How to convert PD's variable delay [vd~] into Max/MSP
*EDIT*
Second question:
Another issue, this one pretty glaring…
I am wanting to do real-time granular synthesis.
In the puredata patch, it uses variable delay (in pd: [vd~]). Its quite a simple implementation.
In Max/Msp, there seems to be no straight comparison
I am referencing to this patch:
http://www.pd-tutorial.com/english/images/kap3/3-7k.jpg
I know that there is both Benny’s vdb~, as well ad ej.vdb~, however I am completely lost as to implement these in the patch.
What I have done is replace the vd~ with tapin~ -> ej.vdb~, and then replace the delwrite~ with another tapin~. However, this simply produces an audio feedback loop without any granular synthesis.
In what way am I mis-using the ej.vdb~ external? Is there a better way I could be translating this patch?
puredata's [vd~] allows for you to patch a number of different data lines into it, allowing for flexible feedback control, whereas the [tapin~]/[tapout~] delays as well as [ej.vdb~] seem to require a straight [tapin~] -> [tapout~] or [ej.vdb~] connection. I would really love to start using this patch; sorry for being new to all this, I love the transition just wanting to have my creature comforts come along with me from PD :)
for [wrap~] I have replaced it with [pong~ 1 0 1], and for [samphold~] I have used [sah~]
*first question, already solved*
hello
acquainting myself to msp after awhile in pd
i am trying to convert a simple pd patch that i use into msp; one of the objects contains an expr function given here:
expr (pow(2, ($f2/12)) - 1) * (44100/$f1)
while this object is easily recognized in puredata, it does not work in msp. am i missing a simple "grammar" rule in the code? msp feels so familiar (if not easier to use) but there are little hiccups here and there
Numerical values need to be specified as floats...
expr (pow(2, ($f2/12.)) – 1.) * (44100./$f1)
forward slash with the (pow(2\,) was what caused the problem, just straightened it out :)
Another issue, this one pretty glaring...
I am wanting to do real-time granular synthesis.
In the puredata patch, it uses variable delay (in pd: [vd~]). Its quite a simple implementation.
In Max/Msp, there seems to be no straight comparison
I am referencing to this patch:
http://www.pd-tutorial.com/english/images/kap3/3-7k.jpg
I know that there is both Benny's vdb~, as well ad ej.vdb~, however I am completely lost as to implement these in the patch.
What I have done is replace the vd~ with tapin~ -> ej.vdb~, and then replace the delwrite~ with another tapin~. However, this simply produces an audio feedback loop without any granular synthesis.
In what way am I mis-using the ej.vdb~ external? Is there a better way I could be translating this patch?
for [wrap~] I have replaced it with [pong~ 1 0 1], and for [samphold~] I have used [sah~]
If someone can help me figure out the simple tweak I need to make this PD granular patch functional, I will dedicate the rest of the year to trying to help other people's beginner mistakes on this forum to pay it forward :)
I have every bit of code I need to translate it, I just need to understand the programming grammar & difference in operation between PD's [vd~] and [delwrite~] compared to [ej.vdb~] and [tapin~/tapout~]
You will get more replies if you post examples of your trouble. Create the patch, select all, edit -> copy compressed. Paste into the forum message window. Also, be sure to give the help file and ref docs a look before posting questions as many here will read your post and won't reply assuming that you will figure it out soon enough because it's right there in the documentation.
re: your question, I have never used PD and I suck at MSP... really. I try but I am baffled by audio synthesis. Yet, somehow, in a matter of minutes, I found the following. I have no idea what it means, yet, paradoxically, I suspect it will help you.
[vd~] - The delay in milliseconds is specified by the incoming signal.
[ej.vdb~] - requires number message to change delay time.
[ej.vdb-sig~] same as ej.vdb~ but accepts signal as delay time
[tapin~] - You must connect the outlet of a tapin~ object to the tapout~ objects you want to use with the delay line.
Note that this is not a signal connection, since no signal travels between the objects. It is merely a way to indicate that the objects share the same delay memory.
metamax obviously does not "suck at MSP". Good advice. And he's helped me out in my own experiments in this domain in the past, even if indirectly.
I am currently in the process of converting my own "granary" granulator from Max to PD (the reasons are irrelevant here), and if you are patient I will post the results soon; even though I'm not currently familiar with the vb or ej objects.
