Kernel Streaming is NOT a small improvement in SQ

[Dawnrazor]

Yes, I’ve made various OS tweaks and have examined detailed performance metrics pre and post change.

I’m either an anonymous guy on the internet talking out of my hat or an anonymous guy on the internet with decades experience in platform design, implementation, performance tuning, and monitoring various systems, many of which need to meet stringent operating thresholds due to regulatory compliance where failure could impact safety and/or incur significant financial penalty who finds claims made by the audio “optimization” software vendors professionally offensive. Charging people for software is what I take issue with. Making information available for free to experiment with is a different story.

is it impossible that system tuning could fix issues? No. But in those rare situations in 2019, the problem would be with the original system configuration which is where the root cause should be addressed, not with targeted OS tweaks.

Some of the Focusrite recommendations are just basic common sense (disabling system sounds for example). Some are not only not going to impact audio reproduction, they are awful recommendations that put your system and data at significant risk (disabling firewalls and anti-virus software)

Good. You don't sound biased at all....

What is the root cause by the way? I bought a new computer in 2019 and hooked up the dac downloaded programs, etc. It sounded terrible. I did the Focusrite optimizations and it sounded way better. What would cause that?

What interface are you using to your dac? Why couldn't the computer affect that?

 

[bfreedma]

Good. You don't sound biased at all....

What is the root cause by the way? I bought a new computer in 2019 and hooked up the dac downloaded programs, etc. It sounded terrible. I did the Focusrite optimizations and it sounded way better. What would cause that?

What interface are you using to your dac? Why couldn't the computer affect that?

The two most likely root causes for audio issues would be something running that is over utilizing system resources or something running that is utilizing the windows audio stack in addition to your playback software of choice. The former scenario would be identifiable via Resource Manager. The latter is easily avoided by running WASAPI in exclusive mode.

Since you didn’t mention running WASAPI, the most likely reason the focusrite recommendations improved things was by disabling something also using the Windows audio stack.

I use USB but don’t understand the question. How would disabling services and changing thread priorities impact an interface?

 

[Dawnrazor]

It's not how my mind works, it's how computers and operating systems work. That isn't the big mystery you seem to believe it is.

The two most likely root causes for audio issues would be something running that is over utilizing system resources or something running that is utilizing the windows audio stack in addition to your playback software of choice. The former scenario would be identifiable via Resource Manager. The latter is easily avoided by running WASAPI in exclusive mode.

Since you didn’t mention running WASAPI, the most likely reason the focusrite recommendations improved things was by disabling something also using the Windows audio stack.

I use USB but don’t understand the question. How would disabling services and changing thread priorities impact an interface?

Well I was using ASIO. So the windows audio stack shouldnt matter,The computer is completely dedicated to music. Thats all. Which is basically the player and the asio driver. Everything else is turned off as much as I can. Nothing uses the windows audio stack and since I disabled the windows audio process, I can't play anything else if I wanted to.

Anyhow if you have usb, the power situation of the computer will affect the cleanliness of the signal to the usb dac, and tweaking the os for audio can affect that. Heck, just adjusting the USB port power settings on all usb ports can change things. What is crazy is that audio is a real time thing and even with an internet dac, settings matter.

you never did a cicsmemory player build where the OS is gutted. Winlog is replaced with minlog, only 2-3 services run, etc. Its HOURS of tweaking and sometimes you can go too far and destroy enough where you need to start over from a fresh install. But it makes a big difference in sq. I was hoping with an ethernet dac to be able to get good sound without any of that hazzle. Well I was wrong.

 

[bfreedma]

Well I was using ASIO. So the windows audio stack shouldnt matter,The computer is completely dedicated to music. Thats all. Which is basically the player and the asio driver. Everything else is turned off as much as I can. Nothing uses the windows audio stack and since I disabled the windows audio process, I can't play anything else if I wanted to.

