It's 2015: Why does Onboard audio STILL suck?

Page 5 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

chucky2

Lifer
Dec 9, 1999
10,038
36
86
So has someone tried something like a O2+ODAC from JDS and actually noticed a benefit when listening to music and/or playing games?
 

Cerb

Elite Member
Aug 26, 2000
17,484
33
86
So has someone tried something like a O2+ODAC from JDS and actually noticed a benefit when listening to music and/or playing games?
Different DACs and amps, but I've been doing that since USB audio was suppoorted well enough, and used sound cards with external analog boxes before that. I still haven't heard a PC with a decent video card that doesn't make obvious noise on the onboard analog outs when the video card is not at idle, even some Asuses that have the separate traces and layers, shielded chips, and all that jazz. Often, the CPU cycling around power states will do the same thing. While you won't usually hear anything with included speakers, it really doesn't take much for it to start being apparent, and a lot of people don't hear it, even when it's obvious to me. With sensitive headphones, it's clear as day, though.
 

RampantAndroid

Diamond Member
Jun 27, 2004
6,591
3
81
Yes, it is. Higher frequencies do not follow the path of least resistance. DC will. A fan should not be doing much in terms of AC, and everything it does should take the 'shortest' paths from power and ground. Meanwhile, your mouse is working at, IIRC, 60MHz.
They do seek the path of least resistance. Specifically, what you're getting at is the path of least impedance...which will be inductance if you're talking about AC power (and on how the wiring is set up).

You will in the opamp, which there will be at least two of, if not four (bipolar inputs). Not to mention some chance of antennae being connected.
Huh? Are you getting at how the opamp is more or less giving a single ended output from differential inputs (they don't have to be fully differential - you can do it with the signal and a reference of 0)?

Not if it contaminates the input.
I maybe wasn't clear here - my point was that you're going to totally destroy the incoming signal. You don't put a rectifier on this.

Explain GSM noise, then, or picking up AM stations, with unshielded opamps (or insufficiently shielded, or unlucky cable lengths). Each follows a different path to just that end result. GSM noise is closer to most of what we can hear from inside our PCs.

Which is it? It's change is the fields that can causes noise, not their static state.
I maintain that an audio amp isn't going to take something subsonic and make you suddenly hear it. Crappy amps have 60hz hum because the AC coming from the wall is 60hz (or it's 120hz for a full wave rectifier with no smoothing.)

What opamp do you have that picks up AM signals (let alone FM signals)? A half wavelength antenna for AM is what, 800 some odd feet? 400ft for the quarter wavelength. Now, I get that AM antennas in use are *not* that long, but that's only because they have a variable cap or similar to change the frequency the antenna resonates at. At AM and FM frequencies and powers, you should NOT be picking them up even in an unshielded opamp.
 

RampantAndroid

Diamond Member
Jun 27, 2004
6,591
3
81
Different DACs and amps, but I've been doing that since USB audio was suppoorted well enough, and used sound cards with external analog boxes before that. I still haven't heard a PC with a decent video card that doesn't make obvious noise on the onboard analog outs when the video card is not at idle, even some Asuses that have the separate traces and layers, shielded chips, and all that jazz. Often, the CPU cycling around power states will do the same thing. While you won't usually hear anything with included speakers, it really doesn't take much for it to start being apparent, and a lot of people don't hear it, even when it's obvious to me. With sensitive headphones, it's clear as day, though.
What video cards have analog audio on them? Only ones I see throw digital audio out over HDMI, and leave it up to other hardware to do the DA conversion.

So has someone tried something like a O2+ODAC from JDS and actually noticed a benefit when listening to music and/or playing games?
I never tried an ODAC because NwAvGuy is a [insert some unpleasant adjective in here] (perma-banned from Head-Fi last I knew), but I have built AMB's (Ti Kan's) gamma2 DAC with a good amount of success. I ended up with audio that sounded more than good enough for what I have in speakers and headphones. I plan to build his newer gamma1.5 which has a USB DAC that can go to 24/192 and a small-ish headphone amp in it (of course, this HP amp is powered by USB and it's PRETTY limited. He lists the Vpp it'll clip at.)
 
