Sound Cards vs. Expensive Motherboards vs. Cheap MOBOs

[Guest]

So I've never really fully understood the use of soundcards. I've read that they were a necessity back when motherboard didn't have audio headers, but even today there's apparently different qualities and specifications of the onboard audio on mobos. I've always thought all headphone jacks are the same and all USB headphone/speakers sound the same no matter what device you plug them into.

 

So what difference would it make by plugging in some standard headphones into a very old, cheap motherboard, to the most expensive and high-end sound card on the market?

[rice guru]

the sound difference betweena  high end mobo, and a low end mobo is pretty minimal from my experience as long as they are made in the same year . sure they have improved over the years. but the differences on sound quality alone isn't massive. of course it varies by brand but they all typically use the same chips. the main differences though is generally power and shielding , they typically have higher quality caps and sometimes a better chip. higher end mobos are generally used by higher end and noisier parts thus shielding becomes more of an issue. in regards to sound cards depending on the quality of the sound card likea soundblaster ae 9 which is at this point top tier in sound card. they aren't really nessecary for most people. those that care about audio most take their set up externally for those that care about audio have higher end gear, need software controls for some reason, or want soemthing for a htpc and want surround sound then these applications they would benefit from a modern soundcard. otherwise if you are just using a standard non studio or high end headpone just stick to your mobo. 

[Moonzy]

7 hours ago, RonnyDouglas said:

So what difference would it make by plugging in some standard headphones into a very old, cheap motherboard, to the most expensive and high-end sound card on the market?

im no audiophile nor am i very interested in the audio market, but i can try to explain this in whatever i know (from a theoretical standpoint), feel free to correct me and teach me stuffs!

 

you can skip to the TL;DR

 

Sampling Rate

audio signal is an analogue signal, and computer stores music in digital signal

so you would need a DAC (Digital-to-analogue) converter, something to convert the signal from digital to analogue

ideally, the conversion process would produce a smooth sine wave without steps

but because computer works by digital signal, the wave it produces is a "simulated sine-wave"

 

picture for sine-wave vs simulated sine-wave:

Spoiler

the audio signal is distorted, compared to the intended input

a smaller step allows the digital signal to better follow the curve of the real intended signal, to more accurately simulate a sine wave

this is also called sampling rate, higher sampling rate = smaller step size

 

as for how a higher quality audio solution will improve this, this is mostly affected by the "input" side of things, when recording is done, since you can't really change the sample rate after recording it. (other than some AI stuffs, but wont get into that)

 

 

Noise

when a current is flowing in a wire, it produces EMI (electro-magnetic interference), which is a changing magnetic field that can induce currents in parallel wires.

when a wire carrying a pure sine wave is near other wires carrying other signals, small currents can be induced in the signal carrying wire producing noise, which distorts the signal

 

picture of noisy sine wave:

Spoiler

wiring is not the only source of noise, though

choice of capacitor matters as well, as not all capacitors are created equal, they have different frequency response characteristics (the reason of which is very complicated, but just know that each type of capacitor reacts differently as frequency changes)

a higher quality DAC will have better precaution when designing their circuitry to minimalise the noise-to-signal ratio, which will produce a cleaner audio signal

 

Impedance

Impedance is the electrical resistance equivalent for an inductor, which is what drives most headphone drivers

Some headphones have large impedance, up to 300Ohm (unit of measuring resistance) or even higher

while some headphones have only 18 Ohm or even lower.

so DAC output may not be sufficient enough to drive some higher impedance headphones, thus requiring an AMP (Amplifier) to boost the signal to the appropriate levels

You can plug the headphones into your computer and it may work, but the volume will be small if the output capability is insufficient.

So some people may buy a separate AMP to boost their signal to the correct amplitude.

 

now, difference between a cheap and highend motherboard isnt that different nowadays from what i've gathered (i am in no position to judge this personally because i know jack about audio)

all i know is that if you have a hard to drive (high impedance) headset, you need to get an AMP

as for dedicated DAC, im not too sure about those

another reason to get a separate DAC and AMP is that if your PC is very far, you could use a table top DAC AMP combo, you can transmit the signal to the DAC digitally from your PC (which doesnt distort the signal as long as it doesnt corrupt the data) and then convert the signal near the listener.

[The Flying Sloth]

5 hours ago, Moonzy said:

 

Sampling Rate

audio signal is an analogue signal, and computer stores music in digital signal

so you would need a DAC (Digital-to-analogue) converter, something to convert the signal from digital to analogue

ideally, the conversion process would produce a smooth sine wave without steps

but because computer works by digital signal, the wave it produces is a "simulated sine-wave"

 

picture for sine-wave vs simulated sine-wave:

  Reveal hidden contents

the audio signal is distorted, compared to the intended input

a smaller step allows the digital signal to better follow the curve of the real intended signal, to more accurately simulate a sine wave

this is also called sampling rate, higher sampling rate = smaller step size

 

as for how a higher quality audio solution will improve this, this is mostly affected by the "input" side of things, when recording is done, since you can't really change the sample rate after recording it. (other than some AI stuffs, but wont get into that)

In theory, of course you are right, in practice there is no measurable difference between 44.1khz and anything higher. Most studios mix down to either 44.1 or 48khz for this reason and many studios don't even produce at above 48khz. 