Anyhow if you have usb, the power situation of the computer will affect the cleanliness of the signal to the usb dac, and tweaking the os for audio can affect that. Heck, just adjusting the USB port power settings on all usb ports can change things. What is crazy is that audio is a real time thing and even with an internet dac, settings matter.

you never did a cicsmemory player build where the OS is gutted. Winlog is replaced with minlog, only 2-3 services run, etc. Its HOURS of tweaking and sometimes you can go too far and destroy enough where you need to start over from a fresh install. But it makes a big difference in sq. I was hoping with an ethernet dac to be able to get good sound without any of that hazzle. Well I was wrong.

How would adjusting the USB power settings impact audio? The only settings typically available, and the only ones focusrite lists is one that allows/disallows the OS to turn off the USB device to save power and one that enables USB selective suspend setting which aren't in any way going to have an impact during playback over the USB. It's either powered or it isn't. I'm also at a loss as to how "the power situation of the computer will affect the cleanliness of the signal to the usb dac". This is a digital signal, not an analog. And if the worry is the infamously overhyped "dirty PC power situation" resulting in some electrical element riding on the cable to the DAC's USB port, the simplest and most effective solution would be a basic powered USB hub in the middle of the chain - random changes to PC settings aren't going to be effective.

You've buying into a lot of audiophile hacks that aren't based on reality. Again, if the computer isn't resource constrained, reducing load on an already unburdened system isn't going to change anything. But if you have measurements that show a big difference in sq, feel free to post them, as big differences in sq would be blatantly obvious. Or analyse the data stream from a basic PC and your optimized/gutted/tweaked PC utilizing the same audio file and show the differential in the data.

 

[Dawnrazor]

How would adjusting the USB power settings impact audio? The only settings typically available, and the only ones focusrite lists is one that allows/disallows the OS to turn off the USB device to save power and one that enables USB selective suspend setting which aren't in any way going to have an impact during playback over the USB. It's either powered or it isn't. I'm also at a loss as to how "the power situation of the computer will affect the cleanliness of the signal to the usb dac". This is a digital signal, not an analog. And if the worry is the infamously overhyped "dirty PC power situation" resulting in some electrical element riding on the cable to the DAC's USB port, the simplest and most effective solution would be a basic powered USB hub in the middle of the chain - random changes to PC settings aren't going to be effective.

You've buying into a lot of audiophile hacks that aren't based on reality. Again, if the computer isn't resource constrained, reducing load on an already unburdened system isn't going to change anything. But if you have measurements that show a big difference in sq, feel free to post them, as big differences in sq would be blatantly obvious. Or analyse the data stream from a basic PC and your optimized/gutted/tweaked PC utilizing the same audio file and show the differential in the data.

This is a digital signal, not an analog.
This is going to be our disconnect. Please do some research. There are 2 things most people miss in computer audio...one is that in most cases the digital signal is riding on an analog signal and there is a timing element to the signal. . When you look at the problem with those 2 things factored in, changes with the os can make a difference and a bunch of small things can add up to big improvements....

 

[bfreedma]

This is a digital signal, not an analog.
This is going to be our disconnect. Please do some research. There are 2 things most people miss in computer audio...one is that in most cases the digital signal is riding on an analog signal and there is a timing element to the signal. . When you look at the problem with those 2 things factored in, changes with the os can make a difference and a bunch of small things can add up to big improvements....

Where, specifically, is anything in the PC or going to a DAC's USB port analog and changed in a consistent and predictable manner by "optimizations"? Are you suggesting that issues in the PC cause a consistent and repeatable alteration in that signal that result in SQ improvement instead of manifesting themselves as gaps, clicks, or pops? And that these changes aren't measurable? What do you think occurs audibly if/when the timing gets so bad that it would be noticeable?

Instead of speaking in generalities, show (or at least discuss) a specific aspect you believe supports your position.

 

[Andrew_WOT]

Timing was more of an issue in the days of Isochronous USB transfer, nowadays with asynchronous USB receivers this pretty much a thing of the past.