Last edited:

IlllI

Diamond Member
Feb 12, 2002
4,923
9
81
i remember a long time ago some motherboard manufacturer made a mobo with vacuum tubes.
 

Cerb

Elite Member
Aug 26, 2000
17,484
33
86
What video cards have analog audio on them? Only ones I see throw digital audio out over HDMI, and leave it up to other hardware to do the DA conversion.
The motherboard has the analog audio. The video card makes the most easily heard interference, IME, on gaming PCs.
 
Last edited:

Cerb

Elite Member
Aug 26, 2000
17,484
33
86
They do seek the path of least resistance. Specifically, what you're getting at is the path of least impedance...which will be inductance if you're talking about AC power (and on how the wiring is set up).
AC power is far away. The interference from a fan would be from it turning on and off, which would come from its normal operation (2+ times per cycle), controlling it by PWM on power (1+kHz), or the fan driving the motor using the 4-pin PWM signal (25kHz+harmonics).

In the case of USB devices themselves creating noise, it's going to be in the tens of MHz (about 60Hz for USB 1-2, IIRC), where inductance is most of what matters.

Very slowly changing stuff, like DC power draw, can be put on the same traces as audio power and 0V, so long as it has a better path to follow than through the audio devices. It would be easiest just to make them separate, but sharing a 0V somewhere else isn't going to hurt anything.

Huh? Are you getting at how the opamp is more or less giving a single ended output from differential inputs (they don't have to be fully differential - you can do it with the signal and a reference of 0)?
No. Generally, if you don't have a balanced output, your PC's audio is single-ended the whole way.

I maybe wasn't clear here - my point was that you're going to totally destroy the incoming signal. You don't put a rectifier on this.
That's not what I mean. I mean that as an effect.

See Op-Amp Applications Handbook. In the 2005 one, the relevant section begins on p. 720. It's on Google Books, with that section allowed for preview, too.

I maintain that an audio amp isn't going to take something subsonic and make you suddenly hear it. Crappy amps have 60hz hum because the AC coming from the wall is 60hz (or it's 120hz for a full wave rectifier with no smoothing.)
217Hz is not supersonic, and is where the biggest spike will be seen with 3G and older GSM. 4G-only phones FTW! But, anyway, high energy spikes of high frequency signals can get turned into lower frequency signals, which, if contaminating your signal or ground, make noises as you do stuff, like with your mouse, or a PSU, or any number of other switching PSUs, and chips with quickly varying loads.

What opamp do you have that picks up AM signals (let alone FM signals)?
The AD797 is known for being able pick up AM, with the "right" wiring. FM would take some pretty interesting accidental circuitry :).
 

chucky2

Lifer
Dec 9, 1999
10,038
36
86
snip

I never tried an ODAC because NwAvGuy is a [insert some unpleasant adjective in here] (perma-banned from Head-Fi last I knew), but I have built AMB's (Ti Kan's) gamma2 DAC with a good amount of success. I ended up with audio that sounded more than good enough for what I have in speakers and headphones. I plan to build his newer gamma1.5 which has a USB DAC that can go to 24/192 and a small-ish headphone amp in it (of course, this HP amp is powered by USB and it's PRETTY limited. He lists the Vpp it'll clip at.)
I thought that was the point of the O2+ODAC, that he went through the effort to debunk the BS of audio and get something on the scope that was as good as something like the Benchmark, but not with insane price markups and hype?

I'm just learning about it now, was his effort BS or did he actually accomplish what he set out to do? What I know is no one has heard from him it sounds like in a while (drop off face of earth online)...

Chuck
 

RampantAndroid

Diamond Member
Jun 27, 2004
6,591
3
81
I thought that was the point of the O2+ODAC, that he went through the effort to debunk the BS of audio and get something on the scope that was as good as something like the Benchmark, but not with insane price markups and hype?

I'm just learning about it now, was his effort BS or did he actually accomplish what he set out to do? What I know is no one has heard from him it sounds like in a while (drop off face of earth online)...