 

If you can't measure the audible difference between 44.1 and anything higher the argument that the sample rate produces higher 'definition' audio is pretty dead in the water as if it's not measurable you for sure cannot hear the difference.

[Spuriae]

7 hours ago, Moonzy said:

ideally, the conversion process would produce a smooth sine wave without steps

but because computer works by digital signal, the wave it produces is a "simulated sine-wave"

This is true for the resistor ladder DACs that are usually taught in most EE classes, but not in practice. Virtually all modern audio frequency sampling uses a different process which samples at an incredibly high rate then trades time resolution for level resolution. Although switching does occur, it usually happens at a frequency so far above the cutoff of the analog output filter that steps do not appear in the output signal.

8 hours ago, Moonzy said:

Noise

when a current is flowing in a wire, it produces EMI (electro-magnetic interference), which is a changing magnetic field that can induce currents in parallel wires.

when a wire carrying a pure sine wave is near other wires carrying other signals, small currents can be induced in the signal carrying wire producing noise, which distorts the signal

EMI is the most well-known source of noise, but in practice is rarely an issue; it only tends to be significant at high (non-audible) frequencies. The most common audible noise in audio circuits is due to grounding issues.

 

In state-of-the-art audio amplifiers, noise is usually dominated by quantum effects. (e.g. thermal noise, flicker noise, and shot noise)

8 hours ago, Moonzy said:

choice of capacitor matters as well, as not all capacitors are created equal, they have different frequency response characteristics (the reason of which is very complicated, but just know that each type of capacitor reacts differently as frequency changes)

At audio frequencies, parasitic inductance (the main source of differing frequency responses) is rarely significant enough to cause audible differences between capacitors. Though different capacitors may have audible effects in bypass/decoupling and feedback loop applications where ultrasonic performance can affect the performance of other components in the signal path, differences between capacitors placed directly in the signal path are usually due to nonlinearities.

 

Polarized electrolytic capacitors, for instance, predictably conduct electricity differently depending on the current direction, and tend to produce audible even-order harmonic distortion. In nonpolarized capacitors with high dielectric constants, capacitance may vary with the applied voltage and will be dominated by odd-order harmonic distortion (in the absence of DC bias).

8 hours ago, Moonzy said:

Impedance is the electrical resistance equivalent for an inductor, which is what drives most headphone drivers

Some headphones have large impedance, up to 300Ohm (unit of measuring resistance) or even higher

while some headphones have only 18 Ohm or even lower.

Surprisingly the inductive contribution to impedance for dynamic drivers tends to be quite low, only becoming significant at higher frequencies. Through most of the audible range, impedance is dominated by a resistive (real) term and a peak due to back-EMF from the physical resonant frequency of the driver.

8 hours ago, Moonzy said:

so DAC output may not be sufficient enough to drive some higher impedance headphones, thus requiring an AMP (Amplifier) to boost the signal to the appropriate levels

You can plug the headphones into your computer and it may work, but the volume will be small if the output capability is insufficient.

So some people may buy a separate AMP to boost their signal to the correct amplitude.

The voltage levels on even standard consumer equipment are sufficient for most listening conditions (usually ~1Vrms+ before clipping). It's the current output that is more limited, which often makes low-impedance loads more difficult to drive. For most motherboards the sweet spot is somewhere between 80Ω and 250Ω: above this, the output is voltage limited, below this, output is current limited. In addition, cheap outputs tend to have high output impedances, which cause larger frequency response issues with low impedance than high impedance loads.

 

Funnily enough, headphone amplifiers are often run in their lowest gain setting, making their main function current buffering rather than voltage amplification.

3 hours ago, The Flying Sloth said:

In theory, of course you are right, in practice there is no measurable difference between 44.1khz and anything higher. Most studios mix down to either 44.1 or 48khz for this reason and many studios don't even produce at above 48khz. 

 

If you can't measure the audible difference between 44.1 and anything higher the argument that the sample rate produces higher 'definition' audio is pretty dead in the water as if it's not measurable you for sure cannot hear the difference.

It depends whether or not ultrasonic noise is present, and how good the analog circuitry is. Spurious signals above the Nyquist frequency but below the Nyquist rate get aliased into the audible band. The Nyquist sampling theorem only promises perfect signal reproduction assuming an ideal brickwall low pass filter (which doesn't exist in real life) at the Nyquist frequency.

