Sound and Windows 10

[Funkerwolf 5]

Hello everyone, I would like to clarify why my editing in the article on sound is constantly canceled?

https://pcgamingwiki.com/wiki/Glossary:Sound

 

 

{{ii}} For clear and accurate sound on any mode, you must always use Windows 10. The audio subsystem has been significantly reworked. And include Windows Sonic. {{Refurl|url=http://www.synthtopia.com/content/2015/07/29/windows-10-promises-better-audio-midi-performance-easier-app-development/|title=Windows 10 Promises Better Audio & MIDI Performance, Easier App Development|date=}} Windows 7 and 8.1 spoil the sound and put the interference, since the presence of DRM-mechanism since Vista and introducing WASAPI. What was not in Windows XP.

 

 

Here is the evidence on the pictures below about the difference in sound on the spectrograms.

As you can probably notice, that on Windows 10 samples are more identical to the original, rather than on Windows 7.

 

Used materials in this tests

Acer Aspire 5333-P462G25Mikk with onboard Realtek Sound Card

Creative SoundBlasterX G1 USB Sound Card

Windows 7 64Bit Home Premium with updates on August 2018

Windows 10 64Bit Home with updates on August 2018

Drivers from Realtek and Creative for this systems, include intel chipsets and etc.

All sound enhancements and other effects disabled, max sound quality with 80% volume.

 

Test sound from Kaspersky AntiVirus tool

 

for capture through Audacity from line input channel

AsRock H110M-HDV r3.0 with onboard Realtek Sound Card

Windows 10 64Bit 1803 Pro with updates on August 2018

 

Source sound

Windows 10 on Realtek

Windows 10 on Creative G1

Windows 7 on Realtek

Windows 7 on Creative G1

 

The peculiarity of this behavior in terms of sound in Windows 7 was caused by the introduction of DRM at the development stage of Windows Vista, which is also easily confirmed by reading the development literature.

 

To carry out this test, recording through the line input via the built-in sound card with the help of Audacity was used and then the results were transferred to the schedule through the Spek.

The conditions were the same (Volume 80% output and input) and the same player Foobar2000 on default settings (no asio/openal and ks with wasapi exclusive for output) was used.

 

For very talented people, I note that the sound system has remained the same since Windows Vista, but has been significantly reworked: https://www.head-fi.org/threads/windows-10-has-better-sound-quality-than-windows-7-on-my-computer-not-necessarily-yours.789224/

Especially it is important for the old and not-so games that use DirectSound.

And for those who like to listen to music, and especially in a high quality, because windose 10 does not force you to use all sorts of workarounds crutches mentioned above.

[Aemony 40]

I think it's mostly fine, although a minor rewrite would be good:

 

Current:

 

 

{{ii}} For clear and accurate sound on any mode, you must always use Windows 10. The audio subsystem has been significantly reworked. And include Windows Sonic. {{Refurl|url=http://www.synthtopia.com/content/2015/07/29/windows-10-promises-better-audio-midi-performance-easier-app-development/|title=Windows 10 Promises Better Audio & MIDI Performance, Easier App Development|date=}} Windows 7 and 8.1 spoil the sound and put the interference, since the presence of DRM-mechanism since Vista and introducing WASAPI. What was not in Windows XP.

 

 

Rewritten:

 

 

{{ii}} Windows 10 is recommended over Vista, 7, and 8.x for clear and accurate sound due to improvements made in the audio subsystem, including a decrease in audio latency.{{Refurl|url=http://www.synthtopia.com/content/2015/07/29/windows-10-promises-better-audio-midi-performance-easier-app-development/|title=Windows 10 Promises Better Audio & MIDI Performance, Easier App Development|date=2018-12-01}}{{Refurl|url=https://community.pcgamingwiki.com/topic/4054-sound-and-windows-10/#entry11404|title=PCGamingWiki Forum - Sound and Windows 10|date=2018-12-01}}

 

 

Leaving the DRM and introduction of WASAPI aspect out of it because it isn't really _that_ important, and it is still sorta covered by the inclusion of a reference to this thread of yours.

 

Please tell me what you think :)

[Funkerwolf 5]

Thank you, of course, especially considering my level of English, but I am afraid this edit will also be canceled.

But it’s not just the latency that are limited.
This user https://pcgamingwiki.com/wiki/User:Suicide_machine has for some reason thought himself to be a moderator and thinks that he knows everything better than others in this regard, although I cited above the evidence, which themselves are easily verified with 2 computers.

https://pcgamingwiki.com/w/index.php?title=Glossary:Sound&oldid=677812

[Andytizer 137]

I think the rewrite that Aemony wrote will be sufficient.

