< Back to IRCAM Forum

Spat CPU/audio optimization

Spat
#1

Hello,
First of all, thanks a lot for creating Spat! I have only started using it quite recently and it offers great tools.
I try to setup a system with quite a large number of convolutions (using spat5.conv, spat5.hoa.decoder and spat5.virtualspeakers). I want to test how far I can push my PC regarding number of virtual loudspeakers as well as ambisonics order/input signals. For that I try to understand the spat5.monitor. How is the ‘audio’ information calculated and how is it related to the CPU? Any general tricks to get it lower other than ‘less convolutions’ and ‘larger blocksizes’?

Best,
Axel

#2

Hi,

spat5.monitor is a patcher, so you can open it. You’ll see that the ‘audio’ workload is derived from Max builtin [adstatus cpu]. This estimates the load of the audio thread (while the ‘CPU’ section reports the CPU load of the whole Max application, including all threads, similar to what is reported by ‘Activity Monitor’ on macOS)

There’s no magic trick to reduce the CPU load in your use case.
Possible things to do :

  • spat5.conv~ : use less convolutions; use shorter impulse responses; increase latency (blocksize); if possible, use one spat5.conv~ with multichannel IR, rather than multiple spat5.conv~ with mono IR.
  • spat5.hoa.decoder~ : use single-band decoder (dual-band decoders are 2x more expensive); use less (virtual) speakers
  • spat5.virtualspeakers~ : use less speakers; use ‘SOS’ (second-order-sections) SOFA files (files labelled SOS12 are cheaper than SOS24; SOS24 are cheaper than SOS36, etc.)

Hope this helps.

Best,
T.

1 Like
#3

Hi guys,
I am currently trying to create an AV playback player which can play back 2 layers of audio and video. At the moment I am using 2 spat5.sfplay~ objects to playback 2 independent 41 channel audio which gets routed to a slider and then to a dac output. I am currently having issues where audio lags to the video and when I observe the CPU usage, around 80% of my CPU are used by the spat5.sfplay objects ( as I have tried to route my video processing to the GPU). I was wondering whether there would be any tips and advice to reduce the CPU usage? In addition, I was wondering on how to clear a loaded file from the object’s buffer? Many thank’s guys and much appreciated.

#4

Hi,
That seems absolutely abnormal : the spat5.sfplay~ object should cost almost nothing (for the CPU).
Testing two spat5.sfplay~ with 41 channels, the cpu load is less than 2% on my computer.
So, I believe something else must go wrong in your patch. Otherwise, please send your patcher and/or the audio files so that we could have a deeper look.

No, at the moment there is no “clear” message; what would be the use case ?

Best,
T.

#5

how do you load/play those video file? i highly recommend to use jitter and play it as a texture to the jit.world.

#6

Hello guys,
Thank’s for getting the advices. Yes I’ve set the @output_texture to 1 on the jit.movie object. I load the movie files using the usual ‘read’ message.

@tcarpent
Thank you for your help. I am attaching a link to my patch with it’s audio and video test files.

#7

First there’s is a bit of optimization in your patch. Use @mc 1 for the sfplay object. (see below)

But as said T, i doubt the problem comes from audio.

For the video itself, you should use jit.world instead of depreciated technics with jit.gl.render, and jit.gl pix to render the texture

Second, you should try playing the audio part only to identify where comes from that cpu anormal uses

then as i understand, you are playing very big resolution file (more than 4k) . Have a look in the file manager, you will find how to manage video composite using jit.gl.pix. It’s a more optimized way to handle texture based work, than videoplane.


