[Ghoostknight]

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?

maybe microsoft just doesnt care about fidelity and uses a crappy resampling algorithm because its light on resources... to compensate for the overall overloaded windows system... and probably thousands of gamers would scream they lost 2 FPS because they dont hear the difference with their 30$ setup ... just maybe...

dysfunctional testing method? i dont know how hard it can be to setup one linux instance and one windows instance and switch back and forth but OK

and no, i havent noticed any particular windows version, they are all on the same level of crappy i would say atleast the ones i actually updated to in my windows career, no i havent documented them..

 

[Ghoostknight]

- It often pushes things out in front of you too much without widening the stage, ending up in a sound that seems like its 'trying' to sound like speakers but distractingly different from speakers for me.

imo you should play with the gain of the crossfeed here to get it where it needs to be for you

- With music that has lots of directly centred content, this can become 'smeared' for lack of a better term, sounding as though it's a bit stretched out left to right now rather than being precisely in the centre. Certain implementations do this more or less, as different crossfeed implementations will have different approaches to phase/timing offset and frequency vs crossfeed level, but it's still a common complaint.

hmmm this is a tough one... IMO most modern music is mixed on speakers (of course checked with headphones also, but i would say the big part happens on speakers most of the time), so you might wanna reproduce "speaker-like" sound even if that means its objectively less precise but i would say its probably closer to what the engineer intended on speakers (tho if you dont like it, you dont like it, nothing wrong with that)

tho i also think there is no one crossfeed fits all, specially if you compare different speaker setups, i think you would probably need to adjust the crossfeed for each of them to get it "perfectly close"

personally i like simple implementations like cmoy but with less gain ( around -4,5db seems to be right for me )
i kinda tend to get lost with plugins that tend to offer more fine tuning options than Hz/Gain, and in the end i come back to the simple implemenations... if you wanna try crossfeed try a simple implemenation first... dont try complicated plugins and wonder why it doesnt sound right specially as a beginner imo

 

[jamesjames]

I'm repeating myself here - apologies - but I think the 'value' of crossfeed might well depend on the types of music and recording you're interested in. I listen to classical music, typically recorded to capture the ambience of the performance space. I think this is a big part of why I find crossfeed almost always improves the presentation in the ways I've outlined above. I occasionally encounter a 'centered' recording, but rarely. It's easy to test the basic idea - try it for yourself with any system you can find - I think most implementations I've encountered are good enough to give the listener an idea of what's going on. As with many matters hi-fi, it might just be the first step in a process of increasingly sophisticated experiments. We shouldn't allow the counsel of perfection to become the enemy of the good.

 

[MrHaelscheir]

imo you should play with the gain of the crossfeed here to get it where it needs to be for you


hmmm this is a tough one... IMO most modern music is mixed on speakers (of course checked with headphones also, but i would say the big part happens on speakers most of the time), so you might wanna reproduce "speaker-like" sound even if that means its objectively less precise but i would say its probably closer to what the engineer intended on speakers (tho if you dont like it, you dont like it, nothing wrong with that)

tho i also think there is no one crossfeed fits all, specially if you compare different speaker setups, i think you would probably need to adjust the crossfeed for each of them to get it "perfectly close"

personally i like simple implementations like cmoy but with less gain ( around -4,5db seems to be right for me )
i kinda tend to get lost with plugins that tend to offer more fine tuning options than Hz/Gain, and in the end i come back to the simple implemenations... if you wanna try crossfeed try a simple implemenation first... dont try complicated plugins and wonder why it doesnt sound right specially as a beginner imo

Following the procedures for setting up CroPaC in the second section of https://www.head-fi.org/threads/anyone-into-crossfeed.961533/page-6#post-18068536 (#78) gives me coherent imaging along a nigh perfect line a meter ahead of me, even with the default HRIR. With this rather more sophisticated plugin (the only "tuning" regards finding an HRTF SOFA file more accurate to your own head or EQing the whole chain to match some reference speakers' tonality), I am not hearing any stretching of individual sound sources including at the center, just an increasing of the distance between them.

For centered vocals, when moving the virtual ambisonic receiver within AmbiRoomSim closer to and further from the speakers, it simply changes in loudness (the actual perceived distance probably stays locked around the distance at which I took my HRTF measurement, though). For something like the following with flanking instruments, I can slide the virtual ambisonic receiver and perceive the outer instruments panning inward and outward without themselves being stretched.



I suppose if an actual stereo pair of mics in the recording process captured a sound source of continuous width, then the above movement of the virtual ambisonic receiver would simply stretch and compress said sound source consistent with one's moving closer to or further away from the physical continuous source, whereas if it was a single track mixed and panned to a specific position, then it will merely be moved inward or outward proportionally while remaining together.

The following is another nice test of imaging. For me, it already sounds "holographic" through even my unEQed Meze Elite through my FiiO K9 Pro ESS for the sound sources closer to the center, whereby this DSP moves the flanking sounds that are still imaging from the headphone drivers to being imaged from where you would expect the virtual speakers to be.

 

[gregorio]

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...

“Reliable evidence” means any evidence which is not unreliable, such as marketing or anecdotes. Isn’t that self evident? So it can mean objective measurements, which obviously can be looked up or you can do them yourself. Reliable evidence can also mean studies that are “tens of years old”, as by definition they can be repeated and verified but it can also mean more recent studies. Again, isn’t that obvious?

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

Why are you telling me to do what I’ve already done? Have you done measurements for yourself?

I'm repeating myself here - apologies - but I think another consideration is the type of music and recording you're listening to. I listen to classical music, typically recorded to capture the ambience of the performance space.

Yes, it’s very much the “recording you’re listening to” because of course, how it’s been mixed obviously defines the spatiality/soundstage of the recording distributed to consumers. And, with the exception of a relatively few recordings of “modern” (post WWII classical music) compositions, it is always recorded “to capture the ambience of the performance space” but it is by no means an accurate capture of that ambience. In fact by design, it is almost always deliberately intended not to be.

I find crossfeed almost always improves the presentation in the ways I've outlined above.

I find crossfeed almost always degrades the presentation and especially with most classical music recordings. Those recording it can somewhat improve IME, is typically the extreme panned pop/rock recordings from the mid 1960’s and earlier.

As with many matters hi-fi, it might just be the first step in a process of increasingly sophisticated experiments.

Indeed and that process of “increasingly sophisticated experiments” took place in the early to late 1950’s (in the case of classical recordings).

And, as with many matters hi-fi, it's useful to resist the temptation I think of allowing the counsel of perfection to become the enemy of the good.

That “counsel of perfection” only exists as a marketing gimmick in the audiophile world. Recordings of acoustic instruments cannot be perfect because neither microphones nor their positioning are perfect and as mentioned above, recordings are deliberately designed not to be accurate/perfect anyway, because subjectively good virtually always trumps perfect/accurate. It is for this reason that as close to audibly perfect playback of recordings is desirable.

G