Chuck
I've no idea where he went; he ruffled A LOT of feathers (and proposed the idea that you cannot empirically measure the quality of audio equipment, that you can only use your ears). I know he compared his stuff to Ti Kan's stuff, notably his gamma2 DAC. I also seem to remember that to make his ODAC seem better, he intentionally built the gamma2 incorrectly. He also touted his ODAC as better because it could do higher bit depth & sample rates than the gamma2 could - which is true. His ODAC used a chip that you and I cannot buy in quantities less than 1000 at a time. The gamma2 uses one of the FEW ICs our there that can be purchased one at a time for USB audio (the TI 2707.) It's the same reason that the Twisted Pear Audio Buffalo DACs are all pre-made (it uses an ESS Sabre chip.) Ti Kan's latest uses a Atmel microcontroller, bypassing the need for a specialized chip.

Generally though, NWAvGuy was an exceptionally abrasive person and not one I could stand to sift through his writing for long periods of time. Will his stuff work? Sure. I've no clue how good it is...and I doubt I'll ever find out.

AC power is far away. The interference from a fan would be from it turning on and off, which would come from its normal operation (2+ times per cycle), controlling it by PWM on power (1+kHz), or the fan driving the motor using the 4-pin PWM signal (25kHz+harmonics).
That stuff is easily dealt with by adding a decoupling cap or similar. I don't know where you're going with this. I don't know how you're saying that high frequency doesn't follow the path of least resistance. Either it's DC and it's following path of least resistance, or it's AC and it's following the path of least inductance. I'm really over simplifying here, but that's the jist of it.

Analog Devices has a doc on this issue, actually: http://www.powerampdesign.net/images/AN-24_Eliminating_Circuit_Noise_From_Cooling_Fans.pdf

In the case of USB devices themselves creating noise, it's going to be in the tens of MHz (about 60Hz for USB 1-2, IIRC), where inductance is most of what matters.
The noise from a USB device will be on the supply side (likely on ground?) An the DAC just isn't decoupled correctly. The polling rate of the sensor should have nothing to do with this.

Very slowly changing stuff, like DC power draw, can be put on the same traces as audio power and 0V, so long as it has a better path to follow than through the audio devices. It would be easiest just to make them separate, but sharing a 0V somewhere else isn't going to hurt anything.
Sharing ground without decoupling it is going to cause a large number of problems.

No. Generally, if you don't have a balanced output, your PC's audio is single-ended the whole way.
Not necessarily. Either way, I only mention that because you seem to think there's a rectifier in a DAC?

That's going to have nothing to do with other components making interference that may end up received, rectified, and amplified into audio range noise.
217Hz is not supersonic, and is where the biggest spike will be seen with 3G and older GSM. 4G-only phones FTW! But, anyway, high energy spikes of high frequency signals can get turned into lower frequency signals, which, if contaminating your signal or ground, make noises as you do stuff, like with your mouse, or a PSU, or any number of other switching PSUs, and chips with quickly varying loads.
It's 216.7 or something, because of TDMA. The noise from GSM isn't outside of human hearing, and it is caused by the frame size used in the older GSM spec. UMTS changed that, iirc.

Varying loads will absolutely cause problems (and that's covered in the analog devices PDF above.) For higher frequency noise, are you getting at the EM inducing voltage in the wiring? If that's what you're referring to, yes - that will absolutely cause problems. :)

It all comes back to needing to properly filter the power of the circuits AND properly shield them as well. I sometimes see people building custom PC cases and cases for audio equipment out of wood and I want to smack them. Same as the people putting transformers next to audio equipment.

The AD797 is known for being able pick up AM, with the "right" wiring. FM would take some pretty interesting accidental circuitry :).
Yeah, that's what I was getting at - it has to be wired in a REALLY bad way without any proper shielding.
 

StrangerGuy

Diamond Member
May 9, 2004
8,443
124
106
I thought that was the point of the O2+ODAC, that he went through the effort to debunk the BS of audio and get something on the scope that was as good as something like the Benchmark, but not with insane price markups and hype?