----------begin_max5_patcher----------
4507.3oc6cs0iaajk941+JHD1GlAaG459kEXArcbFOClMiCb5I6CACDXKwVM
ioHUHoZ6LCl7aeOUQJJJwRRjRkT6fMcBZqtnHq57Um60oJ9udwMitO6yQEiB
9uB9wfat4e8hatw1jogap+6aFsH7ySSBKresQOjjktZwnaqtD7w3zjnR60v0
M9PV9hvRSKh5VVFVN8w3z4SxilVV0aTNdLUnXTphRwDBgSuMPKkikPqbDhPI
BsvzHGMFca.gLFE7O1zsYqJW2uj5VqZp7WVFU0EiFcavn6CSmOp4FimYIhr6
+ouhOpYvkGtHpLJeRTZ38I1aFYtz+9Euv7qa6IxrHpnHbdz5GaYzmsj5nmhm
EkE7efchYjAhPXDB6.hXGAhv6EhbAMJznSf9Si9DbycH+eJtb77jwKi+bfEI
FJLPTrwbLC..kjxYLg1BC.WgToYJhDSjLtPHuMPgOUNEXPNYdxDyfdUdjguw
IvHQiZOPixqgiZ7.X8iShdJJuHNKsEreynvkKa07MstECH9SY1Gj51llhSqZ
B2zTdzSwMO1MMGlC3UI.VlgsYX9YAazlmS1rn7zUw1appQX9rdLYm4RAV+hk
gSi1LWEkt9ALZyr.FakBwTg4ezDo4eDH0FnFXOlmjM8iQV3BstwrkQowoKyi
JhRKCKqoflKOK5gvUIkSdHKsrH9eZGFXXBz00endX57hFBwRCuNONLogBlmG
OKK0LHZM+W075tCHNtk13sIF62HMb4latYxD3D.bYOWr.HxUE2Glaltp0lzz
skYYIaeol6KI5gx5KuLNMcGTrLa49uXd77GOv8deFbwEG5YauRwjUoUWcBvX
TNoH7osQ6xvjjZY7se7eNLMFT3GUFWMEPPMWrRi5iESyyRR1hdqtxSNtxLfS
eZzmhmU9nsiZyL.e83kqYhF0LKOKddTQ41sUFNuX6VJJ+kJPuUSqtuVRFj8W
rLAnB6W3t5+Xx6.1fluMHuDWTV7X1mJpuq0bcsQiM1SaKl2V64VseHsnaqI8
+rcqtTetOUnXIYrlgwJvtgfvELjFjg0igljBkhSEBFUq.Um5w7sUfteaIGwd
xtpNEiV2r0rRspH+CRVq4UVZBP6EvPGFvnTzXpDQ4ZNUqUbAGrtfI7wZBWSv
RvOENgB.lhuqEGuAX7qDfEmFfOUbpx0Lbk2GpKFTfudPA4jYYP7qBVPtRXAL
hN.eA9XJbrtI.pZL+CsOBIH2jKyE4twJcRb59T2ZGVlq6FGJxVkOcMZWGOPv
1CPvbSYbZiOK+3ZEYvrbv+nWyBCcLfGvX.cgFChdNFX6cLT2XUKCLHBiGnOr
rXN3yZSzjt7LbeQdtGGAcyj5HvBIWNVa9wDyERn0LAnkGslM9HwU3vKU2Qk5
L5BwIE10r7rklvNNXH48MtJEEbrmSDDMB..MDjI6Vqu+T6OZjliwPvV2FvPi
EU+vERJgy0mRz4tiLm3wvOgHO.WzmEjGENyOPDXymokHIUoTRoRnjq4P7V.3
BOG+8hrmhiBd0SYIfH6qfwvxUkqC1ETl8pY4gepL6.wmOTPRikUfDztDqkDC
KCQpuvwmqNEbqLa97gK93hpEZ4XrVgkHEFooRkBCjakGArgyZDmV5jJ09NyU
Gfy4SY4Iypcm9UQ4gEQSllkjkC7Ql+C6CXSBgcn0bEQXRmij.+BT5nOFyB8P
LKP7n4wetIAf6iggQOENlvxx7MFnL+0ZWlZIWMTjgANPpobSDYRMVH4lXNbp
rAy8r1F4o.BKlNNI9onwyCiS+00D6SgMgD69KXRsvzGCSSiRllsJ0R6L7PYh
.TgioXNWnATQPXPTqRLzHQPw.dQUTHzV.p.KVVuxE621MaeX0hUIkw0i0h34
oFmIprZ8PRVXo4CIwEtEQoGRDcMkYxvxrIF1m36WUFs4SE6lVwmBSVEk8P2b
Gt4wmjkN+fP+Ve6EvToMA8MIWY6qFa8z6qjHmWu3wr7x92YqwTWOpUowkM4k
