PureRef Wrote:
-------------------------------------------------------
> Is previous needed? Previous could always go back
> to the previous one, since if you want a random
> you can just go forward?
Not exactly, I think. I'm not sure how to explain, so I'll show an example/usecase:
We're viewing images randomly very fast and by mistake we skip one that looked i
I don't know if dominant color is enough, because it's just a single value.
You could test two or three types:
a) dominant color
b) calculating percentages of multiple dominant colors, the 5-20 most popular ones, then sort by comparing them all which, which images have the most similar percentages
c) divide the image into 16 parts, 4x4 grid, then calculate dominant color/colors in
An alternative would be to include only the video thumbnail, since it is an image. This would at least allow for management of videos on a PureRef board, for example if we select and export the thumbnails, it would export the original videos into a chosen folder.
File management option "Move original images to", to move selected images from the original folder to a chosen destination. The internal file paths would change as well to the new location, in case we want to move them again.
It would be very useful in several usecases. If we want to manage our images, or delete some of them.
"Copy" option might be useful too. If we wa
Arrange by color composition. I'm not sure how exactly. Using color percentages in various parts of the image? For example if we have a lot of screenshots from the same website, it would display them near each other. Useful if we have thousands of different images in the same folder.
Some random ideas:
- specify the ratio at which we want the rect in which the images are arranged. For example 2:1 - then the arranged images are composed in a rect that is 2x as long as tall.
- add gaps of a specific size between ROWS or COLUMNS of the arranged space. For example a gap 50px between every row of images.
Could be also up arrow button, if it is not used by anything
UP - next random image
DOWN - previous random image, if we want to return to it (skipped by mistake)
Good idea.
I would use the node connections feature as well.
Connections between texts, images, drawings.
Suggested hotkey to make a connection: Press Ctrl while hovering the mouse over the start element. Release Ctrl when mouse is over the end element. Connect both elements, Start -> End.
One element can be connected to multiple elements.
If Ctrl hotkey is already taken, use Tab, Z or other.
Mode: Cache on disk
Cache takes about 1.4GB on disk.
172 images, 162MB, some as large as 6MB (not stacked on top of each other).
RAM usage: 1.5GB to 2.0GB (without a project)
After saving the project and reloading: ~850MB to ~1100MB
It's nice, better than the memory options.
Could it be that it's not always unloading images? (If RAM should be lower than 850MB). I noticed tha
It helps!
Sometimes the filenames are too long to be displayed fully, but it's not a big issue.
Option to keep them visible on all images would be useful, but it's not a big priority.
Thanks!
I am using PureRef 2.0. I don't have any scenes from previous versions.
It still doesn't export images from new scenes in the original resolution. I tried both "Export all images" and "Export selected images" options.
I tried both with saving the scene first, and without saving (just opening PureRef and loading images, then exporting).
Hi
It would be really useful to have an additional option to export images in original (full) resolution, even when in "Auto downscale" mode.
My usecase:
1. Load images with "Auto downscale" set to 512x512 or 256x256
2. Export images
3. The exported images are not the original resolution, but downscaled
Thanks!
I checked the Cache on disk mode. It still performs worse (in terms of RAM usage, 2-3 GB).
Looks like I will keep using the downscaling mode (lowest RAM usage), but thanks for help!
I can confirm. The error shows up when trying to export as "Original".
The other options work without error (jpg/png/bmp).
I'm on linux.
1. Loaded a few .jpg images into PureRef.
2. Tried to export as Original.
3. Error.
+
I would also use the option to copy/paste images into a file explorer. Currently the copy option works only for copying image data into other apps.
Suggested use:
1. Select image(s)
2. Click Ctrl+C or a popup menu "Copy" (if it can't be done with the current copy, then a separate "Copy file"?)
3. Paste into a file explorer
Edit: Nevermind, there already exists
I tried, but the "No cache" performs worse than Downscaling.
Comparison (170 images):
- "In memory" Performance Mode - RAM usage 2 GB, works smoothly
- "No cache" Performance Mode - RAM usage between 1 to 2 GB, terrible lags whem zooming/moving the canvas, the images look worse than in the Downscaling mode when zoomed out
- Auto Downscaling (512x512) "In me
I missed that, thanks!
A similar option (background color) could be useful for drawing concepts (the draw object), but it's just an idea for later to consider.
1. Text should have an option to set its background color, to differentiate it more from the rest. Transparent by default, possible to change for each text and in the default one in settings.
2. We have 3 line options: norma, dotted, normal arrow. I suggest to add dotted arrow.
The memory optimization works great, but I would love a way to load the image fully after focusing/double-clicking it. The image would remain loaded until we focus another one.
Loading images in a slideshow shouldn't happen, unless we focus one. Alternatively preload them during previous images shown.
Thanks.
A way to further optimize the memory usage is to convert the images from 4-byte colors to 1-byte palette index after loading. Either grayscale or colors.
The current behaviour of the random option makes it possible for images to repeat, for example 3->5->3->1->5.
I want to propose a new "Non-repeating" setting, which would guarantee a non-repeating sequence: 3->5->2->4->1. The sequence can be generated again after it finishes.
It would be useful when we want to make sure we slide through all of the images, but in
