Drawing text ingame using Font.draw

[XPModder]

Hi.

I am working on a mod and wanted to render some text during the LevelLastRenderEvent. I am trying to draw the Text using Minecraft.getInstance().font.draw(), by passing in the PoseStack from the event, as well as my Text.

For some reason I can't get it to actually render anything.
I am not quite sure what the x and y coordinates mean, that the method requires, but I tried many different values. I also looked at some other mods and tried to find an example of this actually being done, but I haven't really found something helpful...
Are the x and y parameters of the function world coordinates (like block coordinates) or are they relative to something else?

Ho do I find the correct coordinates to use?

Also, why is it not rendering anything at all for me right now?

 

Here is how I am trying to render the text at the moment.

PoseStack matrixStack = event.getPoseStack();

String text = "Collectable XP: ";
text += entity.getCollectedXP();

LogHelper.info("Render: " + text);

matrixStack.pushPose();

matrixStack.scale(1, 1, 1);

instance.font.draw(matrixStack, text, 7, 100, 0xffffff);

matrixStack.popPose();

This is inside an eventhandler for the LevelLastRenderEvent and with the LogHelper printing it to the log, I can see that it is definitively getting to this point and executing it.

The logging of this is just temporary, so I know that its actually trying to render...

[warjort]

For 1.18? See OverlayRegistry for how to register an IIngameOverlay as part of the in game gui.

e.g. see ForgeInGameGui.HUD_TEXT that draws the debug text when you press F3 (amongst other things).

 

If you really do want to draw text in the world, look at the some vanilla code that does this, e.g. EntityRenderer.renderNameTag()

[XPModder]

Yes. It is for 1.18.2.

I do want to render the text in world, as the text will show some data about a specific type of blockentity.

Basically I want to draw some text above the block, when it is being looked at by the player. The text will show information about the blockentity.

I have looked at EntityRenderer.renderNameTag() now, but I am still having problems.

Unfortunately all the parameters in both the renderNameTag method, as well as all of the methods of the Font class do not have meaningful names, which makes it a lot harder to find out how something is working. I am still not sure what some of the parameters do.

I have now tried to use the method font.drawInBatch instead of font.draw, as this is what the renderNameTag method uses, but I still can't get any text to actually show up ingame.

[Luis_ST]

  On 8/11/2022 at 7:05 PM, XPModder said:

Unfortunately all the parameters in both the renderNameTag method, as well as all of the methods of the Font class do not have meaningful names, which makes it a lot harder to find out how something is working. I am still not sure what some of the parameters do.

Expand  

https://github.com/ParchmentMC/Parchment/wiki/Getting-Started

[warjort]

This isn't really my area of expertise but..

 

You will need to do more than just drawing the text, you also need to change the pose (translate, rotate, etc.) to match where to draw it in the world.

e.g. for the entity rendering this is done in the EntityRendererDispatcher.render(). But renderNameTag() makes some additional changes so the text is correctly aligned to the camera.

 

For blocks I guess you can get the basic logic by looking at where it calls BlockEntityRendererDispatcher.render() in LevelRenderer.render().

 

[warjort]

If it just for the block being looked at, you might find subscribing for DrawSelectionEvent.HightlightBlock more useful? 

[XPModder]

  On 8/11/2022 at 8:26 PM, warjort said:

If it just for the block being looked at, you might find subscribing for DrawSelectionEvent.HightlightBlock more useful? 

Expand  

I will look into that.

 

I looked further into the renderNameTag method and also where it is being called from.

I still don't know 100% how everything works, but the renderNameTag method only translates the pose only by adjusting the y value of it. It does not modify the y and z translation.

So I tried doing essentially the same: translating in the y direction and leaving  x and z as is. I also adjusted the text to the camera, similar to the way it is done in renderNameTag.

 

I also looked at another project of mine, where I rendered a block in the world, that isn't actually there, but that project uses getActiveRenderInfo().getProjectedView() in the translation. This other project is sitting in 1.16.5 and there it works, but the ActiveRenderInfo class does not exist anymore and I can't find anything that would be the equivalent to getProjectedView() in 1.18.2.