g8h1tRdR9OtGUYvXbV3zecnxjZPljv.uAkToDbGRXS8L3NDiAsnvTEntWST8
HC7ntLImjaNSyVrHBja1kBwiC996d8GtK3a+l69v6CBAWdd6q+ZmNQKNrSzt
rwKUUdE2BGHPLdc7TlPUGwZmCbfMxiq1xKKJyVN30X.Tv1k.43tDX2XYOKUu
Ljmo8v7xgS7f8W1yAwye1m3wXAoiDNgqGCQ9fHqocEy6zNU87OwiAajikOCD
Ol6SMefVu28M2E7C+k29MuOvkBOBZvJ7vBMtqleqyX6pwiTgM5gnwCbxymHv
ad8W+We2Gd+e+u8V+hBNM+oodCD7Javq+6u8urgx68DMDtVWM+.V0gF4DKIh
FFIx9BfDMKbQGcbTuQhTeRhay+1aRTh6Jt5wYQxyOIpgIrK3jH1mlilEWrLq
HZv1ijLG1iXUtdBj3F6QdN.XPa+W.juyYXAxZMFaodIXVlbAndkGS2Xdv6WF
kN4GLYL5OEmDM4t24NIQnCvHnP5NJkgX865QJlRNervlPqpVQi0BlDwUJBkS
zL7s15On5W3VcxV3Gw+32qWMK9jwOBly6HHw4pqB9wM0XfjwQZEhp2.cG.+n
dD+J5K+2A89SQ6pHRwbge8Mt+sAICz.yR0PUeXx3J+CRGmI6vdNQ554jOAog
yIw8pd7WZKtqgZDyIr3LZZsmUi60bIXVurVEsZeVEVhqUg8qyVkGGkOZnRaf
Qttl80Gw0lghX3Kohm27tAu3HBYWO.zrtwii6VfI8QhpRGyV12HGShRbIsss
GH5.l1zNbzmi4NT5PDd1xFZKc1VOCTMZmLWPeLrjeIUgOb1MrIqmc42H9keS
sKrcXLxqICFF2CV+shf5n+VgbXVqawRddZiXZuJpY4LlTj.5ixGtC3N.AoD4
nzXv90HFS8kCHXp6kNdQqDsBHC.Aj5B.BBupn3nfvAzQPociDixXcCJErgO3
E.CSQO2R6HFcL2rQRPJvMDoDTSAb5npbbt14CoYo38t3tWYzuOIb5GGt4Tlt
a1O4LGk.m.6aqoVeQLEfdi+HziXafw8Ojc2vSNgvgUSAU1Us.W4ULC2Di3AC
Tjw7pxiCwXc.8FRLqqdCpijYInmjmE0tuVw9H5SrgT+5YwP0znPfqPlZnDIA
ZmXJmm8jdOeqng50kYEMTBWZx6DVHDLDigDDrIeGWGB2qKx12+oXf1BxReY1
CODTLMOJJMnLqR.oo7lhSiZpqIx4VAy2ECyBAeHaQX5AheFLZ0wMEmhZfCM1
5XdXKM.U5ypwM7iAU0OHdby+OjbLvOkJ81kxH2wUyjW9fFQsypE4XAfSU9Wc
9cCOYebJqKSFwQNZDngqOumV1n5mUM3XAPub0tpvcknJu6qH5YUENF7PtpZ+
2RG9Ugxw++Bc3XtlzO4qSSINi7adk3NUA4d8F7rV7953sWKDhGB9uMG1CnW9
yKhJyyB9C+4+4ezKAuKHN7LndIkGnmATuFOV4p7z2+2lzppU6cRggnXYHHJd
pfKMmXFDfKPKTiwJFgv3JNVoUZPwjR52L0PId057QffCUgtZvKalD..NgPUL
AyT1dJQ8t4BtjP.wLAMpzmRhZvOqFfHfWncJQUNAewSJKG62zQNYYVAYxObR
EEgP20OelqbxhQ3Kg9ORGOYEXs.D6XTX5gv46U0nWyr85Mzm4zlXxhhgu19N
fQNR9a.Xj6Uu.s3WQzzhfgKNBVO5HNB9E1ECkT+Vw39MXniJNdHEtBlpCFX1
.fs8XytasPzSNhnSmWg50Zq86RB+k6AWjC9trhXy9.OHNMXir2VdLSG7dKA2
cuWHX3N.YUJnM910eCWmVJffG68Q4+5YW6CGykV27PBQGRmK54FmsWa.8Mal
spMxVyyzbbQ09o0FHkqoC3tqOVd7nnH.N7wEOrD3y90fWsdS2EvFtXoSOEbs
ncriloBF83H3u+ouT+j2pOQZyl4rZCaVc7ecfcq4d2V1dcKMlGX1YyDvz0je
X39RxXcKiFAqa0jq7cBQGdoORIWFevG9pkxc41CwQl107KwpkNnpMhbgbU5D