Brendan
Whoops about copy compressed - gettin used to all of this still
Oh, I found that info. How it operates on a basic level makes sense, translating the patch is where I am hung up. I posted a visual of the PD patch. As you can see, it creates a feedback loop with the variable b1; no direct signal path connects the [adc~] -> [delwrite~ b1 5000] signal flow to the [vd~ b1] object above, which is where there is a break in the MSP signal flow.
Here is the patch copy compressed as "close" to the PD patch as possible
1688.3oc6ctsaaiCDF95Tf9NHHzKR553JdPmVrXA1mi1hExxxIpUVxPRtM6V
r4YeoH8AIKKmII1oialarCjbH4P98yY3HZ5e712bg8jh6Rprs9cqOZcwE+Pc
kKzWq4JWr9BWXOO5t3rnJ8GztJKcZRo8nU2qXYcVRc8+rHwTL11Ved88lDE+
0aJKVlOs4dNatbQopDhKxJJM+ONicG05E1XmskQU5+pKYFOX7lRHUWf1ES9x
0R4llR5M4EkIwYowesS8snLoJIuNpNsH+uUef500pyntursVmuLS+oXaqy4o
45Rc6UlbSKan4S18ksk1soSmlj2oM807hIc5A3iZ8Rm+67kppV0Gq69YaLpn
53aSyuokAwXBdSceMyq4Mt1rXxNF1rrhnZ0X1hk0C1G04FpZ2LD2s5KJS26G
uLIScwuoGxbzW6+d6aZdW81nyVFSPL1ZFywMPCWm4DlpRmzhvtsN4tZf8zyJ
xqyilqa.1+UYZT1vfZZds8HKEjleyCyr0koUwQYI6.EOJTV0UVkrbQmxsMIy
a2VUCQE48Y4c5K1gg6TcGQ4PS25rn3jtxXn3upuS0GupuqcQtQa2RTEGkq.a
0muaksGIziPW3okDtF4gqoqJniBU0F6zsF32zEDXDjxdTV47HfJHUaLc9x45
Q4+HuHO4Os2u3huoszoi8Piv61n80X21W6ReQ2saCgTnOREJiTnmFEJ2w2nP
4jB83nPS9thXsaCk5qGMM9d6mhVrRAqM2eye8fBxtiV5AKo1tY6JCZovbrOm
0QCIYddZCgTKJ70RD0xHb2Ua.L7M9QmllFAilNhDhH7UIgvOX7MFBwy71yfP
bN5Dx6eVy1bRliQ3SDTurS3KzywDD7LIH1qBBJfHndDjAcb4dDAAff7HBpuW
LM53F5RDD.BxkHn9DzpkOFhNBJ4Ki+1zI2aMggORR9pjjDGbMWbMBICMokvu
e9HHRZOjjfHodyIYdfeROWzQR0QKTkeCIolxzwAwoERvImc8C31wrneMeEHd
xf0wOsPKJxu4dKlkiEBmihQyQ0yamb0h1zN87jHZNpe69lGSC9nHGZBo94o1
rwCLwLIBQDEEWTguUvwYzy5XWFRXBSR5rNOjNHJG.nblHNEZTeJxr4mjlfhD
XJl6qQ37PBhfFhfbEnKSRJT3dDFWMWRdyFhhV4MyGUdyvGAQYzdOwTaHHO78
bYqhtEgLD8b05yP9lXpEAnzWF9XHexS1PLjwSFODWYHBgQCQ6PjA8kIv3dLx
5ZDBQzVcbPHxAePDNSxHsgp6+jyb0UjX0aTRFePJhQyD0mhbLOX+Ua3QJIiG
lfn.hFhfDgdTRFARQzRyFjhV4MiRx3gIHJAQCtE0XLJIifXHIwP8+J1afm.F
kjQPLjK4IaHFx3IiRx3CxPzCteHeYXbdHbljQFs+gFDhbwGDs31nphRD5OiQ
XTOLx2bVuEpqO4OiUlMKqPUTm7CpG84JFpNpdXNzQ0yo4n5g4ENVJ4NLeEZ6
IGykRGY3w3P64mxIzip.llDmNOJaQlZPoZWP8myw2yqVcaHIaOQmAdFuPbo6
w4D1hDqOXrZevRJG+zh32HLOIQpQOHl83RKXqKMg2FWZluZkTHauLS8Kno9O
QQrw72f2LeeJhMJhsinrkN0hOUxV+VxVFIaeQicq1Zh0rmTnaZg4nSZHbztf