In practice, the common sample rates (44.1kHz and 48kHz) were selected to give some frequency buffer for real filters, which have a less abrupt transition. ADC's are supposed to have analog antialiasing filters at their inputs to remove the ultrasonic components from the signal before it can be sampled, but these being real filters, they seldom remove all of the noise. In addition, high frequency components can interact with lower frequency ones within analog circuits to produce IMD inside the audible band that cannot be removed by such filters.

Philosophical objections aside, there are some arguments to be made for recording at higher sample rates purely for the benefit of being able to more effectively filter out noise with digital filtering, should it be significant. The downside is that ADCs tend to perform worse at higher sample rates. Though the uncorrelated noise performance may be recoverable by resampling, distortion cannot. (I'm splitting hairs here: these effects are absolutely measurable but dubiously audible in most well-designed equipment/setups).

For resistor ladder DACs different sample rates are always measurable and sometimes audible, as resampling the signal to a higher frequency also pushes additional noise and aliasing (and IMD!) into higher ultrasonic frequencies. On Delta-Sigma DACs this is less of an issue as signals are already upsampled internally by design.

[Derkoli]

11 hours ago, RonnyDouglas said:

So I've never really fully understood the use of soundcards. I've read that they were a necessity back when motherboard didn't have audio headers, but even today there's apparently different qualities and specifications of the onboard audio on mobos. I've always thought all headphone jacks are the same and all USB headphone/speakers sound the same no matter what device you plug them into.

 

So what difference would it make by plugging in some standard headphones into a very old, cheap motherboard, to the most expensive and high-end sound card on the market?

The difference?

 

Quite minimal actually.

 

MOBO audio has come quite far since the day's of the first sound cards. The one huge limitation is output voltage/current (depending upon load impedance)

 

The reason why we use external AMP's is to get more power. (In the case of solid-state (transistor) amplification amplification it is current amplification. Tubes are voltage amplifiers by design, but thats another discussion)

 

We also use external gear due to lower noise, which gives us better Dynamic Range and no hissing crap. Most motherboards won't hiss anymore. But coil whine is still a thing with some GPU's, so onboard audio still isn't going to win.

 

Alot of people also use Soundcards for all the effects that you can get with them. Creative Labs in particular bundles ALOT of software with their soundcards.

 

Not all headphone jacks will sound the same, Atleast on higher end/more revealing systems.

 

Cheaper headphones might not actually show the user flaws in the headphone output and the user may not care. Nor may the user actually be able to hear small differences.

 

With a higher end system, Flaws are more obvious. Especially with studio gear, which by design is supposed to show all flaws and be as transparent as possible.

[HumdrumPenguin]

11 hours ago, Derkoli said:

The difference?

 

Quite minimal actually.

 

MOBO audio has come quite far since the day's of the first sound cards. The one huge limitation is output voltage/current (depending upon load impedance)

 

The reason why we use external AMP's is to get more power. (In the case of solid-state (transistor) amplification amplification it is current amplification. Tubes are voltage amplifiers by design, but thats another discussion)

 

We also use external gear due to lower noise, which gives us better Dynamic Range and no hissing crap. Most motherboards won't hiss anymore. But coil whine is still a thing with some GPU's, so onboard audio still isn't going to win.

 

Alot of people also use Soundcards for all the effects that you can get with them. Creative Labs in particular bundles ALOT of software with their soundcards.

 

Not all headphone jacks will sound the same, Atleast on higher end/more revealing systems.

 

Cheaper headphones might not actually show the user flaws in the headphone output and the user may not care. Nor may the user actually be able to hear small differences.

 

With a higher end system, Flaws are more obvious. Especially with studio gear, which by design is supposed to show all flaws and be as transparent as possible.

I haven't come across a single motherboard that doesn't sound like ass, even if on paper they should be fine. Do you remember any motherboard model you've tried that it was ok in your opinion sound wise? I'm curious about that now. My guess is that you haven't tried anything on a motherboard in a long time and you're looking strictly on numbers. If by any chance that's the case, grab any amplifier you have (ideally less than £500) and do a side by side comparison. I think you will be surprised at the difference in overall sound quality.

[Derkoli]

5 hours ago, HumdrumPenguin said:

I haven't come across a single motherboard that doesn't sound like ass, even if on paper they should be fine. Do you remember any motherboard model you've tried that it was ok in your opinion sound wise? I'm curious about that now. My guess is that you haven't tried anything on a motherboard in a long time and you're looking strictly on numbers. If by any chance that's the case, grab any amplifier you have (ideally less than £500) and do a side by side comparison. I think you will be surprised at the difference in overall sound quality.

I've always noticed a difference between onboard and external solutions.

 

I'm just trying to say that most people won't hear the difference, either due to not caring, or their gear can't show the flaws.

 

Personally I think the majority of modern, mid to high end level Motherboards will sound good to an average person. However I've never been happy with any onboard sound solutions and most audiophiles can hear a difference between a motherboard and an external solution.