JyHrtaY8K.nrKrIOipLpwoHlFQt0UkbuOjR7aEWkvDvsHYeDaOEekH7ubhRy
ojFW5HyUH7yunF6KYmLwXkqxXxQtSd98xjP8hWlWfPdk5tE8AUd4i3kfu.Bk
urHJ5iCt3jkcKNYlxwdJQ44MKr9YM0iJpirtgIW9xAU4SqRevTM++or7n44f
wmYA1pTO3SwkOFDlr7wvf5PtN3Zf0+cFh.0MFLti8Mb85NPoCJYzXuiLuALX
2FYFr4Jc2stG7mNxo3IsPK99339m28DAKJcNLgaNlmlkj.weGN8QPQIXMBFv
AuJbUYl8.Fv8gGVRV1RyUVkF+yqh1bjPsEOD6bWG0drORbUHLbsySoD6tdyX
uymGDYCeIC8+wrt8j5ZYRHLaZ1Fpfynli2snkkOVGLpYlBZMJ2L0A+c5r0W.
96hogluwXAG9kzzxx0dw9UnwTR8uQeQMEaRXVmr6Pp1cUm5T7liOryX5U97u
YzbtCObdBe32Mil1qUv9dyNXvqVL8DNC5T5wLAxbBNy.e33.frOEEGM5B5vN
8vNZ5oN2MYjh77lrpm+sSnyMQGDObGVdOuaBOsJU5zN7IoHRGGr.ByQI058C
eRFoum9j16d6Sn4JDY2Sj3Zbo6IQrr8QQ7dNFhUaNFh6.88titR8ipG8i4UL
xY2QLZ0gT8A6IgO5IcOHI6n4b6n03xkum3jdzSBevNftV7C79vf6iNh1iNh3
CnqOyQ9fYvTgWWGRxr8f6wrjW5J80pmz8Q0pYS+5AdbbeXI7hBh9H2h8gZ79
LOw8hAi9XZZGDNKeV0K.Kzkuq2SOiOudl2KlSsOv29nEwTkkWBxre8M5hL4h
6ED6CqoLzUa1jptZZ0L0cyQ8jix8QOYTMxNZWI8QW0GNR4kferW87EQJjzGd
FhOXYH8w4ehOD4H8R51GQNg6kDmOjsMmEf8Q31KTkPc0TOZNk0tR9GYNP7tN
NHYNAmtdSVWOCLX9Uy.Cte9ew7BYw5U3.9grX82FyY1Szd4e01pUVDOaYVbZ
Yw5M2liBxihjN1+e87ad1TUujrrzNa+TkRMF2YeiJcbhthERGGD6c+lmMYg7
.YIL69xNaDXon2jkv2jU+7HkxNDOnjvbjTYG0bUO+hmMM0KkEVJeuSURgZLq
ygOAvU187Kh2Mk5N9hmMQwOehRs98w1VreaUi6GjnD9lnNIMsUKOvNuXrMcw
NuPr24kgsiWD16+kf8tu.rsu7qqdO.uyZRrob5LGVVeu8k37juMJcU0pcr9U
M81Hw8yeHNIoY86ZurZqWIk5U2q4ctc605yd.1nLKwmsvUsmhMz0aqmMuiKs
2Cd8MwZpuUVSATypO8jPseAVauMxl9Bg0U8fYsDYUeBZZ65hEvxz40uzpa8x
+Yzx7rkY4Mu8tGS0Me+UkYyyCmEGUUE.au1g2VyHkCWc2W+xUKQ556oMi21S
FKLKD+4MIzL.cioTyoHjcsVMqHL1BNrcvkSDN2YB2ySB9DqeaE3BxTuIKYVv
2kYVOv3Cg7sWDWSclNtQ.2SCkuk8+Lz92wZF+i00Hafgt1rllMZ258TDZrh6
eJ7DPZzuMozuB6wYyMj2uq.3ENpK6eb8KoX+MCF9YYeXV+84C+nPFv6.LI3C
QyWkDV6Bwo+zpmplTcqU2Ucwmr0rWTRjoNY2BljJgPxq1lMXkp5SLNp9vqpk
j3V2lfJXFWSgOwTBt0oEgx7BZb6ayw4BY6BqqMXVWgXaMhM0Xj8QLIN03KYz
9eTtFm316+us+dgSmtKTHLkhkkPL+TQb09g0Yvd+pxxZ+ZaIbr8iSQPzJfA3
+pfHI.ZXwAfnpc.tcRnoHwrkLF0wf3XLL38wv7S2ee+UWq+cs0WYsCer5778
+0D3yglmNBmtWFKURkuIYUzgMw6PdhZOAOsB.a13d0yu9d.9Nyw80fGgVEHb
a71VAMaoY50Amcif8NvksAO11b5mxMEGpEGws4PG9.rJrbaAJZFgEKq8mvVG
iu3e+h+O.nFoHHB
-----------end_max5_patcher-----------


