got quick-and-dirty ports of these games).
Thing is, the context here was the quality of the sound circuit on various computers of that era. Their popularity is not relevant here.
However, as I already mentioned earlier, I literally paid exactly the same for an SB Pro 2 card, as I did for my entire Amiga 600, both at about the same time.
The Amiga has very crisp and noise-free sound, and excellent stereo separation. The SB Pro 2 does not. There's considerable hiss at all times, worse than cassette tape even. And there's the 'computer thinking noise' on top of that. And let's not get into the stereo DAC. If you play a hard panned sound to one side, you can clearly hear it in the other speaker as well. The Amiga doesn't have any of that. There's no audible hiss, even when you have your amp quite far open. And given the fact that the *entire machine* was the same price as just the SB Pro 2 card alone, it's obvious that the sound circuit in the Amiga couldn't have been built from more expensive components than the SB Pro 2.

The SB Pro 2 is so noisy that I'd compare it with my Commodore 64. In fact, I actually went back to the store because I thought mine was broken. Then they let me listen to the test model in the shop, and it was the same. It's just how these cards are.

yawetaG wrote:Sound Canvas is General MIDI.


But not *cheap* General MIDI.

yawetaG wrote:Sound cards with wavetables also use General MIDI.


Not necessarily. The GUS and AWE32 may be wavetable sound cards, but they are not General MIDI-compatible.
Conversely, MIDI synthesizers are not necessarily wavetable-based.
So again, you are confusing many things, and I have no idea what you're actually trying to say.

yawetaG wrote:And by "cheap" I did not mean "50 dollar"-cheap, I meant "cheap compared to a professional-grade sound module".


Which makes no sense whatsoever, because although the MT-32 and Sound Canvas may be considered 'cheap' in the professional music world, they were the most expensive audio solutions available for gaming by far (where the SB Pro 2 cost the same as an entire Amiga, an MT-32 or Sound Canvas with MPU-401 could buy you 2 or 3 Amigas).
Scali
l33t
 
Posts: 4351
Joined: 2014-12-13 @ 14:24

Re: The "12-bit" Sound Blaster 16 Myth

Postby yawetaG » 2019-5-11 @ 18:43

Perhaps you could provide some references to prices back in the day to back up your points. By preference not only from the USA.

Let me provide some numbers:

Looking at my Yamaha MU10 box (Japanese), the original price in 1996 was 19,800 yen (not corrected for inflation, about 200 bucks), which is not excessive and certainly not excessively more than a sound card.
AFAIK, the Yamaha TG-300 was around 900 bucks.

Looking at the old Escom catalogue from 1994 I have here shows:
- a Soundblaster 2.0/CD-ROM drive pack for 399 NLG (Dutch Guilder)
- ESCOM Powersound 16+ for 249 NLG,
- Raysound Mozart 16 for 249,
- Mediavision Pro Sonic 16 for 199,
- SB 2.0 Value for 119,
- SB16 Value for 279,
- SB AWE 32 for 789,
- Power Media 32 Super Wave for 499,
- an ESCOM sound kit (sound card/CD-ROM/speakers) for 549,
- Creative SB16 with CD-ROM and 6 games for 1199,
- SB 2.0/CD-ROM/speaker kit for 449,
- Mediavision Kit (sound card, SCSI CD-ROM, speakers, games) for 899.
To get prices in Euro's (not corrected for inflation), divide by 2.2.

So, um, I don't see the "Creative is overpriced"-thing here. :neutral: I also don't really see how GM modules are excessively more expensive than the more expensive sound card options that you'd need to have to get the same sound quality.

Furthermore, I'd personally like to see where, precisely, you could find a complete brand-new Amiga for those prices. :-/

(You can also claim as much as you want that economic considerations are not important, but it is highly unrealistic to do so. That Creative managed to pretty much win the marketing game is enough evidence - they still exist, most other companies don't)
yawetaG
Oldbie
 
Posts: 1451
Joined: 2016-6-14 @ 14:36
Location: Netherlands

Re: The "12-bit" Sound Blaster 16 Myth

Postby Scali » 2019-5-11 @ 19:46

This is pretty pointless. Do your own homework.
Scali
l33t
 
Posts: 4351
Joined: 2014-12-13 @ 14:24

Re: The "12-bit" Sound Blaster 16 Myth

Postby rasz_pl » 2019-5-12 @ 02:48

yawetaG wrote:Looking at my Yamaha MU10 box (Japanese), the original price in 1996 was 19,800 yen (not corrected for inflation, about 200 bucks), which is not excessive


which is 2x the price of CD-ROM, by 1996 most games switched to providing cdaudio

yawetaG wrote:Looking at the old Escom catalogue from 1994 I have here shows:
So, um, I don't see the "Creative is overpriced"-thing here

have you looked at the prices? SB16 is the most expensive 16 bit card

yawetaG wrote:That Creative managed to pretty much win the marketing game is enough evidence


evidence of what? Creative had >200% margin on their cards, they won the market by being first to deliver digital sound product and due to aggressive marketing - plastering all press with full page ads, paid by ridiculous margins on low quality products.
rasz_pl
Member
 