 

I also believe that there's several issues with the Sound page, it perhaps needs to be subdivided into different OS in the future as the page rather misses the point (who is it for, what problems does it solve, etc.)

 

Thank you, of course, especially considering my level of English, but I am afraid this edit will also be canceled.

But it’s not just the latency that are limited.
This user https://pcgamingwiki.com/wiki/User:Suicide_machine has for some reason thought himself to be a moderator and thinks that he knows everything better than others in this regard, although I cited above the evidence, which themselves are easily verified with 2 computers.

https://pcgamingwiki.com/w/index.php?title=Glossary:Sound&oldid=677812

[Aemony 40]

Thank you, of course, especially considering my level of English, but I am afraid this edit will also be canceled.

But it’s not just the latency that are limited.

This user https://pcgamingwiki.com/wiki/User:Suicide_machine has for some reason thought himself to be a moderator and thinks that he knows everything better than others in this regard, although I cited above the evidence, which themselves are easily verified with 2 computers.

https://pcgamingwiki.com/w/index.php?title=Glossary:Sound&oldid=677812

 

Sui is a PCGW moderator. See the https://pcgamingwiki.com/wiki/PCGamingWiki:Staff page for confirmation (or just visit the Discord channel).

 

That said, the reason I included the latency is because I believe it has the highest likeliness of translating over to PC gamers. Generally speaking, audio improvements is not something a lot of gamers really understand unless they have knowledge or interest in audiophile or professional audio setups or equipment. Latency, however, is something that translates well due to its importance for PC gamers in pretty much everything (monitors, input, connectivity, etc).

 

 

I think the rewrite that Aemony wrote will be sufficient.

 

I also believe that there's several issues with the Sound page, it perhaps needs to be subdivided into different OS in the future as the page rather misses the point (who is it for, what problems does it solve, etc.)

 

Much like the rest of the Glossary pages, it could use an introduction and general information about sound in video games etc. A bunch of these glossary pages don't really have a "point" beyond being a dumpster used for stuff semi-related to the name of the page.

[Funkerwolf 5]

Yes, for sure, I somehow completely missed this moment with delays for myself, but I myself had such an incident when for some reason the sound came to Windows 7 in some games a little later, especially in network shooters. True it was a long time, but the fact. Also, thanks for helping to understand the situation, I hope this information will not be removed from the article.

[Mirh 95]

Windows 10 also has (supports) DRM for audio, and in no way I'm seeing how that would apply to analog outputs nevertheless.
Why in the world would you be so sure about it responsible for quality degradation?
For as much as I know it may be as well you are using different drivers.

EDIT: or the fact different Windows versions come with different APOs enabled by default.
 
Then I must admit I had initially underestimated W10 improvements.
Every application should indeed save off some millisecond, by merit of a reworked audio engine.
 
In addition to that, querying and selection of lower buffer sizes should also be possible.
But I'd be wary about greatly underlining this, since it requires for software to be updated with IAudioClient3, and for driver to support DEVPKEY_KsAudio_PacketSize_Constraints

My hats off for the testing then.

Edited December 2, 2018 by Mirh

[Funkerwolf 5]

No, the drivers are the same absolutely and the conditions were all the same for 7 and 10. In the first message of this topic I left a link where it was clearly stated what exactly they had done. And again, everything is easily rechecked by yourself.  And if I understand correctly with my level of English, it is possible that hardware acceleration and processing on hardware has returned to Windows 10, maybe partially. Perhaps the hardware acceleration was still in Windows 8, because when I used such devices, there was a tick in the sound settings that turned it on. But still this does not mean that the same EAX came back with Windows 10, since the API was broken starting with Windows Vista.

 

Also during testing all sound improvers were turned off and the maximum sampling quality was set to avoid distortion of the results, however in Windows 7 tests it lost in sound quality. Basically, I started this idea even when I was fond of Linux and noted for myself the exceptional clarity of sound, later with the arrival of Windows 10, I also noticed a high quality of sound that was comparable again with linux, but once I tried to return to Windows 7, I found for myself that the sound somehow sounds like through a cotton pad, which is not in 10. The same thing happened with Windows 8.1, the sound there is also disgusting for me personally just. And it was absolutely not important what exactly I was listening to, whether it be a game, be it a video on the Internet or an ordinary mp3, I am silent about the flac.And starting to study some material, everything turned out that Windows 7 really spoiled the sound, which made it necessary to crutch, i.e. use alternative audio source outputs of the OpenAL type, but this only worked mainly with audio players. But with the advent of Windows 10, such a need has disappeared and you could safely enjoy the sound in any application, and not just through a specially tuned audio player for my music library on the computer. https://habr.com/post/115092/ Actually the very article that explains the reason for the bad sound in windows 7. (need translation from russian language) But what was most interesting was when I asked the same comrades (audiophiles) to conduct an experiment on sound quality difference, they sent me far away and generally said that these differences just seem to you, since in their opinion the sound system remained the same as me and This test led me to the test itself. But of course, the strangest thing is that the sound quality in Windows 7 and 8.1 was distorted by the independence of the sound card and headphones, which at that time were the most usual Panasonic RP-HX350. While I now have practically studio headphones SuperLux HD681.