----------begin_max5_patcher----------
4507.3oc6cs0iaajk941+JHD1GlAaG459kEXArcbFOClMiCb5I6CACDXKwVM
ioHUHoZ6LCl7aeOUQJJJwRRjRkT6fMcBZqtnHq57Um60oJ9udwMitO6yQEiB
9uB9wfat4e8hatw1jogap+6aFsH7ySSBKresQOjjktZwnaqtD7w3zjnR60v0
M9PV9hvRSKh5VVFVN8w3z4SxilVV0aTNdLUnXTphRwDBgSuMPKkikPqbDhPI
BsvzHGMFca.gLFE7O1zsYqJW2uj5VqZp7WVFU0EiFcavn6CSmOp4FimYIhr6
+ouhOpYvkGtHpLJeRTZ38I1aFYtz+9Euv7qa6IxrHpnHbdz5GaYzmsj5nmhm
EkE7efchYjAhPXDB6.hXGAhv6EhbAMJznSf9Si9DbycH+eJtb77jwKi+bfEI
FJLPTrwbLC..kjxYLg1BC.WgToYJhDSjLtPHuMPgOUNEXPNYdxDyfdUdjguw
IvHQiZOPixqgiZ7.X8iShdJJuHNKsEreynvkKa07MstECH9SY1Gj51llhSqZ
B2zTdzSwMO1MMGlC3UI.VlgsYX9YAazlmS1rn7zUw1appQX9rdLYm4RAV+hk
gSi1LWEkt9ALZyr.FakBwTg4ezDo4eDH0FnFXOlmjM8iQV3BstwrkQowoKyi
JhRKCKqoflKOK5gvUIkSdHKsrH9eZGFXXBz00endX57hFBwRCuNONLogBlmG
OKK0LHZM+W075tCHNtk13sIF62HMb4latYxD3D.bYOWr.HxUE2Glaltp0lzz
skYYIaeol6KI5gx5KuLNMcGTrLa49uXd77GOv8deFbwEG5YauRwjUoUWcBvX
TNoH7osQ6xvjjZY7se7eNLMFT3GUFWMEPPMWrRi5iESyyRR1hdqtxSNtxLfS
eZzmhmU9nsiZyL.e83kqYhF0LKOKddTQ41sUFNuX6VJJ+kJPuUSqtuVRFj8W
rLAnB6W3t5+Xx6.1fluMHuDWTV7X1mJpuq0bcsQiM1SaKl2V64VseHsnaqI8
+rcqtTetOUnXIYrlgwJvtgfvELjFjg0igljBkhSEBFUq.Um5w7sUfteaIGwd
xtpNEiV2r0rRspH+CRVq4UVZBP6EvPGFvnTzXpDQ4ZNUqUbAGrtfI7wZBWSv
RvOENgB.lhuqEGuAX7qDfEmFfOUbpx0Lbk2GpKFTfudPA4jYYP7qBVPtRXAL
hN.eA9XJbrtI.pZL+CsOBIH2jKyE4twJcRb59T2ZGVlq6FGJxVkOcMZWGOPv
1CPvbSYbZiOK+3ZEYvrbv+nWyBCcLfGvX.cgFChdNFX6cLT2XUKCLHBiGnOr
rXN3yZSzjt7LbeQdtGGAcyj5HvBIWNVa9wDyERn0LAnkGslM9HwU3vKU2Qk5
L5BwIE10r7rklvNNXH48MtJEEbrmSDDMB..MDjI6Vqu+T6OZjliwPvV2FvPi
EU+vERJgy0mRz4tiLm3wvOgHO.WzmEjGENyOPDXymokHIUoTRoRnjq4P7V.3
BOG+8hrmhiBd0SYIfH6qfwvxUkqC1ETl8pY4gepL6.wmOTPRikUfDztDqkDC
KCQpuvwmqNEbqLa97gK93hpEZ4XrVgkHEFooRkBCjakGArgyZDmV5jJ09NyU
Gfy4SY4Iypcm9UQ4gEQSllkjkC7Ql+C6CXSBgcn0bEQXRmij.+BT5nOFyB8P
LKP7n4wetIAf6iggQOENlvxx7MFnL+0ZWlZIWMTjgANPpobSDYRMVH4lXNbp
rAy8r1F4o.BKlNNI9onwyCiS+00D6SgMgD69KXRsvzGCSSiRllsJ0R6L7PYh
.TgioXNWnATQPXPTqRLzHQPw.dQUTHzV.p.KVVuxE621MaeX0hUIkw0i0h34
oFmIprZ8PRVXo4CIwEtEQoGRDcMkYxvxrIF1m36WUFs4SE6lVwmBSVEk8P2b
Gt4wmjkN+fP+Ve6EvToMA8MIWY6qFa8z6qjHmWu3wr7x92YqwTWOpUowkM4k
