Anime Vision support for 2025 models

[seerge]

Hello fellow community,

I want to add support for "Anime Vision" introduced in the new Strix 2025 models. So I need help from someone with actual Strix 16 or 18 2025.

There is a good chance, it works very similarly to anime-matrix in older Zephyrus devices.

What you need to do

Step 1

Download and run this anciently-looking USB Deivce Viewer

Post a screenshot from it, showing all devices with Vendor ID 0b05 (it's Asus). I need to see if there are new Product IDs for Anime Vision here.

Old Anime Matrix has Product ID 193b, RGB stuff is 19b6

Step 2

Then we need to figure out what commands does Armoury send to "Anime Vision"

1. You will need to install https://desowin.org/usbpcap/
2. Run USBPcapCMD.exe
3. Select correct USB port (containing Anime Vision USB device)
4. Run capture
5. Go to Armoury and change Anime Vision setting to something else (for example turn it ON and OFF, etc)
6. Upload capture log here

Thanks in advance.

[PunkyBust]

Hello @seerge

Thank you very much for your support.
Here is the elements:

USB viewer Vendor ID 0b05 :
The Product ID seems to be the same as older devices.

USB dump:
anime_vision.zip

I stay available if you need anything else to progress.

[seerge]

@PunkyBust ok, amazing. Thanks.

It's indeed seem to be same device ID and furthermore, same protocol.

Can you try this build and check if you see AnimeMatrix block and able to change brightness and set built-in animations - Binary banner / ROG logo (they can be called differently on 2025 device tho).

GHelper.zip

Custom picture and other modes won't look correctly for you most likely, cause matrix size is clearly different from G14 on old devices.

Do you happen to know the actual dot-dimensions of the matrix on your model ? Can you send me some photos of it with some pattern set in Armoury ?

Thanks!

[PunkyBust]

The build is working fine, here is my tests result

Brightness : OK nothing to report every level work properly as the disable too.

Animation:

1. Binary Banner : Diplayed with some glitches

2. ROG logo: Displayed perfectly nothing to report

3. Clock: Not correctly displayed

4. Audio Vizualizer: Noting displayed at all, but in AC I just have the title name in AC not the specter. So, I would says it is normal maybe.

Matrix info and size:
810 leds, 9152 dots.
For the rest I share the complete folder of my Anime Vision, you will find all the animations, config file, & mask.
MatrixDeviceData.zip
In the folder there is some config file with size like below, I don't know wich one is the good :

Pattern in Armory Crate :

[seerge]

@PunkyBust ok, great

1. Binary Banner is built in animation (it's not configured by the app). So I assume it looks same if set in AC ? (it's intentionally "glitchy" by Asus design on G14 too)
2. How many built-in animations does AC offer ? (for old models it was only 2, the ones you can select already)
3. Can you try to set this pattern on Matrix (via Picture button) and send me the photo of how it looks ?

Thanks

[PunkyBust]

Oh ok understand for the binary banner.

2-Detail AC built-in

Animation (6 in total, the first one is default one)

System:

Audio:

3- Pattern Test:
After the upload I have nothing at all:

But if I move the image arround this place it's blink just once:

On the latop the blink give that (we needed two people to take this photo at the right time)

[PunkyBust]

Also, I've test to upload with the build in gif as picture, and the dot are whay too big. I have approximately 10-15 dots filled, for normally one dot.
Even if I lower the size of the image to the minimal it's change nothing.
It give the fill the picture is ultra zoomed.

[seerge]

@PunkyBust ok, thanks

Animation (6 in total, the first one is default one)

Do 2 default animations in G-Helper (banner / logo) have a match in those in AC ? Cause if not, it could be that all modes in AC are actually software-controlled (i.e. it just plays a GIF). A good way to verify that - would be manually stopping LigtingService in windows Services app. If pattern will stop, then it's a software. If not - it's firmware.

Also, I've test to upload with the build in gif as picture, and the dot are whay too big. I have approximately 10-15 dots filled, for normally one dot.