3vhE4VwBpNtTQMT4QP0ACbBqPUxcKJstbQw2uj+oOMx5x2Mi+AFe7UWcMa7U
u+Roj43L9CuaF6pmD4cZfM5q52d1GWgle2NV4d28ns1SSAXmkl262SHcCo4F
6PdUEKKiWi.q+o7wpUSYZRUcZ9lVxG2DKn0g+cOA3P87zoKJRyqW0ds97vxF
3lf.hI3gZSfCwDZNbQsXn0FXfrAeTaCNfrATKGz8vm61fGHaHDyrzivFv63f
KXMMdsgfWM1f2u.rDtsAPQKs6nEtrANGrMf14VYgfrAWT6eP.1FPKKwf4iSf
Z8.r4kjX1F7N6WDGnIkPsOZXgcicJhcVaA9Pr.b6U.BEgZeBfr.TGgDHGBnd
cafVt.yAyl.nH7XLT6SCT1jX3NRaev1.diRElb.0JZFnIkXnNJUFnUOyPcVt
YfBTkg5njXBviC30F.8zFX7eArAIp0CfrANpC0fAyOMpi5lCxGGG21.nX933
N6pdfGGPqllCJlONp8OvA4iii6bhAauYf6rqx.aCrya+Cb+y+XMDn1Os.1ty
.04DP.SOD9KfMf50wAzFdwXolKnd6+ExtINN
Here is the patch where I try to do a "simple" signal flow for the tapin~/tapout~ objects
1766.3oc6ct0aahDE.94To9e.Y0GR553N2.LqVsR6ui1pUXLIgVLXA31ra0l
e66vL9BXLNmjXmbbx4ErEWFlyLemKyYFO9Wu+cmMXR9swkCb9cmO6b1Y+Rel
yLmq9Lms5DmMXV3sQogklabPV7Oym7sACWdsp3aqLmuJbdR1cqO+U4YUYgyh
MW6uJRBSWeo7EUowUU+y7X6at9QixyxhizkzWWcWSBi990E4KxlVeWrUmNJO
Muv9brQiGVePUezmOzgOhs44SLO2.cc8R03VU2VEAaXiCsKfqyxKhiRSh9dq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.IDY
vaRBQr23aVRHAiehDB6fSHe7IYs4nXiQ5SDTmrSLVYHHWwSjf3uIHnwDA00K
koZ5FvIBB.A4QDTWuX9VaPADAAffbIBpKAYxyiqMcOnhfh+1neLcxcNS33ij
TuIIIIDuYd1IG1GQyJLlIIIQRcmUXqMI67AiJRpJbtt7qIIGWFig3zBIEjyt
NCYyFosOyTaGKezf0gOsPyyyt9NGtCyAg1n3jMptd67rALEXSzHhrQ8a2UOM
M3ihXjAoNTjcEtHBL93jAHhhhxKw2H3DbZtN1lgjdlLYqXqxjMCQ4..kVhDT
nQcoH6heRYCJRhoXtuDg1gjDA0GA4huLIoQg6PXb0BE4MqOJZo2LeT4MCeDD
kQ69xCox2Ge1gBuAgLDMuZcYHeaL0xwnzWF9XH521TuLj0SlH.WYHBgQCQqP
jcjeHFVsCoGa+kHDhnk5XuPDCePDNSxHsfp6NyYt10nlKmRxHLJhSVh5RQLy
jloDdTRFAPPT.Q8QPx.OJIi.oHZnY8RQK8lQIYb+DDkfntKQM6u5REmSIYDD
CoHFpCCYWorxwbJIifXHWxSVeLj0SFkjw6kgnItuOeYXzNDNSxHmV+P8BQt3
ChleSXYdAB8mwILp61NjcudytP8UuDiL6pzbcQcz2ndL6qXnZq5gynspmiyV
0C2KXjRIXbeMZ6oFITJlJ3Pro87hrC8nKfowQIyBSmmp6TJ2FTeY19ddyp2F
PpsGo8.OqWnkalPjx5yvds0mbTpQOtH9sJlGkH0nIhYGtzFuwklzasKMoGEx
1ymoeIY5+HEwF2eMdy88oH1nH1NfpsztV7wRs0ugZKmTaeVicqxYhyUOpP2L
JlCOpgvQqBh8qrn1nrfpsKUTCUdDTs2.mvJTEe67Bmymm+yyEe4KCcN+CWI9
DWL5hKtjO5hOdtRwYrQe5CWwu3QQdGGXi9o9si0wUf8+sikt2cOXi8zV.CRS
x57eblohTegsHux7EEQqPfU6.SNMpJSiKqRxVWS97lsRNm8+GeBv95YISmmm
jUsrB67090aNzxfDyxP8egam58C0+PeueYXaIEY8C9m9rT8ZgDP+fOp6GTPj
AApEAIDQvC0h.LKq0x.Gsx.GjL3iZYfARFPs5fz+UfL.xxZ8ekE3kkd.x.d6
GbAqSiWYX7aFYv6U.KgaY.TzRRTG4sP.VFPqsUd.HYvE09GjfkAzxRdm7Cd.
