[jamesjames]

Thank you for that. I listened to some of the Vitruoso sample tracks, and they were very interesting - more towards the "expansive" style, where the overall width increases (compared to traditional CF decreasing the width). But depth-wise, the sound stayed within the line running between my ears. My preference is to bring the sound stage forward in front of me. Also, listening to the classical pieces, I felt that the more expansive presentation came at the cost of image focus, with a slightly out-of-phase feeling.

This was on my laptop directly driving mid-level headphones and it was overall a good experience, so it would be great to try on my main rig. Sadly, my roon server is linux-based, so I won't be able investigate this any further. Based on what I'd heard, this gave some nice benefits over my DAVE's CF, but also some downsides, which is why it's so important to hear each CF solution yourself - you simply cannot rely on another person's opinion on this topic. Based on what I've read, I'd expect @GoldenSound to much prefer Virtuoso over DAVE's CF (I wonder how the Wandla DAC's implementation compares to Virtuoso?) , but right now, I still prefer DAVE's CF. There's no right or wrong answer here.

The impressions of @TheAttorney re the Virtuoso sample tracks are very close to my own with all the digital approaches I've heard (I haven't used Virtuoso or the Wandla - but have heard many others). For whatever reason, with the acoustic music I listen to, I find the analogue CF I've tried more effective in putting me at some distance from the performance space. As @TheAttorney has pointed out, this actually involves some decrease in the impression of width. Far from finding this detrimental (in reducing the sense of being 'inside' the music), I find it an improvement in terms of focus and the avoiding the 'slightly out-of-phase feeling' (a good way to put it I think). I fully accept my goal here may not be shared by many - it's specific to the kind of music I listen to and the sense of realism I'm looking for. So I couldn't agree more that listening to these various implementations is critical for working out individual preferences. I'm increasingly convinced that there's a very wide range of preferred playback presentation indeed - and many of us aren't looking for remotely the same thing.

 

[GoldenSound]

Based on what I've read, I'd expect @GoldenSound to much prefer Virtuoso over DAVE's CF (I wonder how the Wandla DAC's implementation compares to Virtuoso?)

I've not tried virtuoso though the Wandla spatial enhancement works quite differently.

Virtuoso and other binaural translation tools like BACCHDSP's solution for headphones are very cool, but are mostly beneficial with binaural or surround sound source material. The Wandla spatial enhancement is not intended to make everything sound as though it was binaural as I don't feel there is a good way to do that for most stereo content. But rather to provide a moderate but not drastic/party-trick-esque improvement to the spatial presentation and localisation info within existing stereo tracks.

For me those other tools are a neat thing to have if you have suitable source material, but with most regular tracks can be very much hit or miss and can get quite wonky if they are trying to extrapolate info that simply was not present in the recording.

I'm quite excited for the future of spatial audio, because as stuff like atmos recordings become more common, there are some cool things you can do especially if your headphones have head tracking such as airpods or the audeze maxwell, but the approach for spatial audio recordings or binaural recordings inherently has to be very different to normal stereo recordings and personally I feel that doing 'too much' to stereo recordings more often than not ends up with a bit of a weird result

 

[MrHaelscheir]

The impressions of @TheAttorney re the Virtuoso sample tracks are very close to my own with all the digital approaches I've heard (I haven't used Virtuoso or the Wandla - but have heard many others). For whatever reason, with the acoustic music I listen to, I find the analogue CF I've tried more effective in putting me at some distance from the performance space. As @TheAttorney has pointed out, this actually involves some decrease in the impression of width. Far from finding this detrimental (in reducing the sense of being 'inside' the music), I find it an improvement in terms of focus and the avoiding the 'slightly out-of-phase feeling' (a good way to put it I think). I fully accept my goal here may not be shared by many - it's specific to the kind of music I listen to and the sense of realism I'm looking for. So I couldn't agree more that listening to these various implementations is critical for working out individual preferences. I'm increasingly convinced that there's a very wide range of preferred playback presentation indeed - and many of us aren't looking for remotely the same thing.

If ever you are able or have not done so yet, I recommend trying out the foobar2000, foo_record, and bs2b combination I had described, at least for Windows.

