lekt player

Anything to do with computer audio, hardware, software etc.
lekt
Posts: 1102
Joined: Thu Oct 10, 2013 4:52 am

Re: lekt player

Post by lekt »

jesuscheung wrote:retuned OS based on 3.19.1 pop and html5
3.25 256 sys turns out to be the best sound, as it seems. power, rythm, focus, notes, depth... never better
v3.25 256 sys have good vocal focus, like it.

now problem in compatibility with win8.1/ws2012R2, they share the system dll architecture of api and returned rax, but syscall numbers and other returned registers are different vs win8/ws2012 (means rcx, rdx, r8, r9). need build version for win8.1, and can't use tricks of registers. am trying learn more.
User avatar
goon-heaven
Posts: 312
Joined: Sun Sep 15, 2013 9:38 pm

Re: lekt player

Post by goon-heaven »

... and win10 too ? hee-hee
lekt wrote:
sbgk wrote:...have tried ntDelayExecution instead of waitforsingleobject, but couldn't get it to work.
have read this link, but i think x64 still use interrupts, not sure better than syscall or not. see in ntdll.dll win10 many "int xxh", code of ntDelayExecution win10:

; Exported entry 310. NtDelayExecution ; Exported entry 1715. ZwDelayExecution
public ZwDelayExecution
ZwDelayExecution proc near
mov eax, 60034h ; NtDelayExecution
mov edx, offset sub_4B2FCA80
call edx ; sub_4B2FCA80
retn 8
ZwDelayExecution endp

sub_4B2FCA80 proc near
mov edx, large fs:30h
mov edx, [edx+254h]
test edx, 2
jz short loc_4B2FCA98
int 2Eh ; DOS 2+ internal - EXECUTE COMMAND
; DS:SI -> counted CR-terminated command string
retn

loc_4B2FCA98:
jmp far ptr 33h:4B2FCA9Fh
sub_4B2FCA80 endp

have not seen syscall. this "jmp far ptr 33h:4B2FCA9Fh" is secret, you need copy data from this address, think 128 bytes is enough, write into file.txt, put into global variable area of you application, disassembler this last file, and can discover what is this.
---------------------------

share to you NtDelayExecution win8/8.1/ws2012/ws2012R2:

;NtDelayExecution(IN BOOLEAN Alertable, IN PLARGE_INTEGER DelayInterval);
lea rdx, [rsp+32] ; i use some valid pointer on stack
xor r10d, r10d ; Alertable=FALSE. If TRUE, execution can break in a result of NtAlertThread call
mov qword ptr [rdx], -xxxh ;Negative value means delay relative to current
mov eax, 32h ;win8/ws2012
;mov eax, 33h ;win8.1/ws2012R2

syscall ;return rax = 0

1 millisecond = 1 million (1,000,000) nanoseconds.
1 nanosecond = 0.000001 milliseconds.
1 nanosecond = 0.000000001 seconds, ie. 1 billionth of 1 second (1 / 1,000,000,000).

be careful with DelayInterval, it's Negative value.

for win10 you may try:
mov eax, 60034h
or use int 2Eh, not sure. or get syscall number by your method.
tried use this function instead of NtWaitForSingleObject in render loop, but difficult determine DelayInterval, sound have sand, need correct calculate for 16/44100, seems 56.69ms for my 256 buffer size. do you know how calculate?
lekt
Posts: 1102
Joined: Thu Oct 10, 2013 4:52 am

Re: lekt player

Post by lekt »

uploaded v3.27 256 swt

tested different tricks with import address table of PE format, i've seen benefit. used prohibited area of stack memory (SQ always better than, interseting, but risk), maybe work for you, JC.
if it works, think i can build for win8.1/ws2012R2 later, need little remote help of goon-heaven.
in my case v3.27 256 swt very good musicality.

------------------------
and try v3.27.1 256 , file size now reduced, only 3KB, maybe better.
jesuscheung
Posts: 2491
Joined: Mon Oct 07, 2013 11:09 pm

Re: lekt player

Post by jesuscheung »

3.27x throws exception!!!
lekt
Posts: 1102
Joined: Thu Oct 10, 2013 4:52 am

Re: lekt player

Post by lekt »

jesuscheung wrote:3.27x throws exception!!!
doesn't work, think risk in api fuction pointers. OK, don't use this method.

try v3.27.2 256 , v3.27.3 256 , v3.27.4 256

prefer v3.27.4 256, most clarity sound.
jesuscheung
Posts: 2491
Joined: Mon Oct 07, 2013 11:09 pm

Re: lekt player

Post by jesuscheung »

lekt wrote:
try v3.27.2 256 , v3.27.3 256 , v3.27.4 256

prefer v3.27.4 256, most clarity sound.
3.27.2 has 'lagless' tempo here
this sound is a treat!

----
3.27.2 musicality is lesser. maybe some bass/mid/treble is missing
lekt
Posts: 1102
Joined: Thu Oct 10, 2013 4:52 am

Re: lekt player

Post by lekt »

v3.27.3 256 and v3.27.4 256 works for you?
User avatar
goon-heaven
Posts: 312
Joined: Sun Sep 15, 2013 9:38 pm

Re: lekt player

Post by goon-heaven »

Hey lekt - what is your OS? (easier to ask than plough back through millions of posts)
lekt
Posts: 1102
Joined: Thu Oct 10, 2013 4:52 am

Re: lekt player

Post by lekt »

goon-heaven wrote:Hey lekt - what is your OS? (easier to ask than plough back through millions of posts)
win8 pro. think win8 and ws2012 can compromise coz they share syscall numbers, but win8.1 and ws2012R2 use different numbers, = (number of win8) + 1. trying build good version, so then will build it for your win8.1/ws2012R2. but still not good builded, last methods seems not good as i hope.
how do you think, win8.1/ws2012R2 better than win8/ws2012 ? maybe i will install win8.1, but JC use ws2012, hehe...
jesuscheung
Posts: 2491
Joined: Mon Oct 07, 2013 11:09 pm

Re: lekt player

Post by jesuscheung »

lekt wrote:
goon-heaven wrote:Hey lekt - what is your OS? (easier to ask than plough back through millions of posts)
win8 pro. think win8 and ws2012 can compromise coz they share syscall numbers, but win8.1 and ws2012R2 use different numbers, = (number of win8) + 1. trying build good version, so then will build it for your win8.1/ws2012R2. but still not good builded, last methods seems not good as i hope.
how do you think, win8.1/ws2012R2 better than win8/ws2012 ? maybe i will install win8.1, but JC use ws2012, hehe...
my 2012 still has much room for improvement hehehe...

must we update OS i would prefer if we install the same OS
prefer non-pro version of 8.1

how about x86 8.1?

lekt wrote:v3.27.3 256 and v3.27.4 256 works for you?
yes.

here,
3.27.2 has superior tempo
3.27.4 has more power depth

3.27.3 = (3.27.4+3.27.2)/2
Post Reply