I'm just learning about it now, was his effort BS or did he actually accomplish what he set out to do? What I know is no one has heard from him it sounds like in a while (drop off face of earth online)...

Chuck
Nwavguy is absolutely correct, a sane voice in a sea of madness. He is spot on at the fact while it is trivial to design something that sounds good by ear the actual measurable performance is a big unknown without dedicated test equipment such as 5 digit Audio Precision dScope like he did, and he also proved expensive != better.
 

Cerb

Elite Member
Aug 26, 2000
17,484
33
86
That stuff is easily dealt with by adding a decoupling cap or similar. I don't know where you're going with this. I don't know how you're saying that high frequency doesn't follow the path of least resistance. Either it's DC and it's following path of least resistance, or it's AC and it's following the path of least inductance. I'm really over simplifying here, but that's the jist of it.
And that's too oversimplified. Low frequency AC may as well be DC. The faster it is, the closer it will follow the path of least inductance; the slower it is, the closer it will follow the path of least resistance. It's even easy to Google pretty pictures of it.

Page 3, top paragraph, is just what I was talking about. No matter what you have for a cap, if you route the ground so the easiest path is through the audio section, you'll get lots of noise, effectively like an AC signal, even if it's low frequency. Make it so it can go from the PSU, to the fan, to ground, without having to go through there, and it won't be a problem.

The noise from a USB device will be on the supply side (likely on ground?) An the DAC just isn't decoupled correctly. The polling rate of the sensor should have nothing to do with this.
The polling rate has nothing to do with anything, and generally maxes out at 1kHz, not some MHz. Decoupling is not going to be some simple thing for many-MHz signals.

Not necessarily. Either way, I only mention that because you seem to think there's a rectifier in a DAC?
There may be, or not, but that's generally trade secret stuff. And, again, it's not a specific component dedicated to the task, but an effect. I cited quite possibly the most well-known source, and you can go read several pages of the relevant section for free.
 

chucky2

Lifer
Dec 9, 1999
10,038
36
86
I've no idea where he went; he ruffled A LOT of feathers (and proposed the idea that you cannot empirically measure the quality of audio equipment, that you can only use your ears). I know he compared his stuff to Ti Kan's stuff, notably his gamma2 DAC. I also seem to remember that to make his ODAC seem better, he intentionally built the gamma2 incorrectly. He also touted his ODAC as better because it could do higher bit depth & sample rates than the gamma2 could - which is true. His ODAC used a chip that you and I cannot buy in quantities less than 1000 at a time. The gamma2 uses one of the FEW ICs our there that can be purchased one at a time for USB audio (the TI 2707.) It's the same reason that the Twisted Pear Audio Buffalo DACs are all pre-made (it uses an ESS Sabre chip.) Ti Kan's latest uses a Atmel microcontroller, bypassing the need for a specialized chip.

Generally though, NWAvGuy was an exceptionally abrasive person and not one I could stand to sift through his writing for long periods of time. Will his stuff work? Sure. I've no clue how good it is...and I doubt I'll ever find out.

snip
I have no history with him, just was looking at what a good DAC+headphone amp would be, and came across a bunch of O2+ODAC references. Followed them to his blog, http://nwavguy.blogspot.com/2012/04/odac-released.html, but after 2012...he just disappeared.

To me it sounds like he built his DAC to compare to the Benchmark DAC1 which he used as a reference. Are you saying that this gamma2 is better? I'll have to go check it out...

Chuck
 

RampantAndroid

Diamond Member
Jun 27, 2004
6,591
3
81
I have no history with him, just was looking at what a good DAC+headphone amp would be, and came across a bunch of O2+ODAC references. Followed them to his blog, http://nwavguy.blogspot.com/2012/04/odac-released.html, but after 2012...he just disappeared.

To me it sounds like he built his DAC to compare to the Benchmark DAC1 which he used as a reference. Are you saying that this gamma2 is better? I'll have to go check it out...

Chuck
The gamma2 is...different. The ODAC is meant for, I believe, USB only. The gamma2 has USB, but it isn't ideal. If DIY AND USB is what you want, look at the gamma1.5 for USB. If SPDIF is in your future (and not USB) the gamma2 (NOT the gamma1) is just fine.