I still havent been able to get it to work.

My code looks like this now:

                    PoseStack matrixStack = event.getPoseStack();

                    String text = "Collectable XP: ";
                    text += entity.getCollectedXP();

                    LogHelper.info("Render: " + text);

                    matrixStack.pushPose();

                    matrixStack.scale(1, 1, 1);

                    matrixStack.translate(0, 1, 0);

                    matrixStack.mulPose(instance.getEntityRenderDispatcher().cameraOrientation());

                    MultiBufferSource bufferSource = instance.renderBuffers().bufferSource();

                    int packedLight = instance.getEntityRenderDispatcher().getPackedLightCoords(playerIn, 0);

                    //instance.font.draw(matrixStack, text, 7, 80, 0xffffff);
                    instance.font.drawInBatch(text, 7, 10, 0x20ffffff, false, matrixStack.last().pose(), bufferSource, false, 0, packedLight);

                    matrixStack.popPose();

 

[warjort]

  Quote

So I tried doing essentially the same: translating in the y direction and leaving  x and z as is. I also adjusted the text to the camera, similar to the way it is done in renderNameTag.

Expand  

It moves it up by the height of the entity model + half a block and it also does the mulPose and scale.

You are missing the part where it initially translates to the block position, see my comments above. 

 

This doesn't make sense to me.

  Quote

matrixStack.mulPose(instance.getEntityRenderDispatcher().cameraOrientation());

Expand  

This is just the value of camera.rotation() in the last call to EntityRenderDispatcher().prepare()

I doubt this is intended to be used outside the entity rendering calls themselves.

 

Also, if you use the HighlightBlock event you will be given a buffer source in the event state.

You also get a camera.

 

See this thread for tools/info to help translate old mods to the newer versions.

https://forums.minecraftforge.net/topic/115137-1192-translating-classesmethods-from-1165/

 

[XPModder]

  On 8/11/2022 at 9:43 PM, warjort said:

You are missing the part where it initially translates to the block position, see my comments above.

Expand  

Can you show me where that is actually done in the game code?

If I look at the EntityRenderer class I can only find a single call to translate(), which is the one that moves it to the correct height.

 

  On 8/11/2022 at 9:43 PM, warjort said:

Also, if you use the HighlightBlock event you will be given a buffer source in the event state.

You also get a camera.

Expand  

Yes, I see that. I have now moved my code to the HighlightBlock event and inside the mulPose() call I now have event.getCamera.getRotation(). I also adjusted everything else to use the additional information availbale through that event.

 

I also tried to translate the pose to the block position, but it didn't fix the problem. I also couldn't see that being done anywhere in the vanilla code, so I assumed it was not neccessary....

[warjort]

Forget the EntityRenderer (it was EntityRendererDispatcher anyway). I was just using that to explain the translation is in 2 steps.

* Step 1 - translate to the relative entity/block position

* Step 2 - do additional block/entity specific changes like moving the name plate upwards (inside the Renderer)

 

It is done this way so people that write BlockEntityRenderers or EntityRenderers don't have to deal with step 1. It is done for them.

You have neither, so you need to do both steps.

 

As I said above, an example of the code you want is in the LevelRenderer for the BlockEntityRenderDispatcher (d0, d1, d2 are the x,y,z of the camera)

If you are going to force me to feed it to you on a spoon...

  Quote

            BlockPos blockpos3 = blockentity.getBlockPos();
            p_109600_.pushPose();

 

            // HERE
            p_109600_.translate((double)blockpos3.getX() - d0, (double)blockpos3.getY() - d1, (double)blockpos3.getZ() - d2);
 

          this.blockEntityRenderDispatcher.render(blockentity, p_109601_, p_109600_, multibuffersource$buffersource);
            p_109600_.popPose();
 

Expand  

 

[warjort]

As I said above, this is not my area of expertise. There is something of the blind leading the blind here. 🙂

[XPModder]

Ok. I think I understand it now.