Is it in G-Helper or AC ?

Try this build, I have lowered amount of (assumed) LED dots for your model, so grid-pattern should show. It will still show incorrectly tho.

GHelper.zip

The main problem is that any custom picture is set dot-by-dot by just supplying a long list of dots all together. And since matrix shape is not a rectangular, you need to (mathematically) calculate all this diagonal offsets, etc.

Current logic is built around matrix shape in G14/G16. It's (almost) a rectangle with one cut on the bottom-left

On G835 it's different, it's 2 cuts and who knows how many actual dots are in each row. So I need to guess a lot 🗡️

[kalimaenergy]

Thank for implementation 🥳
I suppose something on the firmware because I have debloat w11 with g815lw without this service and laptop show animation "binary banner" by default. Even when you install os for example

[PunkyBust]

Thanks
Oh so, complicated to count this dot, can I help in any ways ?
Counting them at least on each border ?

Default animations
Yes I have this animation in AC but only on powermaangmeent animation

Clearly the banner/logo is firmware related. When I shut the LightingService it continue to display (no shut at all).

Test AC gif on Ghelper:
Yes, I have test them in Ghelper.
On the last build still too zoomed. Let me share.
If I configure this AC gif in ghleper

I have:

If I test another one AC's gif:

I have:

Test Pattern
I still have the same behavior nothing deplayed, just a blink when I move in certain region.
But if I shut the LightingService, it work and display something

[seerge]

@PunkyBust ok, thanks!

Image blinks / disappears because AC Lighting Service simply overrides it. So for further testing, you will have to stop it.

Can you try this build and set a test grid pattern again? Please use default picture positioning, if you have adjusted it - click Reset under Picture

GHelper.zip

Thanks

[homer-aty]

Hey thanks for working on this. Here is mine with the test grid pattern using that build. Hope that helps

[seerge]

@homer-aty @PunkyBust thanks, can you check if this build and post a picture ? Thanks

GHelper.zip

[PunkyBust]

Hello,
Yere is 👍

Thanks.

[PunkyBust]

If I reduce the size of the pattern and place it like that in ghelper:

I have:

[seerge]

@PunkyBust ok, looks like a move in a wrong direction. Can you try this build instead ?

GHelper.zip

[PunkyBust]

Here is:

By keeping the pattern at it defaut location in ghelper:

Rotation in diagonal:

Rotation in Straight:

[seerge]

@PunkyBust ok, and this ?

GHelper.zip

[PunkyBust]

Straight:

Diagonal:

[seerge]

@PunkyBust thanks, one more attempt

GHelper.zip

[PunkyBust]

Oh way better 🙂

Straight:

Diagonal:

Don't hesitate if you want to test other one.

[seerge]

@PunkyBust hooray. Looks like we have nailed the basics. Can you try different pictures and see if they are displayed correctly w/o glitches ?

[PunkyBust]

Work pretty good on simple pattern, but on complex pattern it's another thing (the two last one below)
I would say that the dots are perhaps too large, which gives this impression of drooling and lack of detail.

Result:

Result:

Result:

Result:

Result:

Result (I got the better result with Bicubic:

[seerge]

@PunkyBust well, it's all displayed correctly. Please keep in mind that a physical resolution of the matrix on G18 is very low (in comparison to G14 or G16 from 2023). So it simply can't display anything detailed. You can try to increase contrast and lower brightness, but don't expect miracles.

Can you check how does Clock mode work ? I think I need to adjust position here, to fit a narrower stripe.

[PunkyBust]

Ok, understand thanks.

Here is the clock:

[seerge]

@PunkyBust thanks. I think to keep things simple it makes sense to show only clock in this case.

Try this build
GHelper.zip

Can you also post a picture how does Audio Visualizer mode look like (you need to play some music in order to trigger it)

Thanks

[homer-aty]

Clock still looks a tiny bit off with the spacing. Basically it's squished to the bottom left too much.
Also the spacing between the 1 and 7 in 17:26 is off.
Any way to make it not military time?

