Statistics of Xournal++ files (.xopp
)
#4862
Replies: 7 comments 8 replies
-
Hello! |
Beta Was this translation helpful? Give feedback.
-
Hi I´m a student and use xournal a lot. Almost all my notes are in xournal and I do homeworks, practice and much more in it. I gathered some documents in here: https://1drv.ms/f/s!ApoA9il2jxFHg_F2kld8L9sLZc0taA?e=jqJdal With big documents it starts to bug and slow down, I usually download what I had to a pdf and create a new document with that so it doesn´t slow down as much. Hope it helps. |
Beta Was this translation helpful? Give feedback.
-
Sorry for interrupting the interesting discussion here, but can we just clearly distinguish between Xournal and Xournal++? Although Xournal++ is a rewrite of Xournal, the two codebases are almost completely distinct and what applies to Xournal++ may not apply to Xournal and vice versa. I guess this discussion is all about Xournal++, but I could be wrong. |
Beta Was this translation helpful? Give feedback.
-
An example of the way I use xournal++ to study math https://drive.google.com/file/d/18jYy4cM3EE3j547lDJvDggMNepvljqAQ/view?usp=drive_link |
Beta Was this translation helpful? Give feedback.
-
8° Ano Mat. Bianchini 2015 Cap. 1 - Retas e ângulos Atividades.pdf |
Beta Was this translation helpful? Give feedback.
-
I don't know if what I'm going to ask this linked , but if it's about the pressure sensitivity but when I write I often have the feeling that it doesn't feel so natural , previously I had a wacom ctl4100 tablet and the strokes came out distorted , or when I tried to write with a small calligraphy they came out big strokes or if I tried to write fast like on paper , the strokes didn't have the shape I wanted , then I found out that it could be related to the response times of the pen , the wacom I mentioned had 133PPS, so I bought another graphic tablet a little more expensive a Huion dial 2 that claims to have 300PPS and to my surprise it improved a little, but I still notice that some strokes visually look distorted, I do not want to be wrong because I do not know what would be the cause, or maybe it has to do with some issue of smoothing, or rendering tambien i don't know why whenever i start or finish drawing some letters start or end with a sharp point like an l or p . |
Beta Was this translation helpful? Give feedback.
-
Would it be possible to share the set of files? |
Beta Was this translation helpful? Give feedback.
-
For the developers and curious people. Do you know how many
strokes
are in a typical Xournal++ file? How the pen pressure is distributed? How large are the images that people attach to?Working on #937 , before designing the new file format, the first step is to know the current one.
Here are some plots from a 48 files dataset collected from 9 different public sources.
How large are the texts, images and Tex-images?
As expected images and Tex-images are quite heavy, ~10K and ~100K. Note however that they are stored as Base64 so the numbers are ~1.33 times the real sizes.
Texts are much smaller but they are quite more numerous.
Note: both axes are in log scale
![image](https://user-images.githubusercontent.com/2665522/236442011-cb7fff59-4d57-4967-9597-00d91addb168.png)
How many coordinates a stroke has?
Most of the strokes have around 100 coordinates but some outliers almost reach 1000!
Do you see the weird pattern of missing vertical lines in the left part of plot? This is because the strokes have a even number of coords (they are x-y points)
Note: the x axis is in log scale
![image](https://user-images.githubusercontent.com/2665522/236443338-3e478d3b-2a4d-4737-836c-c5a9ad28cb91.png)
And how many strokes do we have?
If you ask per page you will get this:
![image](https://user-images.githubusercontent.com/2665522/236443494-6c5cd206-6728-43f7-bd5b-e18298dd91dd.png)
![image](https://user-images.githubusercontent.com/2665522/236443896-6d3bb174-e0e5-49d4-a110-942e7f9d984c.png)
If you ask in total per sample file you will get this (x axis in log scale):
Some pages have more than 1000 strokes and some samples docs have more than 10000 strokes (and it may reach 100000)
Talking of strokes, how are theirs x-y coords distributed?
Mean and standard deviation for x-coord (y-coord is similar):
![image](https://user-images.githubusercontent.com/2665522/236444675-049563ca-f8bd-4308-b7ef-b030ec1b2967.png)
Is any relationship between x and y?
No strong correlation but there are some features: it looks that certain y values are more common than others (look carefully the plot and you will see some darker horizontal lines)
You may also spotted some weirdness: why there is a missing chunk in the plot? in the range of (600 - 800) x and y axes?
My intuition says that most of the pages are 800x600 and sometimes the pages are in landscape orientation giving the apparent size of 600x800.
What about the pen pressure?
The darker point at mean and std equals to 0 corresponds to the strokes that they don't have pressure points
Distribution of background page types.
Looks like
![image](https://user-images.githubusercontent.com/2665522/236441570-6a6b1ee9-7ee6-4bca-ac62-54b8187b93e3.png)
solid
is most common butpdf
is not far away.Do you want contribute?
Share your
.xopp
docs and I'll add it to the dataset. With more samples we can get a better understanding of how you and others use Xournal++ and we can design the software better.Beta Was this translation helpful? Give feedback.
All reactions