Which is the best-sounding Realtek driver so far?

[Dr. Dro]

I strongly recommend people check out Alan's universal audio driver here on TPU, it works flawlessly for me even on a laptop that has AMD, NVIDIA and Realtek audio endpoints all at once

AAF Optimus DCH Audio Modded Driver for Windows 10/11 - For ALL HDAUDIO Enumerator Chips

ALL MEMBERS - Before posting please be aware of the following : https://www.techpowerup.com/forums/members/aaf-optimus.170292/#profile-post-22235 Dear TechPowerUp members. I would like to inform you hereby that I will no longer be continuing the Realtek driver modding projects due to lack of...

www.techpowerup.com

It has lots of DSP and APO options available.

There is also another battleground, in there they also debating about required samples so one digital storage oscilloscope this to accurately reconstruct a sine wave over it screen.

Anyway, my own battleground ended today after of me wasting again several hours in research, it appears that Realtek it loved Win7 64bit up to the version R2.81
Realtek HD Audio Manager: Loads successfully
Realtek CPL loads at control panel
Realtek icon loads successfully at taskbar, mixer appears too.

The talk about ultimate sound fidelity this never ends, when I am feel thirsty, I just use my Japanese (90s) heavy weight of sound system, and my special 2x150W pair of three way drives these designed and made in Holland.
When and if the Europeans will return at sound speakers development, then our youth will have back the appropriate tools so them to listen quality.

R2.82 has been available for almost 6 years now and should work decently, the problem is that you're still running Windows 7, which limits your useful range of audio processing software due to the obsolete OS.

 

[bug]

Are you using Windows 11 or Windows 10? Because my Xonar DGX broke once Windows update updated to Windows 11, no matter how many times I would uninstall and reinstall the drivers. So I ended up with the Creative Sound Blaster Z SE. But in order for that one to work with Windows 11 I had to install an older driver software. Now that works great.

One would think that Windows 11 and Windows 10 are driver similar but they are not. Micro$oft messed that part up quite heavily.

Judging by this: https://www.asus.com/uk/supportonly/xonar dgx/helpdesk_download/ it would seem Asus hasn't bothered with support for that card since 10 years ago. I doubt this one's on Microsoft.

 

[Super XP]

Judging by this: https://www.asus.com/uk/supportonly/xonar dgx/helpdesk_download/ it would seem Asus hasn't bothered with support for that card since 10 years ago. I doubt this one's on Microsoft.

I think both are to be blamed for this issue. Microsoft's problem is that Windows 11 mostly not compatible with Windows 10. , because the card and the drivers worked perfect with Windows 10.

 

[ThrashZone]

Judging by this: https://www.asus.com/uk/supportonly/xonar dgx/helpdesk_download/ it would seem Asus hasn't bothered with support for that card since 10 years ago. I doubt this one's on Microsoft.

Hi,
I blame ms seeing they insisted realtek console be installed through ms store and not normal driver install.

 

[bug]

I think both are to be blamed for this issue. Microsoft's problem is that Windows 11 mostly not compatible with Windows 10. , because the card and the drivers worked perfect with Windows 10.

Being as old as they are, they are probably WinXP drivers repackaged and Win11 pulled the plug on that.
It happens sometimes, I also have a webcam that Windows won't recognize anymore. Works perfectly fine on Linux tho.

 

[Kodehawa]

The 1220 Codec has actually Impressed me.

It did, in fact, impress me too when I got it now.

 

[Super XP]

Being as old as they are, they are probably WinXP drivers repackaged and Win11 pulled the plug on that.
It happens sometimes, I also have a webcam that Windows won't recognize anymore. Works perfectly fine on Linux tho.

You are right, but I did some research on ASUS cards and they don't apparently play nice with Windows 11. My Sound Blaster is doing fine and I notice a huge difference in sound clarity.