 

[Dawnrazor]

Where, specifically, is anything in the PC or going to a DAC's USB port analog and changed in a consistent and predictable manner by "optimizations"? The data is sent via voltage changes over the usb cable. Its a square wave and voltage thresholds determine what is a ONE or a ZERO. Its basically digital data transmitted on an analog signal or digital modulation. Its not sending actual ONE and ZEROs. So the power situation in the computer has a bearing on things. The processor is also a big RF source and detuning it or turning off spread spectrum can change things. We are talking millivolts I think for the data and even things like vibration and power fluctuations from fans and HDD can mess up the signal. Its one reason why SSDs are recommended and people report better sound vs a platter.

Are you suggesting that issues in the PC cause a consistent and repeatable alteration in that signal that result in SQ improvement instead of manifesting themselves as gaps, clicks, or pops? Yes in general. Electrical situations are different in every computer and some manufactures of "music computers" quit making them because even the same model mobo sounded different at different times because the parts were always changing. This makes it hard to meet your consistent and repeatable benchmark, but I think on the same computer yes you can hear a difference that is not pops and clicks. In my case the new untweaked pc sounded terrible, but much better when the optimizations were done. I think if I could take the same computer and have it boot into the untweaked os and then the tweaked you could hear a difference.

And that these changes aren't measurable? Not sure. One would think anything audible would be measurable. I don't have the time or equipment to do it.

What do you think occurs audibly if/when the timing gets so bad that it would be noticeable? It sounds edgy and harsh. Its fatiguing and lifeless. The classic "digititus" people speak of. I had a dac that sounded great. I ditched the tablet that was running it for a new mini pc. Same dac to new untweaked computer and it was horrible. I would have returned that dac based on that experience. But I had already heard how good it could sound. I did the tweaks on the focusrite site and some others and it sounded good again...it didnt make me want to stop listening. I have a network dac connected with ethernet through a switch. There should be no reason for what I heard. But I heard it and it was clear enough to make me go through the hours of tweaking and occasional screw ups and start overs...

Instead of speaking in generalities, show (or at least discuss) a specific aspect you believe supports your position.

 

[Dawnrazor]

Timing was more of an issue in the days of Isochronous USB transfer, nowadays with asynchronous USB receivers this pretty much a thing of the past.

Check this out. Here is an "interview" with Gordon Rankin who not only wrote the book on Asynchronous USB, but pretty much all the Asynch dacs liscense his tech. Asynch is not the rosy world you are describing. Just because you have an asynch dac, doesnt mean you are done. He talks about the computer and the cable making a difference, and also Thorsten talks about EMI effects and IMHO what you do with the os can affect all that as well. Kernal streaming is just a part of all of that.

https://darko.audio/2016/05/gordon-rankin-on-why-usb-audio-quality-varies/

 

[Dawnrazor]

Where, specifically, is anything in the PC or going to a DAC's USB port analog and changed in a consistent and predictable manner by "optimizations"? Are you suggesting that issues in the PC cause a consistent and repeatable alteration in that signal that result in SQ improvement instead of manifesting themselves as gaps, clicks, or pops? And that these changes aren't measurable? What do you think occurs audibly if/when the timing gets so bad that it would be noticeable?

Instead of speaking in generalities, show (or at least discuss) a specific aspect you believe supports your position.

Here is a post from an engineer who has actually built and designed usb dacs. Note that he specifically says:

"Unfortunately this can make USB DACs quite susceptible to all kinds of things that happen on computers. The player software used, what audio stack is used, kernel parameters, timing on memory chips, all kinds of things can cause subtle changes in the packet timing on the bus"

https://www.audioasylum.com/forums/pcaudio/messages/9/90278.html

Its more involved than getting a dac and plugging it to the computer....if you want the best sound from your gear.

 

[Dawnrazor]

Here is an interesting vid about why ASIO is better than Kernal Streaming and WASAPI.