The ODAC is probably a fine device too - I just don't buy into the crud that he was peddling.

And that's too oversimplified. Low frequency AC may as well be DC. The faster it is, the closer it will follow the path of least inductance; the slower it is, the closer it will follow the path of least resistance. It's even easy to Google pretty pictures of it.

Page 3, top paragraph, is just what I was talking about. No matter what you have for a cap, if you route the ground so the easiest path is through the audio section, you'll get lots of noise, effectively like an AC signal, even if it's low frequency. Make it so it can go from the PSU, to the fan, to ground, without having to go through there, and it won't be a problem.

The polling rate has nothing to do with anything, and generally maxes out at 1kHz, not some MHz. Decoupling is not going to be some simple thing for many-MHz signals.

There may be, or not, but that's generally trade secret stuff. And, again, it's not a specific component dedicated to the task, but an effect. I cited quite possibly the most well-known source, and you can go read several pages of the relevant section for free.
I say you're thinking of AC power & path of least inductance, you tell me you don't mean AC power. Now you're talking about AC power again.

I cannot think of any DAC that would have a rectifier on the inputs...any at all. Either way, this isn't going anywhere (we're more or less agreeing on a bunch of things...)
 

chucky2

Lifer
Dec 9, 1999
10,038
36
86
The gamma2 is...different. The ODAC is meant for, I believe, USB only. The gamma2 has USB, but it isn't ideal. If DIY AND USB is what you want, look at the gamma1.5 for USB. If SPDIF is in your future (and not USB) the gamma2 (NOT the gamma1) is just fine.

The ODAC is probably a fine device too - I just don't buy into the crud that he was peddling.

snip
Oh I want something assembled already, I have zero time to be putting sh1t together.

You don't believe his measurements or just his viewpoints?
 

RampantAndroid

Diamond Member
Jun 27, 2004
6,591
3
81
Oh I want something assembled already, I have zero time to be putting sh1t together.

You don't believe his measurements or just his viewpoints?
It's been a long time since I read all of it, but what I remember is that he was trying to make other equipment measure poorly. He also goes so far as to say the O2 amp is equal to (indistinguishable from) something like a Beta22...which I find to be pretty comical (and is NOT supported by people who have done comparisons.)

So I guess I don't believe anything from him, other than the fact that he created a DAC and amp that works (and for the cost, probably works pretty well.)

For the DIY stuff, it is a little bit of a time investment, but people do build Ti Kan's stuff and sell it - check out the website: www.amb.org.
 

chucky2

Lifer
Dec 9, 1999
10,038
36
86
Fair enough, I'll have to check Ti Kan's stuff out...I'm going to assume the B22 is pretty expensive, and I just don't need that for occasional music and gaming uses.
 

RampantAndroid

Diamond Member
Jun 27, 2004
6,591
3
81
The B22 is a reference amp, and you likely don't want one. :)

If USB audio is in your future, look into the gamma1.5. It's probably more than a O2 though (I'm about to build one, and I think it'll be around $200 USD in the end.)
 

chucky2

Lifer
Dec 9, 1999
10,038
36
86
It's not just an amp I'm wanting though, it's a DAC and an AMP. The onboard sound is hosed (hardware-wise) on my motherboard, and I'm not getting a new board (on AM3+, going Intel next time).

So I figured I'd just tackle an audio upgrade at the same time I got working audio back on my desktop. Basically from what I've been able to determine on the O2+ODAC, is that it is completely neutral, which was one of the major design parameters - it adds nothing to what isn't there in the sound file being sent to it. Some people it sounds do not like that, they are used to something having say bass enhanced based (or whatever is being debated) and find the O2+ODAC lacking vs. other alternatives.

I'm honestly not sure where I stand on this: Purist-wise, I want to hear what the artist intended, I don't want the sound changed from what they've provided me. On the other hand, maybe I'd actually prefer that song played with more base than what the artist feels should be provided - does that make my preference wrong? No, just different from the artist.