 

[Operandi]

You're splitting hairs with that highlighted statement. And it's worth remembering that a theory has lots of evidence supporting it, ie it's not an unverified hypothesis or idea that someone can simply choose to disagree with. Therefore, if someone "doesn't believe" in it, they have no credibility, especially after it's been explained to them.

The finer points of sampling, DAC topology etc all flow from this root theorem.

The Nyquist theorem isn't zero sum notion in relation to how its used in practice in digital audio. It describes the bare minimum digital resolution you need to recreate the analog signal to avoid aliasing artifacts. They key phrase to take away form this concept is "minimum", "minimum" is the reason its a "theorem" and not a "law", its not disagreeing with it or agreeing, its understanding what it is and how it works in digital audio reconstruction filters.

There is no accepted resolution and sample rate to perfectly reconstruct the analog input signal, even Redbook audio saw fit to resample frequencies that are barely audible over what the Nyquist theorem would dictate to be sufficient.

 

[qubit]

The Nyquist theorem isn't zero sum notion in relation to how its used in practice in digital audio. It describes the bare minimum digital resolution you need to recreate the analog signal to avoid aliasing artifacts. They key phrase to take away form this concept is "minimum", "minimum" is the reason its a "theorem" and not a "law", its not disagreeing with it or agreeing, its understanding what it is and how it works in digital audio reconstruction filters.

There is no accepted resolution and sample rate to perfectly reconstruct the analog input signal, even Redbook audio saw fit to resample frequencies that are barely audible over what the Nyquist theorem would dictate to be sufficient.

I have no idea what you're disagreeing about here, as it makes no sense.

All I said to the other member is that they can't claim to not disbelieve the Nyquist theorem as to do so is to lose all credibility, which you know I'm right about. How can you disagree with that and then go on a long attempt at explaining it and still splitting hairs between law and theorem? Your apparent "difference" is wrong, too and the whole thing makes no sense. Are you trying to say that it's ok to disagree with a theorem, but not a law? If so, that's incredibly stupid.

 

[Ferather]

@Kodehawa, indeed.

 

[bug]

The Nyquist theorem isn't zero sum notion in relation to how its used in practice in digital audio. It describes the bare minimum digital resolution you need to recreate the analog signal to avoid aliasing artifacts. They key phrase to take away form this concept is "minimum", "minimum" is the reason its a "theorem" and not a "law", its not disagreeing with it or agreeing, its understanding what it is and how it works in digital audio reconstruction filters.

There is no accepted resolution and sample rate to perfectly reconstruct the analog input signal, even Redbook audio saw fit to resample frequencies that are barely audible over what the Nyquist theorem would dictate to be sufficient.

Yeah, no, your understanding is completely off.
Here's the description of Shannon's Theorem: https://en.wikipedia.org/wiki/Nyquist–Shannon_sampling_theorem
It defines the minimum frequency needed to perfectly reconstruct the analog signal. In this context, "minimum" means adding extra samples is simply superfluous.

 

[qubit]

Yeah, no, your understanding is completely off.

Glad that wasn't just my impression, please see my response in post 59.

 

[Operandi]

I have no idea what you're disagreeing about here, as it makes no sense.

All I said to the other member is that they can't claim to not disbelieve the Nyquist theorem as to do so is to lose all credibility, which you know I'm right about. How can you disagree with that and then go on a long attempt at explaining it and still splitting hairs between law and theorem? Your apparent "difference" is wrong, too and the whole thing makes no sense. Are you trying to say that it's ok to disagree with a theorem, but not a law? If so, that's incredibly stupid.

I'm not sure how else to say the same thing other than to state that there is a difference describing and understanding how mathematical concept like the Nyquist theorem works and its implementation in real world applications. Thats pretty much what a theorem is, a logical proof, not a practical one.

It defines the minimum frequency needed to perfectly reconstruct the analog signal. In this context, "minimum" means adding extra samples is simply superfluous.

There is no perfect when it comes to reconstructing an analog signal with digital samples. Nyquist describes the logical concept but in practical applications there will always be limitations when put into practice in the real world. The limitation and solution depends on the application but it always comes down to the reconstruction filter.

