<LongChair>
if so do all 3 values need to be changed ?
lkcl has joined #linux-rockchip
<LongChair>
ayaka : i'm heading to lunch, so in case you are still around i'll list a few more questions :) (you might be sleeping when i get back here)
<LongChair>
1 - please confirm if the link above is the vcodec freqs and which one should be changed (1st, 2nd all 3)
<LongChair>
2 - is that vcodec patch enough or do i need dtsi files changes as well ?
<LongChair>
3 - if dtsi changes are required, which ones of the values should be changed and which values can i try there (not sure of there are constraints between the frequencies)
<LongChair>
4 - which are the freq that android is using and which lead to HEVC 10 @80 mbits ? is that 500 Mhz. I see also 600 Mhz in the list of freqs in kernel ...
<LongChair>
i'll be back in a few ... 40 mins or so
lkcl has quit [Ping timeout: 264 seconds]
wadim_ has quit [Remote host closed the connection]
lvrp16 has quit [Ping timeout: 252 seconds]
lvrp16 has joined #linux-rockchip
paulk-blaze has joined #linux-rockchip
wzyy2 has joined #linux-rockchip
fungi has quit [Ping timeout: 264 seconds]
fungi has joined #linux-rockchip
nighty has joined #linux-rockchip
wadim_ has joined #linux-rockchip
_hramrach has joined #linux-rockchip
paulk-blaze has quit [Ping timeout: 264 seconds]
paulk-blaze has joined #linux-rockchip
wzyy2 has quit [Ping timeout: 264 seconds]
wzyy2 has joined #linux-rockchip
<rperier>
mmind00: I have found my email issue, I think
<rperier>
"Return-Path: <rperier@rperier-laptop>"
<rperier>
:)
<rperier>
I will resend my series I think
cnxsoft has quit [Quit: cnxsoft]
wzyy2 has quit [Read error: Connection reset by peer]
<LongChair1>
not sure what that aclk and hclk error is about
<phh>
aclk/hclk are normal, but the "failed on clk_get pd_hevc" isn't
<LongChair1>
hmmm
<LongChair1>
seems to be related to the same id as the one ayaka told me to tweak
<LongChair1>
getting same thing with 500M, i'll try to go back to 400M, maybe there is something else wrong
<phh>
in vpu_get_clk, I don't understand why it doesn't retrieve aclk_vcodec when ID_HEVC
<phh>
ok, it looks like the rate is fixed in DTS
<phh>
LongChair1: in hevc_service of rk3288.dtsi, assigned-clock-rates, ACLK_HEVC is the clock you want
<phh>
I have no clue whether hclk_hevc/sclk_hevc_core/sclk_hevc_cabac must change...
<LongChair1>
i came back to 400M
<LongChair1>
still having the same error message in log
<phh>
yup, leave the code that way
<phh>
just change dts
<phh>
I'm trying to look at original android kernel code
<phh>
unless it's done in userland, aclk is 400MHz, hclk is 200MHz, hevc_cabac is 300MHz, hevc_core is 300MHz... compared to 300/100/300/300 in 4.4 kernel
<phh>
LongChair1: so I'd suggest to change hclk actually
<LongChair1>
i'll try to change with the suggested values in the comment above
<LongChair1>
seen the comment in the file ?
<phh>
true, the comment mentions 500 aclk, 125 hclk
<LongChair1>
* The 4K hevc would also work well with 500/125/300/300
<phh>
perhaps there is something about those in the TRM...
<phh>
ah no, there is no doc about HEVC in the trm
<LongChair1>
i patches the dts .... but i don't see much changes either
<LongChair1>
we're probably missing something :)
<phh>
do you do a precise fps measurement?
<LongChair1>
not really ... i use ao=null / vo =null ... and just check the amount of framedrops i'm getting
<LongChair1>
so that should be pure decoding
<phh>
could you -benchmark?
<LongChair1>
is there such an option ?
<phh>
at the time of mplayer, there was
<LongChair1>
i don't see such options ... but there are so many :)
<phh>
so I see --untimed so that it doesn't sync to anything
<phh>
but I don't see anything to display the rendering fps :s
beeno has quit [Ping timeout: 255 seconds]
<LongChair1>
yeah the only thing i can see is the amount of lag that video and audio are getting
<LongChair1>
but it's more likely playing half the pace it should