1. Install https://www.foobar2000.org/.
2. Install https://vb-audio.com/Cable/VirtualCables.htm.
3. Go to Control Panel -> Sound. Under "Playback", find "CABLE Input" and under Properties -> Advanced set the sample rate to that of your streaming application's or other source; I use "24 bit, 44100 Hz (Studio Quality)" with Idagio. Then Under "Recording", find "CABLE Output" and under Properties -> Advanced select the same bit depth and sample rate.
4. Follow https://vb-audio.com/Cable/VBCABLE_SystemSettings.pdf to set the internal sampling rate to match or be a multiple of the ones you set at the endpoints. Conversion from 44.1 kHz to an internal 48 kHz sample rate will cause audible and measurable distortion.
5. Install https://foobar.hyv.fi/?view=foo_record.
6. In foobar2000, under File -> Preferences -> Output select your playback device. It should suffice to set the buffer length to 50 ms for minimum latency unless you come to later hear issues.
7. Under File -> Preferences -> Tools -> Recording, select "CABLE Output" as the device and set the sample rate and bit depth to the same values you had set in Step 3.
8. In Windows, set the playback device to "CABLE Input".
9. Under File -> Add Location, enter "record://", then press the play button on the foobar2000 interface.
10. Play audio through your streaming application on Windows and confirm that you can hear it through foobar2000; the small spectrum bar graph window to the right of the stop and play controls should react to the music. There will be some latency depending on the chosen buffer length.
11. Install https://bs2b.sourceforge.net/.
12. In foobar2000, under File -> Preferences -> DSP Manager, on the right, find the "Bauer stereophonic-to-binaural DSP" and click the '+' button to add it to the list of Active DSPs. You will be able to adjust the crossfeed parameters by clicking on the ellipses button.
13. "Hear the difference."

Possibly better would be to at least try the default HRTF of an actual binaural decoder:

1. To get started, install https://www.reaper.fm/ for the free 60-day trial. Make sure to under the installation options under "Optional functionality" enable ReaRoute.
2. On the left grey box, right click and click "Insert new track".
   
   
   .
3. Press the red "record" circle on the left to activate the track, and press the little speaker icon to its right until it looks like so where it will say "Record Monitoring: ON". Then on the lower right input dropdown list of the track select "Input: Stereo -> ReaRoute 1/ ReaRoute 2".
   
   
4. Click on the upper right part of the window that is in square brackets or Options -> Preferences on the upper left to open the preferences. Go to Audio -> Device and set the "Audio system" to "Dummy Audio", but set the default sample rate to the desired sample rate.
5. In your track, click "ROUTE", then under "Add new hardware output..." select "ReaRoute 1 / ReaRoute 2".
   
   
6. Install https://vb-audio.com/Voicemeeter/banana.htm. I don't recall trying the most basic version, but you at least need A1 and A2. I recommend reopening REAPER first to ensure that the ASIO connection is alive, and then open VoiceMeeter; do this any time that VoiceMeeter appears to stop working, and such in addition to closing the donationware popup.
7. On the upper right, click on "A1" and select "ASIO: ReaRoute ASIO".
   
   
8. Under A2, select the MME version of your playback device; selecting the WDM seems to give VoiceMeeter exclusive control over that device and prevents Equalizer APO from directly EQing the input to that device.
9. Click Menu -> Restart Audio Engine to detect your playback device.
10. Set the "VIRTUAL INPUTS" in the previous image to output to A1 like shown, then set "HARDWARE INPUT 1" (which should show "A1 ASIO input (1 + 2)) on the left to output to A2 like shown.
11. Got to Menu -> System Settings / Options further down, and set the preferred main sample rate to match what you set in REAPER.
    
