Quote from: ilovefb2k on 2024-11-15 19:56:13

AIMP Analog VU Meter deployed, version named 1.0.16 DT-M 16-Nov-24

@ilovefb2k

Thanks for actively developing this script.

I like transparency and since the original foo_vis_vumeter does not support it and the new foo_vis_vumeter is DUI only/not transparent, your script is the only way to  use this nice feature properly.

Nice you mention features, how to and limitations with new versions in your post, but could you also include a todo and version changes in the post. Especially the last one would be an improvement for me, since I then know what to test. Maybe also an idea to include the version name in the attached source rar itself?

I currently use:
fooBar 2.24 preview 2024-11-14 x86 - CUI/PSS 790MB typical memory usage
AIMP Analog VU Meter 1.0.16 DT-M 16-Nov-24

I did some testing on your current version:


ALIGNMENT:

I noticed in this version that the quality of the upscaled displayed image in all modes is now the same as foo_vis_vumeter.
All good except for the needle stuff mentioned below.

BTW. The original foo_vis_vumeter also has a mono choice besides only left and only right. It combines/adds L+R signals before using the left vu meter to display the results.


MEMORY USAGE:

When using any of the H and C scalings except H:Maximize and switching between them your script hardly uses any extra memory. All of these scalings however display part of the needle below L_0 & R_0 (and L_2 & R_2). Only H:Bottom looks fine because it probably paints the lower part of the needle outside of the panel.

The moment you switch to H:Maximize or any of the V and R scalings the needle is displayed correctly within image boundaries.
However memory uses goes up by a lot. Lowest value I have seen is +400MB, but sometimes it goes up by +1000MB. CPU usage stays low.
Any switch in scaling you do after you selected a high memory scaling, keeps the memory at those elevated levels even when playing stops or plugin is not displayed.
I also encountered some very infrequent freak spikes up to +2000MB while testing before fooBar became unstable and had to be killed.


NEEDLE CALIBRATION/MINIMAL ANGLE:

You mention easy tuning in your post. I still struggle a whole lot by trial and error to calibrate the position of the needle and the length of the needle. In the end I find the minimal angle (and maximum angle=-minimum angle). But the needle is a whole lot shorter when reaching the maximum value. Looks ok'ish at best. I hoped that the VU editor I use for making BIN's could provide me with values to use within the skin.ini file, but those values don't work.
How do you easy tune the needle? What tools do you have available?

While not playing I calibrate (by a lot of trial and error) the needle angle and position with one of the H:scalings (not H:Maximize), so the aspect ratio is the same of te L_0 & R_0 dimensions. Works fine when not playing and switching to any of the other H/C scalings except H:Maximize.
The moment I switch to H:Maximize or V/R scalings when not playing the needle moves from the position I calibrated to a slightly different position. Apparently in these modes the value is corrected by the value of PROPERTY HA.RMS.gain (dB). Can you disable this behavior so that the needle stays at 0 also in these scaling modes when not playing?


NEEDLE MOVEMENT:

In PROPERTIES you can set HA.RMS.gain (dB) which alters the movement of the needle. In foo_vis_vumeter this value is dynamic and can be changed by using the mouse scrollwheel.
I noticed that your script does not have a on_mouse_wheel(s) callback.
Can this be added so that you can also dynamically adjust the value of HA.RMS.gain (dB) PROPERTY in your script?


LAMP/LIGHT:

With VU Editor you can add a light/lamp overlay (with has the same dimensions as L_0 & R_0), that starts to be displayed as an overlay from a certain peak level to a maximum with increasing opaqueness.  I use two knots in VU Editor values +0,0% and +4,100% with linear scale in between.
Can this be implemented in your script by adding the lights as L_3 & R_3 with PROPERTIES such as lowlight_dB, lowlight_perc and highlight_dB, highlight_perc?

I attached current version of my transparent AIMP style skin (red needles) with proposed lamp/light overlays included. Also a gradient BIN style (black needles) with red light/lamp. Also a small video with both VU meters running and show the light/lamp on the BIN.

@ilovefb2k : Some things noticed first try with the "1.0.16 DT-M 16-Nov-24" script", DUI, Foobar 2.24 x64 :

1.  If an LVU meter is loaded first, or is switched to from an AIMP meter, it works fine.  If you then try to switch back to LVU from the AIMP meter, you end up with a white empty space, and cannot then change to any meter.  Only a restart of Foobar temporarily fixes it, but behavior immediately reverts.