Posts: 412
Joined: 2017-6-04 @ 00:57

Re: The "12-bit" Sound Blaster 16 Myth

Postby krcroft » 2019-5-12 @ 04:51

Scali,
Regarding noise - I remember the thinking sounds; like faint chatter in the line when the HD was grinding away, or subtle higher pitched tones when the PCI BUS or even CPU was active. I remember solving it in my wife's computer by replacing the power supply and regrounding the motherboard (this was in ~2001 though when this issue had mostly gone away); but it's certainly due to poor engineering of the BUS and/or audio cards. None the less - those tones can be found and corrected away using DSP filters.

All measuring systems have error - deterministic and stochastic; The hiss you and NewRisingSun mentioned sounds largery deterministic and can certainly be filtered and very likely deconvolved (recall old audio equipment had a 60Hz / 50Hz switch to reverse-out the AC buzz from mains power). When you say the PAS16 is silent, their superior engineering managed to bring both components below the threshold of hearing. Error still exists, but that's a job-complete in my book given an audience of humans not oscilloscopes.

Pure-digital systems behave "perfectly" per their design, but because we haven't created infinite frequency response systems, there will still be imperfections (although no doubt way beyond what people can hear or ABX-test), which perhaps can be thought of as 100% deterministic error (or artifacts) due to design limitations. You can witness this today with pure-software resamplers (and hardware resamplers), which are still evolving. http://src.infinitewave.ca/.

Talking about signal to noise - I thought this thread was all about the 12-bit Sound Blaster 16; maybe discussions about price / value / market share / etc.. is decreasing the SnR :-). None the less I looked back at my DOS-PC hey-day: 1993 PC Magazine Buyer's Guide Edition, one of my favorites where I enjoyed the dense PC ads as much as the content. It's tough to find ads that even mention including a sound card! Pure business use back when Corel, Lotus, etc.. packages were $500 - $800. Dell offered a multimedia upgrade with SB16 + external speakers for $199 USD, which is roughly a 10% bump in price for the PCs on offer (not too bad). By this point, PCs had left other systems in the dust in terms of performance.
User avatar
krcroft
Member
 
Posts: 314
Joined: 2017-4-29 @ 15:07
Location: Ogden's Retreat

Re: The "12-bit" Sound Blaster 16 Myth

Postby Scali » 2019-5-12 @ 07:58

krcroft wrote:None the less - those tones can be found and corrected away using DSP filters.


Not sure what your point is though.
Earlier you were talking about remastering CDs recorded in the 80s and 90s. Where I assume you were talking about commercial CDs. Which do not suffer from this problem, since they were not recorded with a cheap PC soundcard.

You could use DSP technology today to suppress this noise, but again, what's the point? DSP technology would not be adequate or affordable back in the early 90s when these cards were around... And why would you bother to use them today? If you're still using this old hardware, wouldn't you want to experience its characteristics? If you want a silent system, just pick one of the good soundcards of the era, or use a newer one.

I mean, I recorded Crystal Dream from my SB Pro 2, warts and all, so people can experience just how lo-fi, noisy and poorly stereo-separated it sounds (as opposed to DOSBox and other emulators, which emulate a 'perfect' audio device): https://youtu.be/Qhym3zCa7Os
That was the 'state of the art' of hardware when that demo came out.
A total laugh compared to how the same music sounded on an Amiga, which sounds like this: https://youtu.be/3W6zVP-tdGU

krcroft wrote:Talking about signal to noise - I thought this thread was all about the 12-bit Sound Blaster 16; maybe discussions about price / value / market share / etc.. is decreasing the SnR :-).