    
12. Set your Windows audio to output to VoiceMeeter Input (VB-Audio VoiceMeeter VAIO).
    
    
13. Play some PC audio and confirm that VoiceMeeter's "VIRTUAL INPUTS" bars are reacting, REAPER's track bars are reacting, VoiceMeeter's "HARDWARE INPUT 1" bars are reacting, and that you can hear the sound. Effectively, you are outputting PC audio into VoiceMeeter which outputs that into device A1 which sends said data into the REAPER DAW via ReaRoute ASIO. The Track in REAPER was configured to take in that audio and play it through the DSP that we are about to set up. That output is then routed back into ReaRoute which conveys the processed signal to "HARDWARE INPUT 1" which finally outputs it to your playback device. Step 3 is required for REAPER to output audio back into VoiceMeeter.
14. Install https://leomccormack.github.io/sparta-site/ or a binaural decoder or choice like APL Virtuoso (free trial) or the IEM plugins; I discuss possible issues with the latter two in https://www.head-fi.org/threads/rec...-virtualization.890719/page-121#post-18027627 (post #1,812); said post also describes issues with how CroPaC handles certain sharp transients. You will need to go to REAPER Preferences, scroll down to Plug-ins -> VST, ensure that the path to your installed plug-in's DLL files or other (the high-level VST or other plugin folder) is in the list of paths, then click Re-scan -> Re-scan paths for new/modified plug-ins.
15. Click "FX" on the lower left of the track, then click "Add" on the lower left of the FX window, then add your desired binaural decoder VST. For this example, we will add sparta_ambiRoomSim and cropac_binaural. You may optionally add ReaEQ and Oscilloscope Meter to facilitate maximizing the dynamic range of the output, and Volume/Pan Smoother v5 for use with checking on the imagine coherence of pink noise across the stereo field.
    
    
16. You want the plugins to be in the exact same order as on the left of the window below. First copy the settings for sparta_ambiRoomSim below. Uncheck "Enable Image Sources" for a clearer sound, though you can experiment with different source placements, room dimensions, and wall attenuation coefficients, but note that this is CPU intensive and may incur buffer underflow noises.
    
    
17. Set the track to use four track channels for 1st order ambisonic rendering.
    
    
18. Duplicate the settings below for cropac_binaural. You may optionally look at HRTF/SOFA file database files online to find one that better matches your ears. I used https://www.earfish.eu/ to obtain my personalized HRTF, though they are slow to respond. You may need to play pink noise or some other centered test sound and adjust the yaw and pitch on the lower right until it sounds centered before you. You can also have fun with those sliders to see how well a given HRTF preset pans the sound source around you. There are all kinds of fancy mixing tricks you can do with sparta_ambiRoomSim and other plugins.
    
    
19. Optional dynamic range maximization step: The track below happens to max the digital dynamic range without sounding clipped.
    
    
    
    You will need to play this clip and adjust the digital preamp gain on the lower right of the ReaEQ plugin as seen below. For the default HRIR, +17 dB seemed to work.
    
    
    
    
    
    Then while playing that track, you will need to look at the Oscilloscope Meter reading (I dragged the "length" downward to 5.00s; make sure that the "range" is +0.0 dB which represents the maximum digital value. You can look on the lower left for the peak values and for whether the output signal is clipping. Adjust the gain in ReaEQ accordingly.
    
    
    
    
    
    
20. "Enjoy".
21. One way to test the quality of a binaural renderer is to play pink noise from a generator of your choice and then use Volume/Pan Smoother v5 (a Pan Law of 4 dB yields decent volume matching of the extreme pans with the centered pan) to move that pink noise left and right, listening carefully for whether certain frequency bands are moving away from the center line or the expected position.
    
    
    
    
22. The default HRIR happens to work well enough for my ears, though calibrating the tonality to match speakers is a more complicated matter discussed in https://www.head-fi.org/threads/rec...-virtualization.890719/page-121#post-18027627 (post #1,812). By default, a "DF-compensated" SOFA file will be mostly flat and should retain the main tonal characteristics of your headphones. Otherwise, I'd say there are definite improvements to sound coherence, tonality, and sense of distance from going through the trouble of actually having your HRTF measured and doing a threshold of hearing EQ against a neutral speaker reference.

@GoldenSound Should you have the time, I would be curious as to what you think of the imaging/soundstaging of the aforementioned CroPaC binaural decoder compared to the "holography" you have described of the Holo Audio May and Bliss which I am currently considering. For me, CroPaC works great with virtually any stereo tracks that weren't mixed to have intentional HRTF manipulations. Symphonies can be presented with huge scale before me, albeit like looking at a super-ultrawide monitor a meter in front of me; that might be remedied with further-distance HRTF measurements. What I do admit is that I don't get much depth and layering queues on the stereo line virtually imaged before me unless the recording had done something to simulate such like with distant trumpets, its being a 1D line before me with the only "height" being the perceptual bloom or bass content among others, or where the HRTF renderer is actually failing to be perfectly coherent. I suppose in the ideal case of perfect speakers in an anechoic chamber, you really would expect just a perfect line of sound sources between the two channels, any additional dimensions requiring more as in surround sound or ambisonics.