2.  Whenever playback begins with a meter active, a file "~JSP_AIMP_VU_Meter_temp.ini" file is created in the beat_audio_vu_meter skins folder with the contents of:
[error]
number=0

3.  Is there a setting that functions the same way as "var peak_hold = 10; // in frames" does in the JSP3 Sample VU Meter?

Quote from: Defender on 2024-11-16 10:37:30

3. NEEDLE CALIBRATION/MINIMAL ANGLE:
How do you easy tune the needle? What tools do you have available?

I completely have no idea about this. 'easy tuning' means playing around with aligning only.
as such, please accept my apology for giving a text that is not clear meaning. i will remove that text and consider have it back when we can enable such truly useful tuning as this vague feature:  there may be a VU Meter, where the wallpaper is none, the background is transparent but meter range. then, the script can change the background/needle color, draw the meter range .

Quote from: sveakul on 2024-11-16 21:07:51

3.  Is there a setting that functions the same way as "var peak_hold = 10; // in frames" does in the JSP3 Sample VU Meter?

hi @sveakul ,
thank you for your feedback,
3- var peak_hold = 10 : can you shed me a light on what you are after  ? we can have this back or make it as an option.
@ilovefb2k

@Jacob91

As requested a working "Text Display + Album Art + Seekbar + Buttons, added Art mousewheel, ReplayGain button and rightclicks" script for JSP3 3.8.5.

EDIT: Needs the cui-configs.rar with stub images to be extracted in the foobar roaming folder, if not already present.

Quote from: ilovefb2k on 2024-11-17 10:11:37

Quote from: Defender on 2024-11-16 10:37:30

3. NEEDLE CALIBRATION/MINIMAL ANGLE:
How do you easy tune the needle? What tools do you have available?

I completely have no idea about this. 'easy tuning' means playing around with aligning only.
as such, please accept my apology for giving a text that is not clear meaning. i will remove that text and consider have it back when we can enable such truly useful tuning as this vague feature:  there may be a VU Meter, where the wallpaper is none, the background is transparent but meter range. then, the script can change the background/needle color, draw the meter range .

@ilovefb2k

Such a tool already exists in the existent VU Editor.

Because of current incompatibility between skin.ini and BIN project files you have to do a lot of tedious transformations by hand/eye in the needle png, adjusting length and finding of PivotPointY. See my previous post about this process and particularly the attached skin.ini.
https://hydrogenaud.io/index.php/topic,110516.msg1054186.html#msg1054186

Instead of putting effort in writing your own VU Editor for this JSP3 script, it might be a better approach to make this script compatible with and accept the values from the existing VU Editor project files as is.

EDIT: I'm pretty sure the compiled BIN files hold no magic. It just contains the L_0, L_2 and L_3 files as is and a number of prerendered (anglecount) needle png's rendered to the same dimensions as L_0. Therefore there's hardly a performance impact on CPU and memory displaying/scaling a running BIN VUmeter. The only thing needed is a decent scaling/rendering engine supporting alpha and brightness, such as JSP3 already has.

EDIT: I'm pretty sure the compiled BIN files hold no magic. It just contains the L_0, L_2 and L_3 files as is and a number of prerendered (anglecount) needle png's rendered to the same dimensions as L_0. Therefore there's hardly a performance impact on CPU and memory displaying/scaling a running BIN VUmeter. The only thing needed is a decent scaling/rendering engine supporting alpha and brightness, such as JSP3 already has.

we have just bothered @BoringName at MB 4rum to compile a command-line tool to extract fb2k Vu meter's BIN. https://getmusicbee.com/forum/index.php?topic=41692.msg229960#msg229960

Am I blind or is there no way to edit posts?

Quote from: Defender on 2024-11-17 10:57:04

EDIT: I'm pretty sure the compiled BIN files hold no magic. It just contains the L_0, L_2 and L_3 files as is and a number of prerendered (anglecount) needle png's rendered to the same dimensions as L_0. Therefore there's hardly a performance impact on CPU and memory displaying/scaling a running BIN VUmeter. The only thing needed is a decent scaling/rendering engine supporting alpha and brightness, such as JSP3 already has.

I don't really know how much time I'll have to monitor this thread so no guarantees on how much engagement I'll have here down the track.

