For me it’s the other way around xD because they take Functionality away, and keep administration of cuts so basic that even VirtualDub’s cut administration (at least it can copy & paste frames around) looks like alien science.
Video AI is a Tool, VirtualDub is a tool.
A Video Editor is a highly functional, highly flexlible, complex, cluttered (most times ^^) Application.
Multiple Tracks - that can be filled with different Video/Audio sources, snippets/cuts, copy & paste, drag and drop, filtering Snippets, blending etc etc
Back to Video Ai
I find the Tab Idea in V6 not bad (especialy for fast Comparing, just by click) because people got less vertical space as horzontal, and because V5 used a Tree List and not just a Simple List (which is plain and simple just better !)
Tree List
- Artemis V13 - Medium
Frame 0-2000
- Proteus V4
Frame 2001 - 2005 (Extreme Blocking Section - Settings 1)
- Proteus V4 ← Bad ! Tree Group Entry for same Frame Range and same Model
Frame 2001 - 2005 (Extreme Blocking Section - Settings 2)
- Artemis Strong Dehalo
Frame 2006 - 2100
Tree List - Much harder to Sort, Vertical Space Waste, less oversight, uses a HashMap with Array/Lists as underlying Backbone ← More Risk for memory Leaks, more wasted Ram etc etc
Simple List
Frame 0 - 2000 | Tag: Stable Quality | Artemis Medium , Settings …
Frame 2001 - 2005 | Tag: Extreme Blocking | Proteus4, Settings 1
Frame 2001 - 2005 | Tag: Extreme Blocking | Proteus4, Settings 2
Frame 2006 - 2100 | Tag: Stable Quality | Artemis Dehalo, Settings
Simple List - Can be sorted after Frames/tags/Settings, no space Waste, more oversight.
Back to Tree List from V5
There was no real Functionality inside.
If i see a Box with a “V” or “X” inside, i think out of reflex (i grow up with MS-DOS and use a lot Checkboxes in my Tools) that, it’s a CheckBox.
I click on the Check Box, nothing happens, so just a indicator (i would Change the Color Letter inside the Entry to Green, and use the Box as Checkbox. Checked by me means → Ok ! i’m happy with the Result.) that it’s done ? Okaaaay.
I try to delete a Entry inside a Tree Entry, doenst work. Realy ? Why is there then a Tree List in the first place ? I dont get it.
I *rew around in the timelime, try things out on different parts, and the Tree View gets cluttered like that.
- Artemis V13 - Medium - Recover 70
Frame 0-2000
Frame 2001 - 2005
Frame 4000 - 5000
- Artemis V13 - Medium - Recover 69 ^^ ← Wow ! Another Tree Entry because of - 1
Frame 2001 - 2005
- Proteus V4 - Settings 1
Frame 2001 - 2005
- Proteus V4 - Settings 2
Frame 2001 - 2005
- Artemis Strong Dehalo
Frame 2006 - 2100
- Proteus V3
Frame 4000 - 5000
No i got 4 Times 2001 - 2005 ! but i can only delete 3 of them ! because i cant just remove a Entry inside a Tree Entry (Group ^^) i can just Remove ““Groups”” (Tree Entrys)
A Hashmap is a good thing when it makes sense.
Normal List (one dimension, no Fields, anything 1 Line of String) with 100k Entrys
----------------- First Entry is 0 ! Computer Count different.
superman1, plz 827267
superman19, plz 827264
papaschlumpf, plz 82736
5uperm4n1, plz 826627
..
..
Now if i want to find out if there allready exist a entry with “papaschlumpf, plz 82736” the Computer must search from Top to Bottom and break the Search when Search Term was found.
Even 100k Entrys are nothing, it’s blazing fast ! (i know that for shure ^^ i saw many big Lists in my life) even back in 2003 something like that was very fast.
HashMap with 100k Entrys
----------------------- Can’t be Sorted !
- superman1, plz 827267
- superman19, plz 827264
- papaschlumpf, plz 82736
- 5uperm4n1, plz 826627
Now with a HashMap, the “Search” and access Time is much faster. It just Calculates the Hash Sum from the Search Term “papaschlumpf, plz 82736” and by Math Magic it know’s were it is located in the Ram.
Sounds fine !
You can even add a Sub List/Array (with or without Tables/Columns) to each Entry from the HashMap ! But you cant use Sub HashMaps in Lists/Arrays (so the other way around) if i remember it right.
So it looks like a HashMap has just advantages or not ?
Not realy.
I can use Tables/Columns (Data Structures) with Lists/Arrays (Arrays are much faster as Lists @ Random Access. 25-30 Years ago, Lists were faster) too !
- superman1, plz 827267, 1981, bla , bla2
- superman19, plz 827264, 1969 ^^, bla30 , bla25
- papaschlumpf, plz 82736, 1988, bla12 , bla22
- 5uperm4n1, plz 826627, 1977, bla87 , bla29
and i can Sort them like i want ! I can Sort after Name, after Zip Code etc etc ! so i can much better, and easier visualize and transfer the Content from the >internal< Backbone to the >external< visible List !
That Topaz has a Problem to transfer (and maybe that some crazy memory leak is going on. Arrays especialy with Tables/Columns are the worst, when it comes to that ) becomes obvious, when many different Group/Tree Entrys with many Sub Entrys are in the Preview List.
The Gui starts to react very sluggish, which is embarrassing, for anything under 5k visible Entrys !
2 Possiblitys:
Insane Memory Leaks & way to Long “Search/Access Term” for the Hash Map Entrys
Or
FileSystem (MFT in NTFS) get’s cluttered and accessing Directory with Preview becomes very slow, because of a Ton a PNG Files ! (The More Files a Folder Contains, the Slower the Access Time. Some Filesystems slow down faster with many files, some slower)
Solution ? Shorter More Effective Filenames, distributed in more SubFolders !
Input File. SomeChick - SomeMovie.mp4
d:\AITMP\SomeChick - SomeMovie.mp4\TH\F0000-1999\
0000.png
0001.png
..
..
d:\AITMP\SomeChick - SomeMovie.mp4\TH\F2000-3999\
2000.png
2001.png
..
..
or like that.
Input File. SomeChick - SomeMovie.mp4 → ID0
Good >Visible< ID in Video AI !
d:\AITMP\ID0000\TH\F0000-1999\
0000.png
0001.png
..
..
d:\AITMP\ID0000\TH\F2000-3999\
2000.png
2001.png
..
..
I did stuff like that and searching stuff in huge Lists, a Ton in my Life. It’s easy !
If they can write an AI they and got a Team of Coders then they can do that to without problems.