Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Director mode not working. #7

Closed
momokrono opened this issue Feb 7, 2017 · 5 comments
Closed

Director mode not working. #7

momokrono opened this issue Feb 7, 2017 · 5 comments

Comments

@momokrono
Copy link

First of all, great software!
The issue I'm reporting is related to the director mode:
I've imagemagick and transcode installed and compiled, but every time i try to render a video it ends up with nothing:
I can see the render screen working, but when it reaches the end nothing happend.
The thing is that i've tryied to modify the directory for tmp images and i can see that pictures are rendered but when all are done the program doesn't produce any video.
Also i cannot see any output file called like "render.avi" but instead only an "animation.fcta".
Plus i cannot write inside the space to select a name to my new animation, i can just select a file from "browse", and even creating an empty file called "animation.avi" and selecting it i end up with nothing but a bunch of pictures.

@edyoung
Copy link
Member

edyoung commented Feb 19, 2017

Thanks for the report. The director code depends a lot on external programs and hasn't been touched in quite a while. I will take a look when i can.

If you have all the individual frames, it may be possible to stitch them together using a suitable package outside of gnofract4d.

@momokrono
Copy link
Author

momokrono commented Feb 19, 2017

Thank you for the reply!
Actually i'm fine having all the frames anyway, and it's not so much an effort convert those into a video.
2 other thing that i've noticed are that when selecting 8 threads (i have an i7) the program crashes, thing that doesn't happen keeping the number of threads low.
Also i cannot increase the number of iterations as long as i want, and after a while i reach a zoom level where i don't have a smooth image, pressing the button to double the iterations or increasing those into the setting area is useless :\

@edyoung
Copy link
Member

edyoung commented Feb 22, 2017

The 'number of threads' thing is interesting, will have to see if I can reproduce that.

Number of iterations != precision.

The issue with zooming is expected and due to the limits of precision offered by processors' floating point implementation. Basically processors don't do exact math, but round things off after 53 binary bits of precision ('double precision'). In most situations that's fine, but in a fractal after you have zoomed in far enough, it's not enough to clearly represent the image, and you start seeing a messed-up image. Some fractal packages provide an 'arbitrary precision' mode which can zoom further at the cost of slowing things down quite a lot, but unfortunately I've not had time to implement that.

@momokrono
Copy link
Author

momokrono commented Feb 22, 2017

Thanks for the explanation!
Basically giving to all the variables involved into the computation of a certain pixel in the picture more that 52 bits at the fraction part should solve the zoom issue, but at the cost of having variables longer than 64 bits with the slowdowns in computing, right?
If it's so i'll try to find a solution by myself.
Anyway I'll keep using this program, and i'll write here if i'll found something new.

@edyoung
Copy link
Member

edyoung commented Apr 22, 2018

I haven't explicitly checked all the issues raised, but was able to generate a video successfully. Please open a new issue if you can still see this using python3 branch

@edyoung edyoung closed this as completed Apr 22, 2018
edyoung pushed a commit that referenced this issue May 23, 2020
* examples: allow custom formula filename in cmake

* [#7] example experiment: formula with arbitrary precision support

* MP example: add swap idiom and other improvements

* add debug information to cpp examples compilation
edyoung added a commit that referenced this issue Jun 7, 2020
* work in progress: create markdown-based manual

* finished first pass through manual.md

* exclude parsetab and lextab from pylint (#102)

* Remove buildtools package (#104)

No executable files ending in get.py are being installed.

* Add pylint singleton-comparison check (#105)

* Remove unused traceback import (#106)

* fix #107: wait until fd has something to read or fail (#108)

* Arbitrary precision formula experiment (#103)

* examples: allow custom formula filename in cmake

* [#7] example experiment: formula with arbitrary precision support

* MP example: add swap idiom and other improvements

* add debug information to cpp examples compilation

* release prep for 4.2 (#111)

* Update http links in README.md and setup.py to https (#115)

* Starting point for a benchmark to compare multiple-precision libs (#112)

* typo

* experiment to benchmark arbitrary precision math

* was running 3 benchmarks

* update #112 (#113)

* install google benchmark globally in exmaples infrastructure

* update exmaples readme

* disable frequency scaling during benchmark

* remove i/o from benchmark loop

* benchmark several different bit lengths

* Basis for discussion on available math libraries

* notes

Co-authored-by: Alberto Gonzalez <[email protected]>

* tweak doc generation process

* Check gnofract4d executable with pylint (#117)

* Install icons into the hicolor theme (#121)

gnofract4d-logo.png is size 640x640.

Individual icons created using ImageMagick, e.g.:
magick gnofract4d-logo.png -strip -resize 256x256 logo/256x256/gnofract4d.png

Install the 48x48 icon into pixmaps.

* Run pylint as a separate job (#118)

Report linting and testing results separately. Only run pylint once.

* C++ Engine refactor (#116)

* tidy worker initialization

* remove unused typedef

* worker: improve AA comparison

* fix docker test script

* remove duplicate extern declaration

* workers: refactor common members into base

* remove some completed todo's

* organize fractfunc initialization and members

* unravel coupling: pointfunc - site

* refactor: extract calculation options into a struct to ease initializers

* remove duplicate

* move rgba type basic operations to its type definition

* remove some temporaries in stworker->work

* review worker member names

* remove unneeded worker alloc members

* remove pointFunc factory

* update some old-fashion idioms on fractfunc and stats

* prefix fractfunc private members

* avoid clearing fates twice when autoupdating iters and period tolerance

* reorganize and narrow down public/private worker interface

* remove unused, mark experimental and review old fashioned code on stfractworker

* calcoptions: add some comments and move asynchronous back to where it belongs

* prevent uninitialized members on fractFunc

* remove initialization success flag on workers

* fix comment

* reorganize fractfunct members and remove unused

* work in progress: create markdown-based manual

* finished first pass through manual.md

* tweak doc generation process

* generate manual with hugo

new directory 'manual' generates standalone HTML for manual

* rest of hugo-based manual

* repoint to our copy of theme

* repoint to our version of theme

* install hugo-generated manual

* update submodule ref

* ignorance

* delete docbook version of manual. so long!

* fix test for doc version

* fix submodule commit

* install hugo in travis

* better errors on doc generation

* more debug output

* jfc

* apt version of hugo is too old, try this

* maybe this

* fold createdocs into setup

* delete old files

* work in progress: create markdown-based manual

* finished first pass through manual.md

* tweak doc generation process

* work in progress: create markdown-based manual

* generate manual with hugo

new directory 'manual' generates standalone HTML for manual

* rest of hugo-based manual

* repoint to our copy of theme

* repoint to our version of theme

* install hugo-generated manual

* update submodule ref

* ignorance

* delete docbook version of manual. so long!

* fix test for doc version

* fix submodule commit

* install hugo in travis

* better errors on doc generation

* more debug output

* jfc

* apt version of hugo is too old, try this

* maybe this

* fold createdocs into setup

* delete old files

* delete gui createdocs

It's more trouble than it is worth.
Getting access violations when calling it from setup.py.
Since this doesn't change much anyway,
easier to maintain commands.html manually

* setup.py updates

distutils to setuptools
make doc generation a custom build step

* merge madness

* build manual first

* custom build command doesn't work :-(

* back to generating docs separately

* checkin generated doc files for consistency

* restore css file

* move css file inside fract4dgui

* fix finding resources after install

* guess which pylint to use

* typo

* disable that darn test

* fix pylint whining

Co-authored-by: Alberto Gonzalez <[email protected]>
Co-authored-by: Chris Mayo <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants