Tír Na HiFi

relisten to 2.8x

sounds much better than i remember

multi-threaded 2.84 is still highest quantity in my setup. (overall quantity feels a little less than recent mqn. sometimes more. depends on frequency range...)
just like old days, it has excess(dirty)
all 3 frequencies vibrate better than most version!

2.86 sounds great too

compare to 2.8x, now i think 2.9x is too "clean"? (2.8x maybe too dirty?)
maybe need (dirty+clean)/2

i think 2.9x is missing some "dirtiness" of 2.8x, so 2.9x becomes a little hard?

lekt, maybe try learn the sound of mqn1.82/2.82/3.82...
all 3 frequencies vibrate. the treble vibrates better than all software on my win2012
all pkshan tweaks work on this version.
it doesn't negate OS jitter. doesn't seem to compromise

last year, its tune is wrong. weight is bright/harsh
this year, this version is best raw analog sound. big width supported by bass

2.82 is my favorite version for playing medium speed rock music like Scorpions!
its weight is sort of correct for drum/piano/vocal/guitar

jesuscheung wrote:...compare to 2.8x, now i think 2.9x is too "clean"? (2.8x maybe too dirty?)
maybe need (dirty+clean)/2

i think 2.9x is missing some "dirtiness" of 2.8x, so 2.9x becomes a little hard?

just only today back from weekend, tired, tired with summer...

last versions used unsigned __int64 for 64bits variables. 2.8x and 2.90.x used unsigned long long.
i know this different for SQ. C++/C compiler makes something in .exe binary file. weird. think need find reason why result of compiling difference for each time. morning one SQ, night other SQ with same code.

JC, legacy HDD/boot vs UEFI HDD/boot? with UEFI SQ better? i've full data HDD in legacy mode, now need much time for transfer to UEFI, and with installation win8.1/vs2013 think need 2 days.

lekt wrote:

jesuscheung wrote:...compare to 2.8x, now i think 2.9x is too "clean"? (2.8x maybe too dirty?)
maybe need (dirty+clean)/2

i think 2.9x is missing some "dirtiness" of 2.8x, so 2.9x becomes a little hard?

just only today back from weekend, tired, tired with summer...

last versions used unsigned __int64 for 64bits variables. 2.8x and 2.90.x used unsigned long long.
i know this different for SQ. C++/C compiler makes something in .exe binary file. weird. think need find reason why result of compiling difference for each time. morning one SQ, night other SQ with same code.

sounds like quantum computing... hmm...

night > morning??

lekt wrote: JC, legacy HDD/boot vs UEFI HDD/boot? with UEFI SQ better? i've full data HDD in legacy mode, now need much time for transfer to UEFI, and with installation win8.1/vs2013 think need 2 days.

"win8.1 with update" iso?

not tried UEFI. would be interested to know its SQ!!

think need learn UEFI, JC. people tell me that it OK for audio, better than legacy.

uploaded:
v2.96.3 256
v2.96.4 160 c1
v2.96.4 160 c2

c1 and c2 have micro difference, think c2 more correct and better balance. their softness as silk, but dirty or not?
coding with 160 more difficult but small buffer size always better (with same algorithm). hmm...

lekt wrote:think need learn UEFI, JC. people tell me that it OK for audio, better than legacy.

uploaded:
v2.96.3 256
v2.96.4 160 c1
v2.96.4 160 c2

c1 and c2 have micro difference, think c2 more correct and better balance. their softness as silk, but dirty or not?
coding with 160 more difficult but small buffer size always better (with same algorithm). hmm...

c2 > c1
i think nos has better weight

v2.96.3 256 = distortion!!

make c2 bass a little tighter?
c2 all things feel ok

uploaded 2.96.4 160 c2 MOD

lekt wrote: c1 and c2 have micro difference, think c2 more correct and better balance. their softness as silk, but dirty or not?

not dirty. clean. typical 2.9x

2.84 more more clean better better everyday

i think 2.84 was dirty because my OS/bios was incorrect

2.84 lacks emotion. only has quantity... hehe

possible to make a multi-threaded 2.90.4 or 2.90.9 or nos? just like 2.84... maybe magic will happen

lekt wrote:think need learn UEFI, JC. people tell me that it OK for audio, better than legacy.

big project. need 1 day! shall see

jesuscheung wrote:2.84 more more clean better better everyday

i think 2.84 was dirty because my OS/bios was incorrect

2.84 lacks emotion. only has quantity... hehe

possible to make a multi-threaded 2.90.4 or 2.90.9 or nos? just like 2.84... maybe magic will happen

think need 2-files model, multi-thread, fibonacci. 3 in 1. all these things very OK by theory, but i used them still not good as needed. now C native can resolve problems for them. maybe and your UEFI mode, hehe...

"v2.96.3 256 = distortion!!"
used specific memory address for buffer, made running for 4/8/16GB RAM. mistake for your RAM? seems works for goon-heaven?