[Mirh 95]

And if I understand correctly with my level of English, it is possible that hardware acceleration and processing on hardware has returned to Windows 10, maybe partially.

Please, really, it seems like you are going all over the place with buzzwords.

"Hardware offload" is a thing again starting from ~W8. But putting aside this should be mostly targeted at power saving needs IIRC (I had written it already somewhere).. could we please stop to fetishize the word hardware? There's nothing intrinsically better (and EAX is also supported in OpenAL games for the records)

 

Windows 7 discrepancies also still doesn't make sense tbh.

If you say ASIO/OpenAL/WASAPI exclusive mode are good, then it's certainly not the driver, and it must be indeed something the OS does.

But I'm pretty on the relative sure side microsoft wouldn't be so dumb to screw up even the simplest of the plain test cases...

 

And please take notice, every "test article" before 2012 is mot since that's when they released an important fix.

There seems to be truth over the mixing being kind of forced though.. Even though I'm not sure anything of that could still explain *noise*.

 

Could you check any whatever combination of bit depth and sampling rate please?

 

p.s. W10 doesn't seem flawless either.

[Funkerwolf 5]

Unfortunately, I no longer have the opportunity to conduct more tests, since this requires a second PC. I also read somewhere in the article that they fixed the incorrect work of the resampler, however this did not help much, since I conducted tests with the image of windows 7 which had all the updates until August 2018. And all that I wanted to show with these tests is that windows 10 does not require any special alterations for less-quality sound against a windows 7 background. As with most usage scenarios (in the same games), and for listening to music.

 

If, of course, you have free time, try to do comparative testing yourself, playing the sound on one laptop and recording from the desktop via the line-in cable using the same 3.5mm jack connectors. (It is worth noting in this place that a lack of grounding during recording can make a significant error. So I personally advise during this test, disconnect the laptop from charging.) Of course, such a test may have some errors, but I assure you that the picture will differ dramatically in Vista/7/8.1 and 10 with XP. What actually I brought above in this topic in the tests.

 

 

 

 

 

 

Just while I was temporarily given a laptop for repair (reinstall windows and dust cleaning), there was an opportunity to test the difference in sound for so long, if I heard it, it turned into a difference in graphs. In normal mode, without resorting to using the same output via Openal and ASIO with WASAPI Exclusive, KS. (Of course, which in Windows 7 also return the sound quality, but only in separate and specially customized programs.)

 

P.S. But after reading the material with my level of English, I concluded for myself that most likely this change affected the sound quality in Windows 10. Updated WASAPI to support RAW mode so you don't have to use exclusive mode and tie up the device. (We did this in 8.1 in prep for Windows 10 as I recall) https://www.reddit.com/r/Windows10/comments/50ty05/improved_audio_in_windows_10/d775w2e/?sh=866ba8dc&st=JP63ISJY

Also, what was interesting, before the release of Windows 10, interesting articles surfaced, which claimed that dolby itself would be engaged in sound. https://www.ixbt.com/news/ht/188909

[Funkerwolf 5]

 

On 12/1/2018 at 10:49 PM, Andytizer said:

 

Hello everyone again, after some time after the publication of the comparison tests for sound quality in Windows 7 and 10, this laptop returned to me again and I decided to conduct better tests that finally show a significant difference.

 

Used soft: RightMark Audio Analyzer

 

See for yourself:

 

Frequency Response Graph 24Bit/48KHz

Frequency Response Graph 24Bit/192KHz

Stereo Crosstalk:

THD:

Noise level:

 

[Expack3 52]

13 hours ago, Funkerwolf said:

 

Hello everyone again, after some time after the publication of the comparison tests for sound quality in Windows 7 and 10, this laptop returned to me again and I decided to conduct better tests that finally show a significant difference.