 

[jamesjames]

If ever you are able or have not done so yet, I recommend trying out the foobar2000, foo_record, and bs2b combination I had described, at least for Windows.

1. Install https://www.foobar2000.org/.
2. Install https://vb-audio.com/Cable/VirtualCables.htm.
3. Go to Control Panel -> Sound. Under "Playback", find "CABLE Input" and under Properties -> Advanced set the sample rate to that of your streaming application's or other source; I use "24 bit, 44100 Hz (Studio Quality)" with Idagio. Then Under "Recording", find "CABLE Output" and under Properties -> Advanced select the same bit depth and sample rate.
4. Follow https://vb-audio.com/Cable/VBCABLE_SystemSettings.pdf to set the internal sampling rate to match or be a multiple of the ones you set at the endpoints. Conversion from 44.1 kHz to an internal 48 kHz sample rate will cause audible and measurable distortion.
5. Install https://foobar.hyv.fi/?view=foo_record.
6. In foobar2000, under File -> Preferences -> Output select your playback device. It should suffice to set the buffer length to 50 ms for minimum latency unless you come to later hear issues.
7. Under File -> Preferences -> Tools -> Recording, select "CABLE Output" as the device and set the sample rate and bit depth to the same values you had set in Step 3.
8. In Windows, set the playback device to "CABLE Input".
9. Under File -> Add Location, enter "record://", then press the play button on the foobar2000 interface.
10. Play audio through your streaming application on Windows and confirm that you can hear it through foobar2000; the small spectrum bar graph window to the right of the stop and play controls should react to the music. There will be some latency depending on the chosen buffer length.
11. Install https://bs2b.sourceforge.net/.
12. In foobar2000, under File -> Preferences -> DSP Manager, on the right, find the "Bauer stereophonic-to-binaural DSP" and click the '+' button to add it to the list of Active DSPs. You will be able to adjust the crossfeed parameters by clicking on the ellipses button.
13. "Hear the difference."

Possibly better would be to at least try the default HRTF of an actual binaural decoder:

1. To get started, install https://www.reaper.fm/ for the free 60-day trial. Make sure to under the installation options under "Optional functionality" enable ReaRoute.
2. On the left grey box, right click and click "Insert new track".
   .
3. Press the red "record" circle on the left to activate the track, and press the little speaker icon to its right until it looks like so where it will say "Record Monitoring: ON". Then on the lower right input dropdown list of the track select "Input: Stereo -> ReaRoute 1/ ReaRoute 2".
   
4. Click on the upper right part of the window that is in square brackets or Options -> Preferences on the upper left to open the preferences. Go to Audio -> Device and set the "Audio system" to "Dummy Audio", but set the default sample rate to the desired sample rate.
5. In your track, click "ROUTE", then under "Add new hardware output..." select "ReaRoute 1 / ReaRoute 2".
   
6. Install https://vb-audio.com/Voicemeeter/banana.htm. I don't recall trying the most basic version, but you at least need A1 and A2. I recommend reopening REAPER first to ensure that the ASIO connection is alive, and then open VoiceMeeter; do this any time that VoiceMeeter appears to stop working, and such in addition to closing the donationware popup.
7. On the upper right, click on "A1" and select "ASIO: ReaRoute ASIO".
   
8. Under A2, select the MME version of your playback device; selecting the WDM seems to give VoiceMeeter exclusive control over that device and prevents Equalizer APO from directly EQing the input to that device.
9. Click Menu -> Restart Audio Engine to detect your playback device.
10. Set the "VIRTUAL INPUTS" in the previous image to output to A1 like shown, then set "HARDWARE INPUT 1" (which should show "A1 ASIO input (1 + 2)) on the left to output to A2 like shown.
11. Got to Menu -> System Settings / Options further down, and set the preferred main sample rate to match what you set in REAPER.
    