It seems this is the conundrum in the O2+ODAC vs others debates...
 

RampantAndroid

Diamond Member
Jun 27, 2004
6,591
3
81
For things to be totally neutral, you need headphones or speakers that are as well (or you need to get a calibrated mic - I have one, it was around $50 I think?) that can take measurements for you to correct with an EQ.

Regardless, an amp isn't just about the thd or the output impedance. There's a lot to be said for what gain level it is (and what headphones you'll be driving), if the DAC is upsampling or not (generally a good thing), what the max voltage peak to peak is before it'll clip, what power it can sustain versus what it can peak and so on.

Like I said in my previous post - check out the gamma1.5. It's a USB DAC (no SPDIF) that does up to 24/192 and has a decent headphone amp built in that will work for many headphones, as long as they're not overly hard to drive. I think a schitt stack will also get recommended often.

Also, for what the artist intended, realize that often (depending on the type of music of course) the sound engineers will master the recording to sound the best on the crappiest headphones out there.
 

chucky2

Lifer
Dec 9, 1999
10,038
36
86
Oh, my headphones aren't awesome by any means, they're just a well treated pair of Audio-Technica ATH-M50, certainly nothing special. I'll check out the gamma1.5, thx for the extra info.
 
Dec 30, 2004
12,554
2
76
I've never had a problem with onboard audio, excepting my Via which I solved by installing a ground loop isolator in line
 

jamesgalb

Member
Sep 26, 2014
67
0
0
Would be nice to see some serious AUdio reviews on this board:

http://www.newegg.com/Product/Product.aspx?Item=N82E16813130842

The main three things that stand out to me over other boards is the additional EMI shielding (real shielding, not plastic 'armor'), the dedicated Audio Power plug, and the gold analog connectors... Along with the more standard core shielding and isolated PCB.

I am interest in the Skylake offerings from Gigabyte, ASUS, and now seemingly MSI as well.
 

Puffnstuff

Lifer
Mar 9, 2005
15,555
4,262
136
Well it looks like MSI finally got it by providing that board with a discrete power plug for the audio section. Actually they've copied asus in quite a few areas with it except they forgot a usb 3 motherboard header for the front panel connector. I wish they'd omit the m2 connector on the z97 boards for the shear fact that it forces you to disable pcie slots in order to make use of the m2 slot.
 

jamesgalb

Member
Sep 26, 2014
67
0
0
The one thing that has stood out with the MSI Gaming series audio is the lack of bad audio reviews by users...

When a manufacture goes out of their way to promote the audio, they obviously attract audio-oriented customers. If they fail at producing, it finds its way onto retailer reviews pretty quickly... The worst I have seen for the MSI Gaming so far are complaints about the drivers, which would seem an user issue.

I wish I could get some hybrid Gigabyte, MSI and ASUS board... Dedicated Power, extra audio/IO shielding, swappable OP-AMP, Soundblaster core, high end DAC, good (low) crosstalk/distortion, etc...

Best I've seen:
MSI: http://us.msi.com/product/motherboard/Z97A-GAMING-9-ACK.html#hero-overview
Gigabyte: http://www.gigabyte.com/products/product-page.aspx?pid=4963#ov
ASUS: http://www.asus.com/us/Motherboards/MAXIMUS_VII_FORMULA/

Right now based on specs, setup, and reviews... MSI seems the best.
 

RampantAndroid

Diamond Member
Jun 27, 2004
6,591
3
81
Well it looks like MSI finally got it by providing that board with a discrete power plug for the audio section. Actually they've copied asus in quite a few areas with it except they forgot a usb 3 motherboard header for the front panel connector. I wish they'd omit the m2 connector on the z97 boards for the shear fact that it forces you to disable pcie slots in order to make use of the m2 slot.
Where do you see a power plug? It looks like the system fan header is right next to the audio section.

Edit: I see it now - that section looks like the networking. The audio is below the shield. They're using a Wolfson 8471, which is one of the best DACs out there (I believe I put an 8472 in my DAC) - adding the shielding should help some, too.
 
Last edited:

ASK THE COMMUNITY