Hi,
I have assumed that the size parameter is the size of the allocated buffer in
bytes just like if you were calling malloc() but you are right that the
documentation is not very clear on that.
I have experienced some undocumented limitations with AERT_ALLOCATE() though
which you might find useful. If I specify size>32767 it will fail and return an
error code. You can call it multiple times with size=32767 to create multiple
buffers successfully however there is no guarantee that they are allocated
back-to-back, so I have yet to find a good solution/workaround for this for
situations where you would prefer a large contiguous buffers, e.g. delay
buffers.
On earlier Windows versions it would eventually fail if you called
AERT_ALLOCATE() many times (way before any system free memory limit), however I
have not been able to reproduce this on newer Windows 10 versions. On Win8 and
early Win10 builds, it would fail after allocating a total of between 5-10M
(which is not unreasonable in case of needing long buffers/delay lines running
at 192kHz sample rate and with 8+ channels) but I recently tried this again and
had no problems allocating a total of 100M (still only 32K buffers) so this
seems to have been improved.
Regarding your question regarding std::vector, I am not sure it is possible
since it uses the new operator to allocate memory and I do not think this can
be overridden. You do not need/want the array size to change during runtime
anyway so I find it just as easy to just use plain C style arrays.
Best regards, Martin
From: wdmaudiodev-bounce@xxxxxxxxxxxxx <wdmaudiodev-bounce@xxxxxxxxxxxxx> On
Behalf Of Culann mac Cabe
Sent: 24. oktober 2018 11:10
To: wdmaudiodev@xxxxxxxxxxxxx
Subject: [wdmaudiodev] AERT_Allocate "size" parameter & how to use AERT_x with
std::vector
Hello wdmaudiodev people.
I am developing an APO and am doing some cleaning up work, including
implementing AERT_Allocate for all my memory allocation. I have a compound
question:
1. The doc
https://docs.microsoft.com/en-us/windows/desktop/api/baseaudioprocessingobject/nf-baseaudioprocessingobject-aert_allocate
defines the "size" parameter as " The number of input connections." Is this
correct? It seems to me to be incomplete, as in without other info (eg audio
sample data format) this can't evaluate to a byte size... so can someone please
clarify what this paremeter should really mean? (I'm guessing Bytes?)
2. I am currently using std:vector and std:queue for convenience to manage
IO buffering. Now that I am converting everything to use AERT_Allocate, can
someone please provide guidance/code snippet as to how I can use AERT_Allocate
in conjunction with eg std::vector - or do I have to revert to pure C arrays in
this case?
Thanks!
Culann