g8h1tRdR9OtGUYvXbV3zecnxjZPljv.uAkToDbGRXS8L3NDiAsnvTEntWST8
HC7ntLImjaNSyVrHBja1kBwiC996d8GtK3a+l69v6CBAWdd6q+ZmNQKNrSzt
rwKUUdE2BGHPLdc7TlPUGwZmCbfMxiq1xKKJyVN30X.Tv1k.43tDX2XYOKUu
Ljmo8v7xgS7f8W1yAwye1m3wXAoiDNgqGCQ9fHqocEy6zNU87OwiAajikOCD
Ol6SMefVu28M2E7C+k29MuOvkBOBZvJ7vBMtqleqyX6pwiTgM5gnwCbxymHv
ad8W+We2Gd+e+u8V+hBNM+oodCD7Javq+6u8urgx68DMDtVWM+.V0gF4DKIh
FFIx9BfDMKbQGcbTuQhTeRhay+1aRTh6Jt5wYQxyOIpgIrK3jH1mlilEWrLq
HZv1ijLG1iXUtdBj3F6QdN.XPa+W.juyYXAxZMFaodIXVlbAndkGS2Xdv6WF
kN4GLYL5OEmDM4t24NIQnCvHnP5NJkgX865QJlRNervlPqpVQi0BlDwUJBkS
zL7s15On5W3VcxV3Gw+32qWMK9jwOBly6HHw4pqB9wM0XfjwQZEhp2.cG.+n
dD+J5K+2A89SQ6pHRwbge8Mt+sAICz.yR0PUeXx3J+CRGmI6vdNQ554jOAog
yIw8pd7WZKtqgZDyIr3LZZsmUi60bIXVurVEsZeVEVhqUg8qyVkGGkOZnRaf
Qttl80Gw0lghX3Kohm27tAu3HBYWO.zrtwii6VfI8QhpRGyV12HGShRbIsss
GH5.l1zNbzmi4NT5PDd1xFZKc1VOCTMZmLWPeLrjeIUgOb1MrIqmc42H9keS
sKrcXLxqICFF2CV+shf5n+VgbXVqawRddZiXZuJpY4LlTj.5ixGtC3N.AoD4
nzXv90HFS8kCHXp6kNdQqDsBHC.Aj5B.BBupn3nfvAzQPociDixXcCJErgO3
E.CSQO2R6HFcL2rQRPJvMDoDTSAb5npbbt14CoYo38t3tWYzuOIb5GGt4Tlt
a1O4LGk.m.6aqoVeQLEfdi+HziXafw8Ojc2vSNgvgUSAU1Us.W4ULC2Di3AC
Tjw7pxiCwXc.8FRLqqdCpijYInmjmE0tuVw9H5SrgT+5YwP0znPfqPlZnDIA
ZmXJmm8jdOeqng50kYEMTBWZx6DVHDLDigDDrIeGWGB2qKx12+oXf1BxReY1
CODTLMOJJMnLqR.oo7lhSiZpqIx4VAy2ECyBAeHaQX5AheFLZ0wMEmhZfCM1
5XdXKM.U5ypwM7iAU0OHdby+OjbLvOkJ81kxH2wUyjW9fFQsypE4XAfSU9Wc
9cCOYebJqKSFwQNZDngqOumV1n5mUM3XAPub0tpvcknJu6qH5YUENF7PtpZ+
2RG9Ugxw++Bc3XtlzO4qSSINi7adk3NUA4d8F7rV7953sWKDhGB9uMG1CnW9
yKhJyyB9C+4+4ezKAuKHN7LndIkGnmATuFOV4p7z2+2lzppU6cRggnXYHHJd
pfKMmXFDfKPKTiwJFgv3JNVoUZPwjR52L0PId057QffCUgtZvKalD..NgPUL
AyT1dJQ8t4BtjP.wLAMpzmRhZvOqFfHfWncJQUNAewSJKG62zQNYYVAYxObR
EEgP20OelqbxhQ3Kg9ORGOYEXs.D6XTX5gv46U0nWyr85Mzm4zlXxhhgu19N
fQNR9a.Xj6Uu.s3WQzzhfgKNBVO5HNB9E1ECkT+Vw39MXniJNdHEtBlpCFX1
.fs8XytasPzSNhnSmWg50Zq86RB+k6AWjC9trhXy9.OHNMXir2VdLSG7dKA2
cuWHX3N.YUJnM910eCWmVJffG68Q4+5YW6CGykV27PBQGRmK54FmsWa.8Mal
spMxVyyzbbQ09o0FHkqoC3tqOVd7nnH.N7wEOrD3y90fWsdS2EvFtXoSOEbs
ncriloBF83H3u+ouT+j2pOQZyl4rZCaVc7ecfcq4d2V1dcKMlGX1YyDvz0je
X39RxXcKiFAqa0jq7cBQGdoORIWFevG9pkxc41CwQl107KwpkNnpMhbgbU5D