12. Set your Windows audio to output to VoiceMeeter Input (VB-Audio VoiceMeeter VAIO).
    
13. Play some PC audio and confirm that VoiceMeeter's "VIRTUAL INPUTS" bars are reacting, REAPER's track bars are reacting, VoiceMeeter's "HARDWARE INPUT 1" bars are reacting, and that you can hear the sound. Effectively, you are outputting PC audio into VoiceMeeter which outputs that into device A1 which sends said data into the REAPER DAW via ReaRoute ASIO. The Track in REAPER was configured to take in that audio and play it through the DSP that we are about to set up. That output is then routed back into ReaRoute which conveys the processed signal to "HARDWARE INPUT 1" which finally outputs it to your playback device. Step 3 is required for REAPER to output audio back into VoiceMeeter.
14. Install https://leomccormack.github.io/sparta-site/ or a binaural decoder or choice like APL Virtuoso (free trial) or the IEM plugins; I discuss possible issues with the latter two in https://www.head-fi.org/threads/rec...-virtualization.890719/page-121#post-18027627 (post #1,812); said post also describes issues with how CroPaC handles certain sharp transients. You will need to go to REAPER Preferences, scroll down to Plug-ins -> VST, ensure that the path to your installed plug-in's DLL files or other (the high-level VST or other plugin folder) is in the list of paths, then click Re-scan -> Re-scan paths for new/modified plug-ins.
15. Click "FX" on the lower left of the track, then click "Add" on the lower left of the FX window, then add your desired binaural decoder VST. For this example, we will add sparta_ambiRoomSim and cropac_binaural. You may optionally add ReaEQ and Oscilloscope Meter to facilitate maximizing the dynamic range of the output, and Volume/Pan Smoother v5 for use with checking on the imagine coherence of pink noise across the stereo field.
    
16. You want the plugins to be in the exact same order as on the left of the window below. First copy the settings for sparta_ambiRoomSim below. Uncheck "Enable Image Sources" for a clearer sound, though you can experiment with different source placements, room dimensions, and wall attenuation coefficients, but note that this is CPU intensive and may incur buffer underflow noises.
    
17. Set the track to use four track channels for 1st order ambisonic rendering.
    
18. Duplicate the settings below for cropac_binaural. You may optionally look at HRTF/SOFA file database files online to find one that better matches your ears. I used https://www.earfish.eu/ to obtain my personalized HRTF, though they are slow to respond. You may need to play pink noise or some other centered test sound and adjust the yaw and pitch on the lower right until it sounds centered before you. You can also have fun with those sliders to see how well a given HRTF preset pans the sound source around you. There are all kinds of fancy mixing tricks you can do with sparta_ambiRoomSim and other plugins.
    
19. Optional dynamic range maximization step: The track below happens to max the digital dynamic range without sounding clipped.
    
    
    
    You will need to play this clip and adjust the digital preamp gain on the lower right of the ReaEQ plugin as seen below. For the default HRIR, +17 dB seemed to work.
    
    
    
    Then while playing that track, you will need to look at the Oscilloscope Meter reading (I dragged the "length" downward to 5.00s; make sure that the "range" is +0.0 dB which represents the maximum digital value. You can look on the lower left for the peak values and for whether the output signal is clipping. Adjust the gain in ReaEQ accordingly.
    
    
    
    
20. "Enjoy".
21. One way to test the quality of a binaural renderer is to play pink noise from a generator of your choice and then use Volume/Pan Smoother v5 (a Pan Law of 4 dB yields decent volume matching of the extreme pans with the centered pan) to move that pink noise left and right, listening carefully for whether certain frequency bands are moving away from the center line or the expected position.
    
    
22. The default HRIR happens to work well enough for my ears, though calibrating the tonality to match speakers is a more complicated matter discussed in https://www.head-fi.org/threads/rec...-virtualization.890719/page-121#post-18027627 (post #1,812). By default, a "DF-compensated" SOFA file will be mostly flat and should retain the main tonal characteristics of your headphones. Otherwise, I'd say there are definite improvements to sound coherence, tonality, and sense of distance from going through the trouble of actually having your HRTF measured and doing a threshold of hearing EQ against a neutral speaker reference.