Well, my point was, and remains, that it is plausible that some people called the SB16 a '12-bit' card, because it suffered from pretty poor SNR in general, and the 'computer thinking noises' to boot.
This was an issue not inherent with PCs or computers in general, since other sound cards and other computers of the era could reach significantly better SNR using similar components at similar prices.
Last edited by Scali on 2019-5-12 @ 18:25, edited 1 time in total.
Scali
l33t
 
Posts: 4351
Joined: 2014-12-13 @ 14:24

Re: The "12-bit" Sound Blaster 16 Myth

Postby Tiido » 2019-5-12 @ 14:09

Noise reduction via methods described earlier work really really well. You can capture this thinking noise and all associated noises to create a noise profile and these noise removal methods can then very effectively get rid of such noises without any signifcant harm to the material you want preserved. I've used them to remove traffic noise and many other things out of voice recordings and whatnot though in really bad cases it kind of starts sounding like a low bit rate MP3. Nothing is perfect but you can achieve amazing results.
User avatar
Tiido
Oldbie
 
Posts: 1053
Joined: 2018-1-14 @ 04:40
Location: Estonia

Re: The "12-bit" Sound Blaster 16 Myth

Postby SirNickity » 2019-5-13 @ 06:17

TBH, I find that "corrected" audio more offensive than the original, in most cases. The problem with the kinds of noise that people generally want gone -- hiss, HVAC, etc. -- is that it's mostly random and broad-spectrum. Hiss is normally caused by things like Johnson noise, which is actually useful to repurpose into the entropy for a random number generator. So by definition fairly unpredictable.

I can't recall ever hearing digital noise (bus activity, etc.) on any of my Creative cards. I did hear it on some cheap motherboard (and laptop) sound chips in the early 2000s. My Sound Blasters were just a little bit hissier than my HiFi audio components. If I had my stereo cranked and flipped over to the PC audio input, I would hear the noise floor. Meh. At that point, it was generally loud enough that fidelity wasn't my goal at the moment.

I don't have any super high-class audio interfaces, but I do have a Presonus Digimax FS rack. I guess I should make an evening out of it, and play back some sample clips on all my cards and record the difference on a card-by-card basis. I just never noticed anything being out of sorts back when I was using these as my daily driver rig. Could be rose tinted glasses I suppose.

rasz_pl wrote:
NewRisingSun wrote:A2. The high quality 16-bit CODEC plays back every detail of the 8-bit sample including the coarse resolution of the staircase waveform.

used car salesman snakeoil. Neo, there are no stair steps.

https://www.youtube.com/watch?v=cIQ9IXSUzuM
https://www.youtube.com/watch?v=pWjdWCePgvA


I admit up-front I haven't followed these links yet. But I'm going to assume this leads to the Xiph guy. This is one of those cases where a little bit of knowledge can cause some great misunderstandings.

The waveform is stair-cased. Digital audio is quantized in the X and Y axis, so I don't even understand how anyone can fathom it as anything else. The lower the resolution on either axis, the more stepping occurs.

If you have a DAC that does not have the reconstruction filter integrated, and you can probe the point where the data becomes an analog waveform again, the stair-case is quite literally right there. I did this on my MT-32, for example, since it uses an old Burr-Brown DAC with an external S&H circuit, and finally the discrete analog reconstruction filter. The point of that filter is to reduce the slew rate and thereby remove the upper harmonics that are not part of the desired signal. They don't go away completely. No brick-wall filter is literally a brick-wall. It's a slope, of varying degree, and there are negative consequences when you push it to higher orders - namely transient response and ringing artifacts.

The Xiph guy is going to show you an oscilloscope image where you can see for yourself that the resulting waveform (with only two data points per cycle!) is not stepped. But you have to understand, this is because the stepping has been filtered out. It becomes a semantic argument "whether they exist at all, then" I suppose, but the difference is mainly important due to the discussion topic here:

Old sound cards play samples at various rates, but didn't always have a filter with a -3dB cutoff at the Nyquist frequency. That can change the sonic profile of sampled audio whose Nyquist is below the fixed filter frequency. You will get aliasing artifacts in this case. Guess where those come from.

Now, most DACs today are more sophisticated than those early ones using resistor ladders and/or S&H latches. So there is no point where you'll be able to measure those steps. It's rather dubious to claim they don't exist at all, though.
SirNickity
Oldbie
 
Posts: 818
Joined: 2018-11-08 @ 04:51

Re: The "12-bit" Sound Blaster 16 Myth

Postby Scali » 2019-5-13 @ 07:36

SirNickity wrote:The waveform is stair-cased. Digital audio is quantized in the X and Y axis, so I don't even understand how anyone can fathom it as anything else. The lower the resolution on either axis, the more stepping occurs.


