okay, i admit, this title might make me look a bit strange, but i assure you my problems are purely max-related. *ahem*.

my [anal] seems to be buggy. I was working with max 4. when i put anal's table size on 16384 (the maximum, according to the anal help file) it tells me "not enough memory for table".

strange, because the [anal] in the anal help subpatch called markov_chain_example does work, using this table size.

then, since i want to get max 5 anyway, i installed the max 5 demo (hoping it was a max 4 bug) but i get the same error message. It can not be that i haven't got enough memory to work with (i have 3gb), so what is the problem? and, why does it work in the markov_chain_example, and not when i load an [anal 16384] in a new patcher?

my [anal] seems to be buggy. I was working with max 4. when i put anal's table size on 16384 (the maximum, according to the anal help file) it tells me "not enough memory for table". 

strange, because the [anal] in the anal help subpatch called markov_chain_example does work, using this table size. 

anal-problems

Works for me! I am having no anal problems whatsoever. It froze my comp for a second (2GB) which is to be expected, but there was no "not enough memory for table" error message.

Quote: bil3y_da_k1b wrote on Thu, 11 September 2008 12:19

----------------------------------------------------

> Works for me! I am having no anal problems whatsoever.

Sorry I am totaly immature today !!! , but yes funny topic !!

Quote: bil3y_da_k1b wrote on Thu, 11 September 2008 12:19
----------------------------------------------------
> Works for me! I am having no anal problems whatsoever. 

Sorry I am totaly immature today !!! , but yes funny topic !!


On 11 sept. 08, at 19:30, bas gebbing wrote:

> okay, i admit, this title might make me look a bit strange, but i 

> assure you my problems are purely max-related. *ahem*.

> my [anal] seems to be buggy. I was working with max 4. when i put 

> anal's table size on 16384 (the maximum, according to the anal help 

> file) it tells me "not enough memory for table".

> strange, because the [anal] in the anal help subpatch called 

> markov_chain_example does work, using this table size.

> then, since i want to get max 5 anyway, i installed the max 5 demo 

> (hoping it was a max 4 bug) but i get the same error message. It can 

> not be that i haven't got enough memory to work with (i have 3gb), 

> so what is the problem? and, why does it work in the 

> markov_chain_example, and not when i load an [anal 16384] in a new 

Believe it or not, my anal doesn't exhibit this behavior. Note that 

with 16384 as argument, the object allocate a matrix of 16384*16384 

stores internally as a short which makes 536 870 912 bytes total. My 

guess is that your anal works fine in the help file, because it's the 

first one that you open, and then you try to instantiate another one, 

which asks again for another 500 MB of contigus RAM space.

> okay, i admit, this title might make me look a bit strange, but i  
> assure you my problems are purely max-related. *ahem*.
>
> my [anal] seems to be buggy. I was working with max 4. when i put  
> anal's table size on 16384 (the maximum, according to the anal help  
> file) it tells me "not enough memory for table".
>
> strange, because the [anal] in the anal help subpatch called  
> markov_chain_example does work, using this table size.
>
> then, since i want to get max 5 anyway, i installed the max 5 demo  
> (hoping it was a max 4 bug) but i get the same error message. It can  
> not be that i haven't got enough memory to work with (i have 3gb),  
> so what is the problem? and, why does it work in the  
> markov_chain_example, and not when i load an [anal 16384] in a new  
> patcher?

Believe it or not, my anal doesn't exhibit this behavior. Note that  
with 16384 as argument, the object allocate a matrix of 16384*16384  
stores internally as a short which makes 536 870 912 bytes total. My  
guess is that your anal works fine in the help file, because it's the  
first one that you open, and then you try to instantiate another one,  
which asks again for another 500 MB of contigus RAM space.

While we are on the subject of anal problems, does anyone

have any good examples that show how to use anal and prob 

to create multi order markov chains? I always seem to get 

While we are on the subject of anal problems, does anyone
have any good examples that show how to use anal and prob 
to create multi order markov chains? I always seem to get 
lost trying to figure it out.


indeed you're right, once i run load a single [anal 16384] there is no problem :)

some good info about probability tables and using them:

the markov_chain_example patch and also mchain by richard dudas (available on maxobjects.com) both use bitshifting to fold two or more values into one, wich you then save in to the probablity table and convert back when reading. mchain can handle 3rd, 4th and i believe higher order markov chains.. and i believe it is (simply) a matter of shifting 2, 3 (or whatever order you would like) values into one.

by the way: it would be great if someone could shed some light on how this bit shifting works exactly, because i haven't been able to find out by myself and it seems like useful stuff.

some good info about probability tables and using them: 

the markov_chain_example patch and also mchain by richard dudas (available on maxobjects.com) both use bitshifting to fold two or more values into one, wich you then save in to the probablity table and convert back when reading. mchain can handle 3rd, 4th and i believe higher order markov chains.. and i believe it is (simply) a matter of shifting 2, 3 (or whatever order you would like) values into one. 

by the way: it would be great if someone could shed some light on how this bit shifting works exactly, because i haven't been able to find out by myself and it seems like useful stuff.