@GoldenSound Should you have the time, I would be curious as to what you think of the imaging/soundstaging of the aforementioned CroPaC binaural decoder compared to the "holography" you have described of the Holo Audio May and Bliss which I am currently considering. For me, CroPaC works great with virtually any stereo tracks that weren't mixed to have intentional HRTF manipulations. Symphonies can be presented with huge scale before me, albeit like looking at a super-ultrawide monitor a meter in front of me; that might be remedied with further-distance HRTF measurements. What I do admit is that I don't get much depth and layering queues on the stereo line virtually imaged before me unless the recording had done something to simulate such like with distant trumpets, its being a 1D line before me with the only "height" being the perceptual bloom or bass content among others, or where the HRTF renderer is actually failing to be perfectly coherent. I suppose in the ideal case of perfect speakers in an anechoic chamber, you really would expect just a perfect line of sound sources between the two channels, any additional dimensions requiring more as in surround sound or ambisonics.

Many thanks indeed for taking the time with this - greatly appreciated. I've often used foobar in the past - including the general crossfeed settings - and experimented with different combinations. But I certainly haven't followed the approach you've described here. I would like to try it one day - the issue will be finding the time and overcoming my relatively limited technical capabilities! As I think I've mentioned above (somewhere ...) I find the analogue crossfeed implementations I've listed to be very effective and simple means of improving my experience of listening to the music I like. That said, I accept they won't appeal to everyone - other contributors have made that abundantly clear!! And I'm not looking to convince anyone of the merits of crossfeed (and particularly the analogue approach) - simply to describe why it appeals to me. I feel the value of sites like this is to allow us to describe our experiences with enough detail and rigour to allow other readers who think they might share our interests and tastes to decide whether there might be different approaches worth trying. Obviously, I think analogue crossfeed is worth trying - and I've tried to describe what sorts of music and presentation appeal to me. I should add - I think I've said this above - that I'm not wishing to suggest there may not be 'spatial' settings that would appeal more strongly even to me - it's just that I haven't encountered them.

 

[Ghoostknight]

I guess you use Windows, for linux there is easyeffects which has a great EQ and crossfeed implemenation, easyeffects works similarly as EqualizerAPO just with much better sound quality imo, you avoid windows resampling (which is probably the biggest L of windows) and the possibly crappy equalizerAPO implementation, tho i never clearly figured out if the bad performance im getting comes from windows resampling alone or a combination of equalizerAPO and windows resampling since there are no real alternatives

i havent looked back to windows since pipewire/easyeffects are well supported under linux now, pipewire lets you choose different resampling qualitys... its pretty good on highest resampling setting, so far just HQPlayer was really able to beat it

 

[jamesjames]

I guess you use Windows, for linux there is easyeffects which has a great EQ and crossfeed implemenation, easyeffects works similarly as EqualizerAPO just with much better sound quality imo, you avoid windows resampling (which is probably the biggest L of windows) and the possibly crappy equalizerAPO implementation, tho i never clearly figured out if the bad performance im getting comes from windows resampling alone or a combination of equalizerAPO and windows resampling since there are no real alternatives

i havent looked back to windows since pipewire/easyeffects are well supported under linux now, pipewire lets you choose different resampling qualitys... its pretty good on highest resampling setting, so far just HQPlayer was really able to beat it

Thanks - I'll look into this.

 

[gregorio]

I guess you use Windows, for linux there is easyeffects which has a great EQ and crossfeed implemenation, easyeffects works similarly as EqualizerAPO just with much better sound quality imo, you avoid windows resampling (which is probably the biggest L of windows) and the possibly crappy equalizerAPO implementation …

That’s interesting, do you have any reliable evidence of that? I believe there was an issue with resampling in an old windows version, to the point of audibility, but that was way back in Windows 95 (or maybe it was as recently as Vista, I don’t recall exactly). So I’d be interested to see if it’s an audible issue again in more current versions.

G

 

[schs]

I've always been against altering the sound with "gimmicky" applications, but I tried the Phonitor crossfeed application and am now hooked. I also have a Rebel headphone amp, which I love and has a very different sound from the Phonitor, but I really miss the crossfeed when listening to the Rebel. ...

Same here. When I first tried the Phonitor and used the crossfeed I was very pleased with the sound. And I liked the fact that I can make small adjustments right away with the different knobs. Very handy.