In relation to audio thats what different oversampling approaches are all attempting to address. Its also the whole point behind HD audio; nobody can hear the frequencies, or the dynamic range HD audio brings to the table, the idea is to push the artifacts and distortion so far past the audible band they become irrelevant. Its analogous to having a tweeter with a frequency response into the 30-40Khz region, its not that there is content there you are concerned about its that you are pushing the mechanical breakup modes well past the audible domain.

 

[bug]

There is no perfect when it comes to reconstructing an analog signal with digital samples. Nyquist describes the logical concept but in practical applications there will always be limitations when put into practice in the real world. The limitation and solution depends on the application but it always comes down to the reconstruction filter.

In relation to audio thats what different oversampling approaches are all attempting to address. Its also the whole point behind HD audio; nobody can hear the frequencies, or the dynamic range HD audio brings to the table, the idea is to push the artifacts and distortion so far past the audible band they become irrelevant. Its analogous to having a tweeter with a frequency response into the 30-40Khz region, its not that there is content there you are concerned about its that you are pushing the mechanical breakup modes well past the audible domain.

You keep misunderstanding.
Here's a better explanation: https://www.allaboutcircuits.com/te...hannon-theorem-understanding-sampled-systems/
That theorem is math. Once proven, there is no argument anymore. What you are talking about is engineering and its limits (imperfections). Raising the sampling frequency will not help you with that.

 

[qubit]

I'm not sure how else to say the same thing other than to state that there is a difference describing and understanding how mathematical concept like the Nyquist theorem works and its implementation in real world applications. Thats pretty much what a theorem is, a logical proof, not a practical one.

Look, once again, I simply said to the other user that to disbelieve the Nyquist theory is to lose all credibility. This cannot be argued with, yet you're trying to counter my point with post after post of irrelevant stuff about it and splitting hairs between theory and law, also irrelevant.

If you want a discussion about the Nyquist theory, start a thread about and I'll engage you there. Can't say fairer than that. However, I can't see you actually doing it, but instead continuing the same thing here. One, two, three...

 

[bug]

Look, once again, I simply said to the other user that to disbelieve the Nyquist theory is to lose all credibility. This cannot be argued with, yet you're trying to counter my point with post after post of irrelevant stuff about it and splitting hairs between theory and law, also irrelevant.

If you want a discussion about the Nyquist theory, start a thread about and I'll engage you there. Can't say fairer than that. However, I can't see you actually doing it, but instead continuing the same thing here. One, two, three...

I don't think there ever was a question of believing or not in Nyquist's theorem. I think the problem is understanding the theorem.
When people read about the minimum sampling rate, they interpret this as a baseline. However, that could not be further from the truth. In math (geometry) we have other similar theorems that are much simpler to understand, but deal with essentially the same problem: two points will define a line and three points will define a plane. Nyquist's theorem defines the same thing for a sine wave. Now, two points define a line unambiguously. You can use three, four or a dozen points, the line will not be defined any more precise. Same for a plane, same for a sinusoid.

TL;DR Nyquist's theorem does not say "use this many samples as a baseline, more for better results". It says "you must use this many samples in order not to lose any info; use less, info starts to go away".

 

[Operandi]

Look, once again, I simply said to the other user that to disbelieve the Nyquist theory is to lose all credibility. This cannot be argued with, yet you're trying to counter my point with post after post of irrelevant stuff about it and splitting hairs between theory and law, also irrelevant.

If you want a discussion about the Nyquist theory, start a thread about and I'll engage you there. Can't say fairer than that. However, I can't see you actually doing it, but instead continuing the same thing here. One, two, three...

If someone doesn't believe in proven math thats not something I'm going to get involved in, but you quoted my post so if I'm not meant to respond to it I don't know whats going on.