The BIN files contain the background image as is. The needle and LED information is stored as pixel data and there are offsets stored for each frame that points to this data which contains what position in the background image it needs to replace. So the process is to start with the background image, look up the offset for the specific frame being drawn, use that data to replace pixels in the background image with the needle/led pixels, then draw the final image. It's a bit more complicated than that but it's not too far off.

There is a zeroFrame value stored which designates which frame should be used when the peak level is 1.0 (zero decibels).

Some frames use the same data and that's how the bin file creates a scale for the meter. If a meter starts at -40 decibels and another starts at -20, you will find the -20 meter will have a lot more frames with the needle pointing to -20 than the other meter will have pointing to -40. I hope that makes sense.
The calculation to determine which frame to use is Peak Value * zeroFrame. Assuming Foobar returns peak values between 0-1.0

Quote from: ilovefb2k on 2024-11-21 04:27:36

we have just bothered @BoringName at MB 4rum to compile a command-line tool to extract fb2k Vu meter's BIN. https://getmusicbee.com/forum/index.php?topic=41692.msg229960#msg229960

Now I'm up to speed on what you want it for, I'm even more sure it's not going to do what you want. Aside from the issue I already mentioned about VUEditor appearing to anti-alias the needle image when it stores the data in the BIN file, the other problem is you are planning on caching thousands of images. Most bin files are 1024 frames. So you're talking over 2000 images if it has a needle and LED components.

It's been a decade since I've dabbled in javascript and I don't know what limitations you have with the jscript panel but for AIMP skins in my plugin, I just cache the background image and the needle image. Then each frame I draw the background, rotate the needle image based on the peak value and draw it over the top.

I might have it wrong but it looks like you are caching all the possible angles of the needle up front, wouldn't that be using large amounts of VRAM?

I initially tried that method with BIN files and it didn't end well.

If you are intent on supporting BIN files I'd suggest you have a read of the info oops provided and work on drawing them with the approach I mentioned at the start of this post. While you might be able to extract a fully intact needle image with the binExtractor program for some BIN files, I think you will find they will be in the minority.

Quote from: ilovefb2k on 2024-11-17 10:17:05

Quote from: sveakul on 2024-11-16 21:07:51

3.  Is there a setting that functions the same way as "var peak_hold = 10; // in frames" does in the JSP3 Sample VU Meter?

hi @sveakul ,
thank you for your feedback,
3- var peak_hold = 10 : can you shed me a light on what you are after  ? we can have this back or make it as an option.
@ilovefb2k

I thought it might help visually with the extremely slow decay rate of LEDs when "LED follows Peak" is selected.  On the JSP3 bar meters it controls how long the peak line holds before beginning the decay of the line.  May or may not be relevant here but maybe worth an optional inclusion.  Just to clarify, the NEEDLE meter decay rate has always looked fine, but I assume that RMS is what is being used by default with AIMP meters?

I have tried to unzip BIN file by 7zip https://7-zip.org, then following @oops 's instruction to manipulate  that file content by utilizing an  ActiveX Object ADODB.Stream (which can read binary content), but no luck except for very strange characters. 

@Defender
"Text Display + Album Art + Seekbar + Buttons, added Art mousewheel, ReplayGain button and rightclicks"
Thanks for this script.

Request:
Could you please add "Custom Text" to the Contestmenu?
Thank you.

Oops, when I right-clicked on the bottom of the image, the original "Text Display + Album Art" context menu appeared.
I am deeply remorseful.

Quote from: ilovefb2k on 2024-11-21 09:17:10

I have tried to unzip BIN file by 7zip https://7-zip.org, then following @oops 's instruction to manipulate  that file content by utilizing an  ActiveX Object ADODB.Stream (which can read binary content), but no luck except for very strange characters. 

Removing compression is definitely the first step. Finding a C# library that could unzip BZIP2 files properly was one of my biggest hurdles.

Let's see if you can decode the header using javascript and go from there. The first 8 bytes (64 bits) of the file contain the background image width, height, number of frames and zero frame. Each value is 2 bytes.

In C# I use BitConverter.ToUInt16(value,0);
Where value is a byte array containing the 2 bytes I want to convert.

If you can find a javascript equivalent of that command and extract the 4 values from the header that will be a good start.

xmas present

But I am puzzled by this post. Were you able to retrieve the back (L_0), needle (L_1) and glass (L_2) from the BIN?
You state, "we then apply a formula to convert decibel into angle" ... how and what?