 

[Ghoostknight]

That’s interesting, do you have any reliable evidence of that? I believe there was an issue with resampling in an old windows version, to the point of audibility, but that was way back in Windows 95 (or maybe it was as recently as Vista, I don’t recall exactly). So I’d be interested to see if it’s an audible issue again in more current versions.

G

i havent heared windows since around 2 years ago the last time, it was Windows 11

you kinda have to trust me on this one (or try it yourself), its not like windows audio is "broken", just that better resampling methods sound a bit better... and windows is on the bottom end

the thing is, even youtube can sound better, its not much better tho keep in mind that you have to set pipewire to the highest quality which is " resample.quality = 15 "
the default value isnt much better than windows or even slightly worse than windows, i dont recall 100%

i mainly use this as a multimedia machine so i dont noticed any performance hits with the higher resampling... tho i guess it uses a few % of my intel 12400

 

[TheAttorney]

The Wandla spatial enhancement is not intended to make everything sound as though it was binaural as I don't feel there is a good way to do that for most stereo content. But rather to provide a moderate but not drastic/party-trick-esque improvement to the spatial presentation and localisation info within existing stereo tracks.

How does the Wandla spacial enhancement compare with Zaal HM1's "sound stage knob". The latter sounds very interesting, but I've not yet found anywhere in the UK where I can audition it. I recall you were very impressed with the HM1 in general.

 

[gregorio]

you kinda have to trust me on this one …

So that’s a “no” then, you don’t have any reliable evidence at all that there is or has been any sort of audible issue with Windows resampling for at least a couple of decades or so. Thx

G

 

[Ghoostknight]

So that’s a “no” then, you don’t have any reliable evidence at all that there is or has been any sort of audible issue with Windows resampling for at least a couple of decades or so. Thx

realiable evidence means measurements , you can look those up m8 but since you probably know those you probably will heavly rely on some old study..

RELIABLE evidence doesnt mean some tens of years old (probably flawed in some way) study made with whatever equipment and whatever people...

Do looki looki on the measurements and hear for yourself if you hear a difference

 

[MrHaelscheir]

i havent heared windows since around 2 years ago the last time, it was Windows 11

you kinda have to trust me on this one (or try it yourself), its not like windows audio is "broken", just that better resampling methods sound a bit better... and windows is on the bottom end

the thing is, even youtube can sound better, its not much better tho keep in mind that you have to set pipewire to the highest quality which is " resample.quality = 15 "
the default value isnt much better than windows or even slightly worse than windows, i dont recall 100%

i mainly use this as a multimedia machine so i dont noticed any performance hits with the higher resampling... tho i guess it uses a few % of my intel 12400

I this morning prior to an amp and DAC audition (as a skeptic curious if things could possibly sound "better") started a Qobuz trial. To my ears, there was absolutely no noticeable difference between Qobuz Wasapi (Exclusive) (Windows desktop app) to my FiiO K9 Pro ESS driving the Meze Elite or HE1000se with the cables listed in my signature or playing the same recording out of Idagio on MS Edge. There was likewise to my ears no sonic advantage to feeding Qobuz Wasapi into Voicemeeter and then my REAPER binaural rendering DSP versus trusting that with Idagio outputing CD quality, Windows would not have a problem feeding that into the 24-bit (probably a good idea for DSP) 44.1 kHz Voicemeeter endpoint. Things are just as textured or clean and so on pending the recording quality or the state of my hearing at a time of the day. Equalizer APO sounds transparent to my ears, solely introducing minimum-phase linear distortions consistent with purely tonal changes, the only harmonic distortion changes being related to the expected changes in driver excursion per frequency. I had in https://www.audiosciencereview.com/...n-susvara-headphone-review.50705/post-1888972 (post #1,183) shown the impressively low multi-tone distortion I could get out of an EQed Meze Elite compared to unEQed HiFiMans. But this of course won't stop you from hearing what you hear.

The only time I have heard issues with resampling was when I unwittingly left VB-Cable or Voicemeeter conducting conversions between 44.1 kHz to an internal 48 kHz, the distortion even for a single tone being audible and visible in an FFT.

