Tír Na HiFi

i feel v3.44.2 256 norel2 also have good sound, thin and mellow. it can't work for PCI card.

uploaded:
v3.46.2 256 r12d
v3.46.3 256 srv

hard to beat ses2, balance, not much softness but good vibration.
try and compare.

3.46.3 256 srv
shiny and energetic and sweet! fast, sharp, tight and clear bass. smooth and very comfy and very relax on the ear. tempo is extremely good(maybe new best)
tempo is much improved for notes of long duration. (fast notes is amazing as always)

only thing is lesser maybe height... who cares... it sounds sooo goood

no need to compare, i like this srv already.... new best for me

let me listen half more hour, and shall test the other version

---
3.46.2 256 r12d
is just as good. drier. different style. music has focus. need more testing.

prefer srv

-----
yea, srv better. srv's focus is on realism. when drum hits, it feels like drum as it is- lekt's drum at its best
r12d's focus is on music

3.46.3 256 srv maybe is the first version can play 'woman in love'?

jesuscheung wrote:3.46.3 256 srv maybe is the first version can play 'woman in love'?

yes, also think it's best, many things have improved. included in code several rewriten functions of wasapi core for testing. only bypass first layer coz bypass second layer can make player not work for PCI card.

Both versions best yet.

I like r12d best - excels at bringing out the subtleties of interplay between the musicians, better matching what I hear live, buried in other versions.

goon-heaven wrote:Both versions best yet.

I like r12d best - excels at bringing out the subtleties of interplay between the musicians, better matching what I hear live, buried in other versions.

yes! r12d has better flow. live convert.

v3.46.2 r12d code lesser, difference using of instruction for manipulation critical section object. sound more free. maybe this way is effective.

should be wary of windows update, rebuilding combase.dll, audioses.dll can make change relative addresses of internal functions, code blocks and player can be not working. OK, if this will happen then we will together make update and i'll make correction for code, not difficult.
not benefit if now use method get absolute address in dll, longer code and SQ lesser.

in win10 i just notice that youtube(html5) changes timerresolution to 0.5ms
interesting... google likes 0.5ms

uploaded v3.47 256 act

rewriten several function of audio core, too much code and file size now bigger (4.5kb), seems negative affect to sound. in my hard disk maybe it is fragmented coz size > 4kb. i think need optimize code for reducing file size.
try it, SQ better. balance have little issue again, more code and more wrong code alignment, need some correction.

lekt wrote:uploaded v3.47 256 act

rewriten several function of audio core, too much code and file size now bigger (4.5kb), seems negative affect to sound. in my hard disk maybe it is fragmented coz size > 4kb. i think need optimize code for reducing file size.
try it, SQ better. balance have little issue again, more code and more wrong code alignment, need some correction.

better again! bass kept with more softness, less earaches(compare to 3.46.3 256 srv is a bit harder than norm), more little details in vocal.