 

Used soft: RightMark Audio Analyzer

 

See for yourself:

  Reveal hidden contents

Frequency Response Graph 24Bit/48KHz

Frequency Response Graph 24Bit/192KHz

Stereo Crosstalk:

THD:

Noise level:

 

Could you please explain the charts and their differences, Funkerwolf? I'm not an audio quality expert, and I doubt others on this forum are, either.

[Funkerwolf 5]

Sorry for the long answer, we had a problem with the input to the forum from another place, but not the essence.  Just say this: the more direct the frequency response line and the lower the noise value with the mixing of stereo channels, and with less distortion the better. And here you do not need to be an audiophile to understand this, and even more so to feel, it will be noticeable immediately when comparing the two systems in real life by ear. And unfortunately, this situation will occur with any sound card.

 

 

[Suicide machine 31]

ZZZzzZZZzzzZZzzzzzzz
 

Just let Windows 7 die in peace with it having pretty much a killdate set up... since it's pretty much the only way Microsoft can "convince" people nowadays.

Even if it's all true, difference is really not that "huge" and making a note "upgrade to Windows 10" is not a solution, either.

[Mirh 95]

Ehrm.. My own limited tests picture another situation.

It's not super fair to compare dedicated soundcards to a Realtek (even though, this fared relatively good here) but benchmarks made already under Vista shows Creative has more to share with the latter category. So.. I'm wondering how much this is up to the driver, and how much actually the OS. Are you sure all enhancements are disabled? Are the outputs all set to stereo? And could you try normal 16-bit @ 44.1/48 khz?

Then I'm also no psychologist, but I feel like you really seem *hoping* too much into clueless explanations (no, Dolby just team up for multi-channel endeavors)

ALC887--->Xonar DX

ALC887-->ALC887

(full disclosure: I didn't post other results, because noise in the garbled mess of cables I have behind my computer is killer)

[Funkerwolf 5]

6 hours ago, Mirh said:

Ehrm.. My own limited tests picture another situation.

It's not super fair to compare dedicated soundcards to a Realtek (even though, this fared relatively good here) but benchmarks made already under Vista shows Creative has more to share with the latter category. So.. I'm wondering how much this is up to the driver, and how much actually the OS. Are you sure all enhancements are disabled? Are the outputs all set to stereo? And could you try normal 16-bit @ 44.1/48 khz?

Then I'm also no psychologist, but I feel like you really seem *hoping* too much into clueless explanations (no, Dolby just team up for multi-channel endeavors) 

 

  Hide contents

ALC887--->Xonar DX

ALC887-->ALC887

(full disclosure: I didn't post other results, because noise in the garbled mess of cables I have behind my computer is killer)

 

Its lol. 16 bit mode should not be used as it adds noise (dither) for smoothing (quantization feature associated with to reproduce a material of lower bit depth on a higher bit depth of the sound card).  Try it yourself to listen to something trite in the 16-bit sound card mode and in 24-bit (but this does not apply to the source of the sound, music or other recordings there ... ). In moments of pause and transition from silence to loud sounds, there will be such background noise similar to telephone.  Specifically now, of course, I already have no opportunity to retest this all again. But I think the other day I will repeat the experiments with another laptop. In general, for the purity of the experiment, I would put windows 7 on my PC, where xonar dx is installed, but it’s hard to install this rudimentary on current platforms (kaby lake). Of course, deviations are possible that the drivers have some kind of fixes for the seven and other magic, but in my case the drivers were the same everywhere with the same conditions.

[Mirh 95]

You are really presenting just too.. odd interpretations.

First of all, I'm not even sure how physically sound cards could be "full" 24-bit, considering I'm not aware of hardware even approaching the complete 144dB dynamic range.

Second Windows automatically outputs 16-bit (even if 24 bit is selected) if the source is 16-bit. Legitly.

Third, and most of all, there are people with thousands and thousands and thousands of dollars of equipment, that still couldn't find differences from the lowest 16/44.

https://pcgamingwiki.com/wiki/Glossary_talk:Sound#x16_bit_is_just_good_2534

[Aemony 40]

On 2/4/2019 at 1:46 PM, Mirh said:

Second Windows automatically outputs 16-bit (even if 24 bit is selected) if the source is 16-bit. Legitly. 

I don't think this is really the case any longer, based on my testing with my Hi-Res Audio certified DAC and headset.

That DAC have a visual indicator that showcases what sample rate and bit depth, and in shared mode it will _always_ output as the selected format according to what I've selected in the Advanced tab of the output device (24 bit, 96 000 Hz). In exclusive mode through WASAPI it'll output whatever format is configured in the player, regardless of the source itself (tested and verified using Foobar2k and the foo_out_wasapi plugin.