The audio visualizer with audio playing is just one line near the top of the display. It seems like it's the bottom of the visualizer?

[homer-aty]

Hmm i'm going down a rabbit hole here:
Can you help me understand the values for the buffer headers that are being written via AsusHid.WriteAura(buffer)?
Here's what I THINK each value means but is this incorrect?

buffer[0] | Report ID / Device ID | AsusHid.AURA_ID
buffer[1] | Command (per-key update) | 0xBC
buffer[2] | Unknown / Reserved | 0
buffer[3] | Unknown / Reserved | 1
buffer[4] | Unknown / Reserved | 1
buffer[5] | Unknown / Reserved | 1
buffer[6] | LED start index (first LED in packet) | 0
buffer[7] | LED count (how many LEDs in this packet) | 1
buffer[8] | Unknown / Reserved | 0
buffer[9] | Red value for first LED | 0-255
buffer[10] | Green value for first LED | 0-255
buffer[11] | Blue value for first LED | 0-255

[seerge]

@homer-aty yeah, but you don't need to change the structure of the packet :) You can just experiment with trying different addresses

This line contains addresses for 6 zones of a lightbar (LB1 - LB6). So far they seemed to work for all models universally. But you can try to change 169 ... 174 numbers to something else (that is not used for other keys) and see if it suddenly starts affecting your lightbar.

/* LB1   LB2   LB3                                                               ARW?  ARWL? ARWD? ARWR?       LB4   LB5   LB6   */
   174,  173,  172,                                                              120,  140,  141,  143,        171,  170,  169,
       

It could be that there is something in a range form 162 to 168 or after 178 as other addresses are used already :)

[homer-aty]

Ok I think I've finally kind of got it?
The protocol for addressing the lightbar in these laptops is horrendous but here's what I have:

    static int [] g8835rogkeys = { 2, 3, 4, 5, 6 };
     static int [] g8835fnrow = { 21, 23, 24, 25, 26, 28, 29, 30, 31, 33, 34, 35, 36 };
     static int [] g8835numrow = { 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54 };
     static int [] g8835querty = { 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116 };
     static int [] g8835numpad = { 80, 81, 82, 83, 101, 102, 103, 122, 123, 124, 144 };
     static int [] g8835arrowkeys = { 121, 141, 142, 143};
     static int [] g8835logo = {21};
     static int [] g8835leftbar = { 42, 126 };
     static int [] g8835rightbar = { 41, 125 };
     static int[] g8835frontbar = { 127, 145 };

These ints should map out for the g835 various zones. You may notice that some of the addresses are shared and I'm pretty sure I'm still missing some addresses here but for what its worth, this does work.

     public static void AddressLED(int ledIndex, Color color, int repeat = 0)
    {

        if (!backlight) return;
        for (int i = 0; i < repeat; i++)
        {
            Logger.WriteLine($"Addressing LED {ledIndex} with color {color}");
            byte[] buffer = new byte[12];
            buffer[0] = AsusHid.AURA_ID;
            buffer[1] = 0xbc;
            buffer[2] = 0;
            buffer[3] = 1;
            buffer[4] = 1;
            buffer[5] = 1;
            buffer[6] = (byte)ledIndex;
            buffer[7] = 0x01;
            buffer[8] = 0x00;
            buffer[9] = color.R;
            buffer[10] = color.G;
            buffer[11] = color.B;
            AsusHid.WriteAura(buffer);
        }
    }
    public static void AddressLEDArray(int[] ledIndices, params Color[] inputColors)
    {
        if (!backlight || ledIndices == null || ledIndices.Length == 0 || inputColors == null || inputColors.Length == 0) return;