If you point is the Nyquist theorem is correct and not refutable yeah I agree and arguing it is foolish. What I'm talking about is the Nyquist theorem in relation to digital audio formats and DACs, all of which are based on principles of Nyquist - Shannon. In order to get accurate performance you need to sample at least (aka more than) 2x the input frequency, thats documented history of all digital audio. Pretty much every field that uses Nyquist is full of similar examples of implementing Nyquist in real world applications and understanding or at least acknowledging the implications of the theory in practice isn't splitting hairs or irrelevant in my book.

You keep misunderstanding.
Here's a better explanation: https://www.allaboutcircuits.com/te...hannon-theorem-understanding-sampled-systems/
That theorem is math. Once proven, there is no argument anymore. What you are talking about is engineering and its limits (imperfections). Raising the sampling frequency will not help you with that.

I've read a bunch of articles at that level (that one probably included), and you are right I am talking about the implementation of Nyquist, and the engineering of products (DACs in this case) that use its principles. In terms of audio it all comes down to the performance of the reconstruction filter and I don't work in digital signal processing and I'm not a math major so I can't speak to the specifics of the filter design but resampling the input frequency more than 2x does help, see Wikipedia 44,100 Hz.

In addition, signals must be low-pass filtered before sampling to avoid aliasing. While an ideal low-pass filter would perfectly pass frequencies below 20 kHz (without attenuating them) and perfectly cut off frequencies above 20 kHz, such an ideal filter is theoretically and practically impossible to implement as it is noncausal, so in practice a transition band is necessary, where frequencies are partly attenuated. The wider this transition band is, the easier and more economical it is to make an anti-aliasing filter. The 44.1 kHz sampling frequency allows for a 2.05 kHz transition band.

 

[bug]

I've read a bunch of articles at that level (that one probably included), and you are right I am talking about the engineering and implementation of Nyquist. In terms of audio it all comes down to the performance of the reconstruction filter and I don't work in digital signal processing and I'm not a math major so I can't speak to the specifics of the filter design but resampling the input frequency more than 2x does help, see Wikipedia 44,100 Hz.

Now we're getting somewhere (starting to speak the same language). Next step, note the "before sampling" in your quote. That is very important. It tells us this (sampling above 44 kHz) is important when you covert analog to digital (i.e. in a studio) to make sure the right info is encoded in the digital form. But for those of us that simply consume the tracks in digital format, 44 kHz is where it's at.

 

[Operandi]

Now we're getting somewhere (starting to speak the same language). Next step, note the "before sampling" in your quote. That is very important. It tells us this (sampling above 44 kHz) is important when you covert analog to digital (i.e. in a studio) to make sure the right info is encoded in the digital form. But for those of us that simply consume the tracks in digital format, 44 kHz is where it's at.

Agreed. Unless you get into HD audio (which I'm not advocating for one way or the other), and there is still argument for benefits of the format in relation to Nyquist.

Fun fact, apparently the 44.1Khz spec is fairly abstract and was chosen because it was compatible with digital audio stored on cassette (VHS?) tapes that were used as the masters.

 

[mclaren85]

The one bundled with SRS Premium sound. I'm very certain about that.

 

[bug]

Agreed. Unless you get into HD audio (which I'm not advocating for one way or the other), and there is still argument for benefits of the format in relation to Nyquist.

Fun fact, apparently the 44.1Khz spec is fairly abstract and was chosen because it was compatible with digital audio stored on cassette (VHS?) tapes that were used as the masters.

Yes, 44.1 kHz encodes all the info in a signal up to 22 kHz. Which is about 2kHz higher than the best human ear can distinguish, but I don't mind having a bit of extra info in there. You know, for those claiming they can hear better than their dogs

 

[emanresu]

Not all Realtek are created the same!
This is what I mean by different drivers sounding different. Also there is something to the Audyssey speaker tuning folder created in ProgramData (at least with HDA).

Regarding Nyquist and such..
We mustn't forget overtones.
Even though a .wav can look exactly the same to us waveform-wise, it doesn't mean they sound the same. A lot of different factors we cannot see on a 2D picture.