 

[Ghoostknight]

I this morning prior to an amp and DAC audition (as a skeptic curious if things could possibly sound "better") started a Qobuz trial. To my ears, there was absolutely no noticeable difference between Qobuz Wasapi (Exclusive) (Windows desktop app) to my FiiO K9 Pro ESS driving the Meze Elite or HE1000se with the cables listed in my signature or playing the same recording out of Idagio on MS Edge. There was likewise to my ears no sonic advantage to feeding Qobuz Wasapi into Voicemeeter and then my REAPER binaural rendering DSP versus trusting that with Idagio outputing CD quality, Windows would not have a problem feeding that into the 24-bit (probably a good idea for DSP) 44.1 kHz Voicemeeter endpoint. Things are just as textured or clean and so on pending the recording quality or the state of my hearing at a time of the day. Equalizer APO sounds transparent to my ears, solely introducing minimum-phase linear distortions consistent with purely tonal changes, the only harmonic distortion changes being related to the expected changes in driver excursion per frequency. I had in https://www.audiosciencereview.com/...n-susvara-headphone-review.50705/post-1888972 (post #1,183) shown the impressively low multi-tone distortion I could get out of an EQed Meze Elite compared to unEQed HiFiMans. But this of course won't stop you from hearing what you hear.

The only time I have heard issues with resampling was when I unwittingly left VB-Cable or Voicemeeter conducting conversions between 44.1 kHz to an internal 48 kHz, the distortion even for a single tone being audible and visible in an FFT.

im not sure if its solely due to resampling but i always had the feeling windows just messes around more with the sound than linux, tho only pipewire (or just alsa for that matter) is great, pulseaudio is kinda on windows level again, maybe better with higher resampling settings but i think i havent messed with that at the point of using pulseaudio

it actuallty went like this....
1. using windows since idk 2008? till 2017 or so
2. using linux with pulseaudio, not for audio quality but just to get rid of windows...
3. switching back to windows because linux still has its limitations and sound quality didnt seem better at that time...
4. switching back to linux for actually testing pipewire vs windows (and switching back and forth a few times to compare) because these "windows resampling sucks" people made me curious
5. actually noticing that even youtube sounds a bit better (WITH FIREFOX... google chrome resamples everything to 48KHZ... its crap, tho in the youtube case it doesnt matter since i think youtube plays natively 48khz, where its important if you use a music streaming web player like deezer which plays in 44,1khz ... if you dont use firefox for this you get twice the resampling.... one from google chrome and one from windows in the worst case... but it depends on input and output samplerates)
6. testing upsampling to 176,4khz with pipewire, further improvement...
7. well and 2 years later and we are here and not much has changed, beside that i tried different distros, but all with pipewire since, its not only for sound quality better... it just beats pulseaudio and windows in any aspect (well maybe "userfriendly" goes still to windows...)

i should notice, that i purely speak about windows resampling, no exclusive mode etc.. i switched to pipewire because it sounds better AND works in a similar way, no strange setups that only allow one app to play etc... pipewire is king for multimedia use imo, my kind of goal was to get audio from a game, a youtube video or possible music AT THE SAME TIME and still get better quality... exclusive mode is just a bandaid regarding playing multiple streams

the thing is, these changes, like using linux/pipewire and using upsampling with the highest resample quality actually made me hear differences more easly... specially with lets say youtube comparisons of soundclips but also in general
this CLEARLY tells me that in the end windows had actually masked details and i account the resampling for it, but there might be more going on under the hood
as with some other tweaks... upsampling is one you actually expierence after longterm use better imo, A/B comparisons make it harder to notice any difference

Also i should notice... HQPlayer i tried recently and it actually defeats pipewire with the right settings, tho for general use pipewire is still better, it just works, without any strange server/endpoint setups etc..

 

[castleofargh]

realiable evidence means measurements , you can look those up m8 but since you probably know those you probably will heavly rely on some old study..

RELIABLE evidence doesnt mean some tens of years old (probably flawed in some way) study made with whatever equipment and whatever people...

Do looki looki on the measurements and hear for yourself if you hear a difference

You're the one bringing up a fidelity concern about an OS(not even one version of it, the all OS), isn't it normal to want to clearly know if it's a real matter that concerns maybe billions, or yet another audiophile anecdote of dysfunctional testing method turned into a global statement for no good reason?