The pose is translated first by the difference between camera and block position to move it to where the block is. Then it is translated upwards, so it is above the block/entity and not inside it.

I am doing that now, but it still doesn't fix the problem...

PoseStack matrixStack = event.getPoseStack();

String text = "Collectable XP: ";
text += entity.getCollectedXP();

LogHelper.info("Render: " + text);

matrixStack.pushPose();

Vec3 cameraPosition = event.getCamera().getPosition();

matrixStack.translate((double)hitPos.getX() - cameraPosition.x, (double)hitPos.getY() - cameraPosition.y, (double)hitPos.getZ() - cameraPosition.z);

matrixStack.scale(1, 1, 1);

matrixStack.translate(0, 1.5, 0);

matrixStack.mulPose(event.getCamera().rotation());

MultiBufferSource bufferSource = event.getMultiBufferSource();

int packedLight = instance.getEntityRenderDispatcher().getPackedLightCoords(instance.player, event.getPartialTicks());

//instance.font.draw(matrixStack, text, 7, 80, 0xffffff);
instance.font.drawInBatch(text, 7, 10, 0x20ffffff, false, matrixStack.last().pose(), bufferSource, false, 0, packedLight);

matrixStack.popPose();

It still doesn't draw the text ingame.

 

I think I understand how most of the game works, but all the graphics and rendering stuff is still somewhat beyond my understanding...

 

  On 8/11/2022 at 10:54 PM, warjort said:

As I said above, this is not my area of expertise. There is something of the blind leading the blind here. 🙂

Expand  

Well, we can hope that someone else, who understands this better, picks up this thread and helps us out...

[warjort]

I had a go at implementing something like what you are doing, this is what I came up with based on your code:

    public static void highlightBlock(DrawSelectionEvent.HighlightBlock event) {
        Minecraft instance = Minecraft.getInstance();
        BlockHitResult hit = event.getTarget();
        Vec3 pos = hit.getLocation();
        Component text = new TextComponent("test");

        PoseStack matrixStack = event.getPoseStack();
        matrixStack.pushPose();
        Vec3 cameraPosition = event.getCamera().getPosition();
        Quaternion rotation = event.getCamera().rotation();
        matrixStack.translate((double) pos.x - cameraPosition.x, (double) pos.y - cameraPosition.y, (double) pos.z - cameraPosition.z);
        matrixStack.scale(-0.025F, -0.025F, -0.025F);
        matrixStack.mulPose(rotation);

        Font font = instance.font;
        float f2 = (float)(-font.width(text) / 2);
        MultiBufferSource bufferSource = event.getMultiBufferSource();
        int packedLight = instance.getEntityRenderDispatcher().getPackedLightCoords(instance.player, event.getPartialTicks());
        instance.font.drawInBatch(text, f2, 0f, -1, false, matrixStack.last().pose(), bufferSource, false, 0, packedLight);
        matrixStack.popPose();
    }

This is not perfect since it actually draws the text kind of inside the block. Something you can see is especially a problem when you look at solid block from above.

I'm sure you can fix this and other shortcomings. 🙂

Edited August 12, 2022 by warjort

[XPModder]

I got it to work now!

The problem apperently was the scaling of the matrixStack. For some reason with the matrixStack.scale(1, 1, 1) it does not create any text at all, but as soon as I changed it to -0.025F, it worked!

That was the only change required to make it work in some capacity. I later made some other changes to, but they were not neccersary to make the text show up ingame.

I have no idea why that is the case, as I thought, the text would simply be way to big, if the scale was of...

Seing as it needs a negative scale, maybe scaling it to 1, 1, 1 made the text super small. So small that it wasnt visible....

 

Also: any translation should be done before the scaling, otherwise its way off...

 

Thanks for the help!

[warjort]

A negative scaling means a reflection and I guess the reflection in the y direction is because world coords are "updside down" compared to screen coords (which I think is what the font drawing uses?)

 

I tried your 1,1,1 scaling and it is drawing the text, just not where you can see it in first person. You can see the bottom of the backwards large text when you press F5. 🙂

Edited August 12, 2022 by warjort