JyHrtaY8K.nrKrIOipLpwoHlFQt0UkbuOjR7aEWkvDvsHYeDaOEekH7ubhRy
ojFW5HyUH7yunF6KYmLwXkqxXxQtSd98xjP8hWlWfPdk5tE8AUd4i3kfu.Bk
urHJ5iCt3jkcKNYlxwdJQ44MKr9YM0iJpirtgIW9xAU4SqRevTM++or7n44f
wmYA1pTO3SwkOFDlr7wvf5PtN3Zf0+cFh.0MFLti8Mb85NPoCJYzXuiLuALX
2FYFr4Jc2stG7mNxo3IsPK99339m28DAKJcNLgaNlmlkj.weGN8QPQIXMBFv
AuJbUYl8.Fv8gGVRV1RyUVkF+yqh1bjPsEOD6bWG0drORbUHLbsySoD6tdyX
uymGDYCeIC8+wrt8j5ZYRHLaZ1Fpfynli2snkkOVGLpYlBZMJ2L0A+c5r0W.
96hogluwXAG9kzzxx0dw9UnwTR8uQeQMEaRXVmr6Pp1cUm5T7liOryX5U97u
YzbtCObdBe32Mil1qUv9dyNXvqVL8DNC5T5wLAxbBNy.e33.frOEEGM5B5vN
8vNZ5oN2MYjh77lrpm+sSnyMQGDObGVdOuaBOsJU5zN7IoHRGGr.ByQI058C
eRFoum9j16d6Sn4JDY2Sj3Zbo6IQrr8QQ7dNFhUaNFh6.88titR8ipG8i4UL
xY2QLZ0gT8A6IgO5IcOHI6n4b6n03xkum3jdzSBevNftV7C79vf6iNh1iNh3
CnqOyQ9fYvTgWWGRxr8f6wrjW5J80pmz8Q0pYS+5AdbbeXI7hBh9H2h8gZ79
LOw8hAi9XZZGDNKeV0K.Kzkuq2SOiOudl2KlSsOv29nEwTkkWBxre8M5hL4h
6ED6CqoLzUa1jptZZ0L0cyQ8jix8QOYTMxNZWI8QW0GNR4kferW87EQJjzGd
FhOXYH8w4ehOD4H8R51GQNg6kDmOjsMmEf8Q31KTkPc0TOZNk0tR9GYNP7tN
NHYNAmtdSVWOCLX9Uy.Cte9ew7BYw5U3.9grX82FyY1Szd4e01pUVDOaYVbZ
Yw5M2liBxihjN1+e87ad1TUujrrzNa+TkRMF2YeiJcbhthERGGD6c+lmMYg7
.YIL69xNaDXon2jkv2jU+7HkxNDOnjvbjTYG0bUO+hmMM0KkEVJeuSURgZLq
ygOAvU187Kh2Mk5N9hmMQwOehRs98w1VreaUi6GjnD9lnNIMsUKOvNuXrMcw
NuPr24kgsiWD16+kf8tu.rsu7qqdO.uyZRrob5LGVVeu8k37juMJcU0pcr9U
M81Hw8yeHNIoY86ZurZqWIk5U2q4ctc605yd.1nLKwmsvUsmhMz0aqmMuiKs
2Cd8MwZpuUVSATypO8jPseAVauMxl9Bg0U8fYsDYUeBZZ65hEvxz40uzpa8x
+Yzx7rkY4Mu8tGS0Me+UkYyyCmEGUUE.au1g2VyHkCWc2W+xUKQ556oMi21S
FKLKD+4MIzL.cioTyoHjcsVMqHL1BNrcvkSDN2YB2ySB9DqeaE3BxTuIKYVv
2kYVOv3Cg7sWDWSclNtQ.2SCkuk8+Lz92wZF+i00Hafgt1rllMZ258TDZrh6
eJ7DPZzuMozuB6wYyMj2uq.3ENpK6eb8KoX+MCF9YYeXV+84C+nPFv6.LI3C
QyWkDV6Bwo+zpmplTcqU2Ucwmr0rWTRjoNY2BljJgPxq1lMXkp5SLNp9vqpk
j3V2lfJXFWSgOwTBt0oEgx7BZb6ayw4BY6BqqMXVWgXaMhM0Xj8QLIN03KYz
9eTtFm316+us+dgSmtKTHLkhkkPL+TQb09g0Yvd+pxxZ+ZaIbr8iSQPzJfA3
+pfHI.ZXwAfnpc.tcRnoHwrkLF0wf3XLL38wv7S2ee+UWq+cs0WYsCer5778
+0D3yglmNBmtWFKURkuIYUzgMw6PdhZOAOsB.a13d0yu9d.9Nyw80fGgVEHb
a71VAMaoY50Amcif8NvksAO11b5mxMEGpEGws4PG9.rJrbaAJZFgEKq8mvVG
iu3e+h+O.nFoHHB
-----------end_max5_patcher-----------
#8