 

[bfreedma]

Check this out. Here is an "interview" with Gordon Rankin who not only wrote the book on Asynchronous USB, but pretty much all the Asynch dacs liscense his tech. Asynch is not the rosy world you are describing. Just because you have an asynch dac, doesnt mean you are done. He talks about the computer and the cable making a difference, and also Thorsten talks about EMI effects and IMHO what you do with the os can affect all that as well. Kernal streaming is just a part of all of that.

https://darko.audio/2016/05/gordon-rankin-on-why-usb-audio-quality-varies/

Gordon Rankin did not "write the book" on Async USB. He was one of the first to use it in audio applications, not any technical accomplishment - he simply applied existing tech standards. The statement that "pretty much all the Asynch dacs liscense his tech" is completely false. He has worked on USB DACs for Ayre and AudioQuest, but his work is limited to the device specific USB implementation. Rankin doesn't even have a patent regarding USB, let alone any kind or async license used across the industry.

Note that his statements in the article are purely theoretical - he can't actually show an issue with running a USB hard drive on the same host, but if that's a concern, it's trivially easy to avoid.

One thing in the article I do agree with: "“A good example of this is when we transitioned from Full Speed USB to High Speed USB DACs. A lot of the really expensive USB cables from audio companies failed miserably; I doubt many of these cables were even tested for High Speed compliance.”

It feels like you're on a fishing expedition to prove your point, yet aren't reading or understanding the articles you've linked.

 

[bfreedma]

Here is an interesting vid about why ASIO is better than Kernal Streaming and WASAPI.

ASIO is certainly better than Kernal Streaming. No argument there.

What in that video lead you to believe WASAPI Exclusive is not as good as ASIO? Aside from the comment that it's limited to 24/96 which is incorrect. WASAPI supports 24/192.

 

[bfreedma]

Here is a post from an engineer who has actually built and designed usb dacs. Note that he specifically says:

"Unfortunately this can make USB DACs quite susceptible to all kinds of things that happen on computers. The player software used, what audio stack is used, kernel parameters, timing on memory chips, all kinds of things can cause subtle changes in the packet timing on the bus"

https://www.audioasylum.com/forums/pcaudio/messages/9/90278.html

Its more involved than getting a dac and plugging it to the computer....if you want the best sound from your gear.

An article from someone who works for a company that sells "solutions" to these "problems". Yet can't actually show demonstrable evidence that the "problem" impacts audio and just uses scare tactics...

 

[Dawnrazor]

Gordon Rankin did not "write the book" on Async USB. He was one of the first to use it in audio applications, not any technical accomplishment - he simply applied existing tech standards. The statement that "pretty much all the Asynch dacs liscense his tech" is completely false. He has worked on USB DACs for Ayre and AudioQuest, but his work is limited to the device specific USB implementation. Rankin doesn't even have a patent regarding USB, let alone any kind or async license used across the industry.

Note that his statements in the article are purely theoretical - he can't actually show an issue with running a USB hard drive on the same host, but if that's a concern, it's trivially easy to avoid.

One thing in the article I do agree with: "“A good example of this is when we transitioned from Full Speed USB to High Speed USB DACs. A lot of the really expensive USB cables from audio companies failed miserably; I doubt many of these cables were even tested for High Speed compliance.”

It feels like you're on a fishing expedition to prove your point, yet aren't reading or understanding the articles you've linked.

Not sure why you seem to be so eager to diminish someone. Sure he worked with Ayre and Audioquest. And he has his own company named Wavelength. But there are other companies that license his Streamlength code. Berkeley Audio, Aesthetix, ART Legato, Hallide all use the Streamlength code. And those are the ones that admit it. There are probably others that license his code instead of inventing their own. Is it the whole industry no, but its more than you acknowledge.

Maybe I don't get the article but he showed errors and said that the set up doesn't record the errors and that he has to press a button to capture them, and well I can understand the conclusion of the article easy enough: "In other words, USB audio isn’t simply a matter of bits leaving the host PC/streamer and arriving AOK at the DAC. The quality of that which sits in-between matters."

And look at the first part about the noise on the signal. Digital audio is more than just bits. That noise won't affect your printer but will your audio in most cases.