        // If the number of colors matches the number of LEDs, use a per-LED color
        // Otherwise, use the first color for all LEDs
        bool usePerLedColor = (inputColors.Length == ledIndices.Length);
        Color singleColor = inputColors[0];
        Logger.WriteLine($"Using {(usePerLedColor ? "per-LED" : "single")} color for AddressLEDArray. Red: {singleColor.R} Green: {singleColor.G} Blue: {singleColor.B}");

        const int ledsPerPacket = 16;
        int i = 0;
        while (i < ledIndices.Length)
        {
            // Start a new run
            int start = i;
            int runLength = 1;
            while (runLength < ledsPerPacket &&
                (i + runLength) < ledIndices.Length &&
                ledIndices[i + runLength] == ledIndices[i + runLength - 1] + 1)
            {
                runLength++;
            }

            // Prepare buffer for this run
            byte[] buffer = new byte[64];
            buffer[0] = AsusHid.AURA_ID;
            buffer[1] = 0xbc;
            buffer[2] = 0;
            buffer[3] = 1;
            buffer[4] = 1;
            buffer[5] = 1;
            buffer[6] = (byte)ledIndices[start];
            buffer[7] = (byte)runLength;
            buffer[8] = 0;

            // Fill color data per LED
            for (int j = 0; j < runLength; j+= 3)//always send 16 colors starting from bit 9. each color takes 3 bytes. the last 7 bytes are empty
            {
                Color c = usePerLedColor ? inputColors[i] : singleColor;
                buffer[9+j] = (byte) c.R;
                buffer[9+j + 1] = (byte) c.G;
                buffer[9+j + 2] = (byte) c.B;
            }
            Logger.WriteLine($"buffer: {BitConverter.ToString(buffer)}");

            AsusHid.WriteAura(buffer, true);
            i += runLength;
        }
    }`

These functions are helpful to just address a specific pixel or array of pixels;

if you then add

            AddressLED(g8835leftbar[0], color[4]);
            AddressLED(g8835leftbar[1], color[4]);
            AddressLED(g8835rightbar[0], color[7]);
            AddressLED(g8835rightbar[1], color[6]);

            AddressLED(g8835frontbar[0], color[0]);
            AddressLED(g8835frontbar[1], color[1]);

to the end of ApplyDirect ambient mode and heatmap will work on the lightbars of these laptops.

(oh and also this line: AsusHid.WriteAura(new byte[] { AsusHid.AURA_ID, 0xBC}); in ApplyDirect needs to have a 1 as the third argument so:
AsusHid.WriteAura(new byte[] { AsusHid.AURA_ID, 0xBC, 1 });

I'm happy to upload these changes for review if you're OK with it @seerge

[seerge]

@homer-aty hello, thanks :) That's a lot of reverse-engineering you did here (wow).

So do I understand correctly that lots of addresses are actually ... changed? It's quite strange, cause so far all other devices seem to use same table, and only introduce new addresses for some new buttons.

[homer-aty]

So I don't know how many changed exactly and it certainly could be the way the packets are being sent or the protocol that I'm doing wrong (hence why I wanted to understand the protocol better) but it seems at least that the left sided caps, shift, Ctrl and fn lock, the right sided enter key, right Ctrl/copilot key, all have different addresses or don't respond to the addresses in the packet map.
The lighbars seem to share addresses with other keys now as opposed to having their own addresses like before. For example the logo is at address 21 for some reason but so is the escape key. I'm sure this is nonsense at the hardware level but aura just doesn't seem to want to do anything with the lighbars / logo unless you send a command to those addresses.

My testing just involved setting a simple timer and having it iterate sending a command to turn every address from 0 onwards to turn that pixel red. I did the same thing iterating through the packet map to identify which keys changed addresses.

My bet is that aura is expecting some kind of 'handshake' packet as part of the protocol that I'm not sending and that's why the results are so wonky.

[PunkyBust]

Last version of this build, is blocked by my AV.
It was not the case of the previous one, same as virus total trigger it:
https://www.virustotal.com/gui/file/b76244609c85b6e5fd11b6ed0a7547461c9e7b11409fcdb55eb9a766b2a43dfa/detection

Do you know what changes in the application's behavior are causing by this?