Hi,
I confirm that the cpu load doesnt come from audio (just disconnect the “playback position” patch cord, and you’ll see that audio plays fine).
I cannot help for the jitter part. But @fraction must have very good advice.
Good luck,
T.

#9

Thank you @fraction and @tcarpent for your help in looking into my patch and have provided me with examples. Much appreciate this. @fraction, I was wondering on whether you would be able to direct me to some reading materials or tutorial videos on how to implement video alpha blending layering using jit.gl.pix object. @tcarpent , I was wondering on what would be an appropriate signal and I/O vector size and and whether I should turn on overdrive and audio interrupt? Thank’s again guys.

Ronny

#10

if you don’t want to spend too much time, and if mainly your needs for visual are generic needs, install the Vizzie package from the package manager. You will get several useful tool for the day to day use. It’s a library of ready patcher for video processing. Highly recommend.

Inside, you will find an alphablender patcher. Patch bellow as example. Within that alphablender, a jit.gl.pix is used. You can learn from it for further projects,


----------begin_max5_patcher----------
770.3oc0VssiaCBD8YmuBD40zHShiSZepR8yX0pHrMaBIXvBiyktZ+2K2rC4
5l8VU6KVlgg4xYNy.O2KBlI1Qpgfe.d.DE8bunHqHifH+5HXIdWNCWaUChUJ
YCENvsUEVkujxWLWRxUNyLNNdX7.v3YlunI1EiFMLF7n+LFKXMknQU0nlqH6
TMRRqI4MkTNinrtC4ERKbmHa02lEnn1BmpoSjZeEwENPH3QyNuzqm4yf6LMW
QUCqX38LZsBdvzgwafSu.LfPiLY9zIAvv33.XfIDUg1nPh2pDVmeuPQ5gxfD
WRTD4bBGmwrwV7EgowWClzI77ErCECST.fEzbEUvwx8vt.u.qvdjyCcQvbFs
piFErgobmUKXMJhFiVZi7rLLiMrTrwG8FkdhxzQdI4lJrlxc4tdKJwHIPAOr
zLAkDihGkLMXuSfZa.K3JBWMuVgUj1rILtCNUqHKEx8iiS4k72lZ46v9RoVI
+eRsX3BMVcSx0UU4tnWeeBRSuPiR9mkdwIa0Ev1xmAz6XcaERVA3mtBI.cig
3nY1w2iGmZ4XHzYSweMBzH3amjThUR5NC4HCyWXIJuqg2Y1LhHaCALmKzEBM
aytc+9feIJynbBXCsfH.M05rGfYUKwf7kZsILPItdsQZ+9sVYCV1whr5pAQd
QmSpjBkvjPcJUQ2w5NLkrcCsllQYT09itv5ompIdX21Jebmb9ZRgt+cQctTv
Xgce5Fi70pkRQyhkgxck2kme.2FaNeirEkhhil5jIjEDYnNWmjfRcriT2pjz
qwRlzJrEf176gA.4PcMT6E3E3Ridi24exDm2EI5Ueni+F9jIeQOzY5G5gN1n
.xzj7SddmMSLxON2qEMx7VS4ebA3PtTPpUTdWOzCGJLFktH7duNp0H21SeBN
J4dyHzGzQytCGk7IjPSuC+jdlebLCbU0Fhr1qr0EZ9+JgkwNafcIk6VZobPI
wL.yoOxIBK0MEJcGg+QKvcotKFglAJRtt+wcCmN65YlXoatLc+0U3b+0w5dv
duz6O7DP1e.
-----------end_max5_patcher-----------
#11

Thank’s again @fraction for the kind help and much appreciated. I will look into it.

#12

Thanks a lot, Thibaut!