jx.p8M.KbOrSQ7SZIvGhDfaqQPnHTaKBjDfZOyfBPB0iW.TXpbFlEAPQVv4n
1mFnrXvwcDd9fkA7leXXpCnVilCxnDG0yWBGzn13nN6pbPApxw87kHA2OfWY
.TVt4hWAxfB05CfjAApC0fCyOMpi5V.xGm.2x.nX9D3Nqddf6GPqNs.TLeBT
6eP.xGm.24DC1ZB.0wZH3fkA9os+Ag+oerFRT6mVBaUAf5bBHgoOD7JPFP83
3.JCOarT8Ize7+znj0.D
You need to turn the bottom right box to 1 to have the feedback turn on (any more than 1 will distort/glitch out), you turn the bottom left box to one to have audio run to the DAC.
Brendan, that's funny you commented. I was just looking at a patch of yours the other day and realizing how little I know about MSP.
Cameron, you need to include the beginning and end of the compressed code including "--------begin max 5 patcher--------" and "-------end max5 patcher----------".
Here you go (patches you posted)
#1
#2
thanks!
I am seeing that in the PD [vd~] object, the b1 is acting as a remote call. For MSP there is the argument "tapconnect" to do remote connection. I will explore with that and see what I can come up with
vd is maybe one of the pd vanilla objects where one might be better off building things from scratch - and eventually different form the original - instead of recreating it 100% in maxmsp. but thats just an opinion^^.
as for a perfect recreation of the samphold object try searching this forum; not long ago somebody posted a max patch for it. (sry i dont have mine here now but i will ul on request.)
-110
max is like pd in that you work out signal flow, look at helpfiles and know what inlets/outlets are what and what goes where. your use of tapin/out is all messed up.
roman is right, vd~ is not the same as the vdb~/vdelay~ abstractions, but for what it's worth, here is 20 mins completely untested patch:
Here's my try to do 1:1 translation from Pd to Max. Pd's [vd~] is just [tapout~] in Max. Nothing fancy. I remember when I was moving from Pd to Max I had a feeling there was a lot of redundancy in Max, but otoh it enabled a much clearer patching. Forget [expr] ;-)
hi AK, this is great! much clearer than mine :-)
the only thing for the OP is interpolation - tapout~ is linear, vd~ is cubic. but for this algorithm i do not think it matters. for cheaper (cpu) use and interpolation options it is easy enough to redo this in gen~ anyway.
also, vd~ guards against lower than 1-signal-vector-size delays, tapout~ does not (i think) which is why i had the [maximum~] in there. of course in gen~ you could use lower delay times down to 1-sample.
the only thing for the OP is interpolation – tapout~ is linear, vd~ is cubic. but for this algorithm i do not think it matters. for cheaper (cpu) use and interpolation options it is easy enough to redo this in gen~ anyway.
Also tapout~ seem to use 4p cubic interp (see my example and zoom in buffer waveform).
also, vd~ guards against lower than 1-signal-vector-size delays, tapout~ does not (i think) which is why i had the [maximum~] in there. of course in gen~ you could use lower delay times down to 1-sample.
[tapout~] doesn't allow delay lesser than VS size, so it's guarded in the similar way.
Hope my methodology is right :)
sorry for delay, i got busy.
yes, it seems it is some sort of cubic. i am pretty sure when i first learnt the tapout~ object it was linear, many many years ago, and i've not used it much in the last years and probably never thought to look into it. same story for clipping vector size as well. i should really try to keep up to date with these things ;-)