Perhaps you need to look at the videos again.
The waveform is sampled, which indeed is quantized in time and dynamic range. However, samples are just points, as in 'infinitely short' in terms of time.
The waveform would only get 'stairstepped' if you stretch out each sample in time in the horizontal direction. In practice this is never the case, not even with the cheapest of DACs.

SirNickity wrote:Old sound cards play samples at various rates, but didn't always have a filter with a -3dB cutoff at the Nyquist frequency. That can change the sonic profile of sampled audio whose Nyquist is below the fixed filter frequency. You will get aliasing artifacts in this case. Guess where those come from.


While it is true that the filtering is suboptimal in most cases, that doesn't mean there is *no* filtering, or that you'd get stairstepping. You'd probably get something 'in between'... as in, there may be extreme cases in terms of frequencies/dynamics where the filtering is inadequate, and you'd get (something close to) stairstepping, but with other types of signals, you do not.
Getting back to the topic at hand though, ironically enough, the SB16 was one of the first cards to implement an adaptive low-pass filter, which responded to the sample rate. So if any card shouldn't suffer from this, it should be the SB16. But as NRS said, apparently there were some issues with early implementations. That's just the story of early SBs all around: their designs/implementations were not making the most of the building blocks they were using. Quality was poor.

The Amiga had a low-pass filter, which could be disabled in code. In practice, most games/demos disabled the filter, especially once 1 mb Amigas became the norm, and more detailed samples could be used.
The SB Pro also has a low-pass filter that can be disabled.
Scali
l33t
 
Posts: 4351
Joined: 2014-12-13 @ 14:24

Re: The "12-bit" Sound Blaster 16 Myth

Postby Cloudschatze » 2019-5-13 @ 17:14

NewRisingSun wrote:I was always very dissatisfied with the CT1740's DOS game sound quality for two reasons:
  1. audible clicks during blocks of single-cycle DMA playback, much more noticeable than on earlier cards I had, such as the Sound Blaster Pro 2;
  2. a constant loud hiss at low sampling rates (such as the 11,025 Hz used by DOS games) whose frequency seems to be a function of the sampling rate, in both 8-bit and 16-bit modes.

To Creative's detriment, it's issues like these - primarily related to 8-bit playback, and separate from any analog stage or build quality discussion - that seem to define a large portion of the negative perception of the SB16. Not that the perception based on those reasons isn't justified, especially for the time; it's just that, even per your experience, it's countered by 16-bit, 44.1kHz playback. To that end, I'm of the opinion that the SB16 was, and is, better suited to Windows than DOS.

SirNickity wrote:I can't recall ever hearing digital noise (bus activity, etc.) on any of my Creative cards.

This is commonly experienced with the 8-bit Sound Blaster cards. I have not encountered system-bus noise with an SB16 though, and where the earlier variants have 4-layer PCBs besides, I would think it ought to be even less of an issue. To be clear on this point, is anyone actually claiming to have heard bus noise through an SB16, and which card(s), if so?
User avatar
Cloudschatze
Oldbie
 
Posts: 1096
Joined: 2005-6-16 @ 14:32

Re: The "12-bit" Sound Blaster 16 Myth

Postby SirNickity » 2019-5-13 @ 17:59

Scali wrote:
SirNickity wrote:The waveform is stair-cased. Digital audio is quantized in the X and Y axis, so I don't even understand how anyone can fathom it as anything else. The lower the resolution on either axis, the more stepping occurs.


Perhaps you need to look at the videos again.
The waveform is sampled, which indeed is quantized in time and dynamic range. However, samples are just points, as in 'infinitely short' in terms of time.
The waveform would only get 'stairstepped' if you stretch out each sample in time in the horizontal direction. In practice this is never the case, not even with the cheapest of DACs.


This is precisely what older DACs do. Like, exactly. Look at the schematic for the MT-32. It has a one-channel DAC and IIRC 6 channels of audio. The DAC outputs a fixed voltage via an analog switch into a S&H latch, which holds that voltage until the DAC comes back around to service that channel again. It's been a while since I've looked at this circuit so I don't remember the exact DAC model number, but it was super common in the late 80s, early 90s. Roland used it, or similar, in all their synths. Some highish-end CD players used it, etc., and that's just the ones I know about.

Then there's the Covox and DSS. Heck even the SB with its microcontroller-based DSPs did this. :-) I know those aren't representative of mass consumer audio, but my point is, it's way more common than you would think. Not anymore, since successive approximation and similar techniques are cheap and easy now.
SirNickity
Oldbie
 
Posts: 818
Joined: 2018-11-08 @ 04:51

PreviousNext

Return to Sound

Who is online

Users browsing this forum: No registered users and 1 guest