Re: Gradual Growth of Memory Use?
"David McClain (as dbm at refined-audiometrics dot com)" <[email protected]>
| Newsgroups | gmane.lisp.lispworks.general |
|---|---|
| Message-ID | <[email protected]> |
This is another case where having modular array indexing would really pay off. (I long for a DSP). > On Mar 15, 2025, at 14:02, David McClain (as dbm at refined-audiometrics dot com) <[email protected]> wrote: > >> Scrolling the WF or deleting old and adding new? >> > > Neither one… We start out with an empty display and add successive scan lines to it with each call to analysis processing. When reach the bottom of the pane, we start overwriting the beginning, but the display shows succeeding lines first, followed by new lines. So it looks like the display is scrolling, but that is just an illusion. > > At any rate, what a scan line contains is floating point values of arbitrary measure. I have to convert this 2D scan array into 0-255 in 4 color channels, turn into a linear vector (by :displaced-to 1D array), and use a high performance vector to GP image using gp:image-access-pixels-from-bgra. (see previous note) > > > > >> On Mar 15, 2025, at 12:59, Paul Werkowski <[email protected]> wrote: >> >> Scrolling the WF or deleting old and adding new? >> >> On 3/15/2025 3:42 PM, David McClain (as dbm at refined-audiometrics dot com) wrote: >>> >>> >>>> On Mar 15, 2025, at 12:33, Paul Werkowski (as pw at snoopy dot qozzy dot com) <[email protected]> <mailto:[email protected]> wrote: >>>> >>>> If I remember correctly, you can convert an image to external image format and then convert that later to an actual image when wanted. >>> >>> Yes, this is what I’m doing for the background logo image. But not useful for the waterfall display, since it is never an external image. >> >