# Weird jump using \`textAscent() and \`textDescent()\`
**URL:** https://discourse.processing.org/t/weird-jump-using-textascent-and-textdescent/39308
**Category:** Coding Questions
**Created:** [October 16, 2022, 9:45pm UTC](https://discourse.processing.org/t/weird-jump-using-textascent-and-textdescent/39308 "2022-10-16T21:45:05Z")
**Posts on this page:** 5
**Page:** 1
### Author:  [@vkbr](https://discourse.processing.org/u/vkbr)
#### Post date: [October 16, 2022, 9:45pm UTC](https://discourse.processing.org/t/weird-jump-using-textascent-and-textdescent/39308/1 "2022-10-16T21:45:05Z")
From the example of [`textAscent()` in reference](https://p5js.org/reference/#/p5/textAscent) I built a littel test. Actually i was going to ask a different question, but then I find a wierd behavior that I can not understand:
There is a jump in values of `textAscent` and `textDescent` when `textSize` changes from `193.13` to `193.15`.
Could not understand why.
the code, heavily based on the example:
```auto
//move mouse to set font size
let base;
let scalar;
let asc;
let dsc;
function setup() {
createCanvas(400, 400);
base = 350;
scalar = 0.8; // Different for each font
}
function draw() {
background(220);
//alter font size - i narrowed dowm the values, but you can change this
textSize(map(mouseY, 0, height, 180, 200));
// asc and desc calc
asc = textAscent() * scalar;
dsc = textDescent() * scalar;
//baseline
stroke(20, 240, 240);
line(0, base, width, base);
//draw asc desc
stroke(240, 240, 10);
line(0, base - asc, width, base - asc);
line(0, base + dsc, width, base + dsc);
//the text
text("dp", 10, base); // Draw text on baseline
}
function mouseMoved() {
console.log(
`text size= ${nf(textSize(), 3, 2)} asc= ${nf(asc, 3, 2)} dsc = ${nf(
dsc,
3,
2
)}`
);
}
```
the sketch in the editor
[p5.js Web Editor](https://editor.p5js.org/v-k-/sketches/fPyL3MmHC)
---
### Author:  [@vkbr](https://discourse.processing.org/u/vkbr)
#### Post date: [October 17, 2022, 1:40am UTC](https://discourse.processing.org/t/weird-jump-using-textascent-and-textdescent/39308/2 "2022-10-17T01:40:42Z")
An static example, am I missing some stupid thing here or maybe this is a bug?
```auto
let base;
let scalar;
let asc;
let dsc;
function setup() {
createCanvas(500, 400);
base = 350;
scalar = 0.8; // Different for each font
}
function draw() {
background(220);
// above 194 weird results?
textSize(194);
// asc and desc calc
asc = textAscent() * scalar;
dsc = textDescent() * scalar;
//baseline
stroke(20, 240, 240);
line(0, base, width, base);
//draw asc desc
stroke(240, 240, 10);
line(0, base - asc, width/2, base - asc);
line(0, base + dsc, width/2, base + dsc);
//the text
text("dp", 10, base); // Draw text on baseline
///// below, 194 (193) the expected result
textSize(193);
// asc and desc recalc
asc = textAscent() * scalar;
dsc = textDescent() * scalar;
//baseline
stroke(20, 240, 240);
line(0, base, width, base);
//draw asc desc
stroke(240, 240, 10);
line(width/2, base - asc, width, base - asc);
line(width/2, base + dsc, width, base + dsc);
//the text
text("dp", 250, base); // Draw text on baseline
}
```
looks like this:
---
### Author:  [@vkbr](https://discourse.processing.org/u/vkbr)
#### Post date: [October 17, 2022, 2:03am UTC](https://discourse.processing.org/t/weird-jump-using-textascent-and-textdescent/39308/3 "2022-10-17T02:03:51Z")
Im done for the night will came back to investigate, weird enough if you change the size of the window, the behavior comes and go… So now there is to many variables to isolate, later I’ll try.
---
### Author:  [@quark](https://discourse.processing.org/u/quark)
#### Post date: [October 17, 2022, 9:35pm UTC](https://discourse.processing.org/t/weird-jump-using-textascent-and-textdescent/39308/4 "2022-10-17T21:35:22Z")
I ran your code using `scalar = 1` so I could check the actual values returned by `textAscent` and `textDescent` but I couldn’t reproduce your result. I got
Although the actual values for `textAscent` and `textDescent` varied with the web browser scale the difference between 193 and 194 was tiny or zero
---
### Author:  [@vkbr](https://discourse.processing.org/u/vkbr)
#### Post date: [October 17, 2022, 10:37pm UTC](https://discourse.processing.org/t/weird-jump-using-textascent-and-textdescent/39308/5 "2022-10-17T22:37:38Z")
Thanks @quark, I haven’t been able to return to this yet. But I’ve noticed a diferent behaviour across browsers. In p5js online editor if you move de division in the page (between the code and the canvas) the lines jumps, even with `scalar = 1`. But I think the editor is not using the last version of p5js… so I’ll need to do some more tests.
Anyway, how the window’s size relate to `textAscent()` and `textDescent()` is a mistery to me.
thanks for your input.