Quick Start
Windows
The command line app for Windows comes bundled with Topaz Video Enhance AI.
If you look inside the “C:\Program Files\Topaz Labs LLC\Topaz Video Enhance AI” folder (or equivalent location if you installed to a non default location) you will see it in the same folder with Topaz Video Enhance AI.exe.
Open the Windows command line interface and type:
cd "C:\Program Files\Topaz Labs LLC\Topaz Video Enhance AI"
veai.exe --help
MacOS
Open the MacOS terminal and type:
cd /Applications/Topaz\ Labs\ LLC/Topaz\ Video\ Enhance\ AI/Topaz\ Video\ Enhance\ AI.app/Contents/MacOS/
./Topaz\ Video\ Enhance\ AI --help
Command Line Options
-h, --help Displays help on commandline options.
-v, --version Displays version information.
-i, --input < filepath > (mandatory) input video file or the path to the first image of an image sequence.
accepted formats for video: [.avi, .flv, .m1v, .mkv, .mov, .mp4, .mpg, .mpeg, .webm, .wmv]
accepted formats for image sequence: [.png, .tif]
-o, --output < filepath | folderpath > (optional) output video file or image sequence folder path. If using image sequences the folder must exist.
-f, --output-format < mp4_h264 | mov_prores | mov_proreshq | jpg | png | 8tif | 16tif >
mp4 - H.264 encoded video file (low-quality)
jpg - for 8 bit compressed jpg sequence
png - for 8 bit lossless png sequence
8tif - for 8 bit uncompressed tiff sequence
16tif - (default) for 16 bit uncompressed tiff sequence
* 16 bit tiff is usually rather large, please make sure you have enough hard drive space available
-b, --begin-frame <begin-frame> (optional) the 1st frame to be process into output image sequence.
-e, --end-frame <end-frame> (optional) the last frame to be process into output image sequence.
-s, --scale <scale> scale the size of output video [1.0 - 6.0] where 1.0 = 100% and 6.0 = 600%
-l, --fill-mode <fill-mode> [fill|letterbox] when -scale is not specified, use fill|letterbox mode to format output frame. default: fill
-d, --width:height <width:height> output video frame size, must be in W:H format
-m, --ai-index <model shortname> the shortname of the ai model
aaa-9 - Artemis Anti Aliasing
ahq-11 - Artemis HQ
alq-12 - Artemis LQ
alqs-1
amq-12 - Artemis MQ
amqs-1
ddv-2
dtd-3
dts-1
dtv-3
dtvs-1
thd-3 - Theia Detail (selected by default)
thf-4 - Theia Fidelity
gcg-5 - Gaia CG
ghq-5 - Gaia HQ
-c, --cuda < 0 | 1 | 2 | 3 | 10 > set cuda device index. 10 will set toCPU. Choose 0 to run on card 0, 1 to run on card 1 and so on.
-p, --sharpen <0-100> [Theia models only] - Amount of sharpening for output video [0-100] Defaults to 15.
-g, --enhance-detail <0-100> [Theia models only] - Amount of compression, affects the amount of detail that can be kept. [0-100]. Defaults to 50.
-r, --reduce-noise <0-100> [Theia models only] - Reduce noise. Values can be from 0 to 100. [0-100]. Defaults to 0.
Lets walk through the options!
Input Flag
-i, --input < filepath > (mandatory)
input video file or image sequence path.
accepted formats: [.avi, .flv, .m1v, .mkv, .mov, .mp4, .mpg, .mpeg, .webm, .wmv]
Video output is always written to the same folder as the input
You can drag and drop your video file onto the command line window to get the command line to fill in the path for you.
veai.exe --input C:\Users\...\file_example_AVI_480_750kB.mp4
Hit enter and it will start processing with the default parameters.
Output Format Flag
This flag is used to set the output format of the video to one of the following options.
-f, --output-format < mp4 | jpg | png | 8tif | 16tif >
mp4 - H.263 encoded video file (low-quality)
jpg - for 8 bit compressed jpg sequence
png - for 8 bit lossless png sequence
8tif - for 8 bit uncompressed tiff sequence
16tif - (default) for 16 bit uncompressed tiff sequence
* 16 bit tiff is usually rather large, please make sure you have
enough hard drive space available
For example we can append this flag to our previous command to change the output.
veai.exe --input C:\...\file_example_AVI_480_750kB.mp4 --output-format jpg
veai.exe --input C:\...\file_example_AVI_480_750kB.mp4 --output-format png
veai.exe --input C:\...\file_example_AVI_480_750kB.mp4 --output-format 16tif
Begin and End Flags
-b, --begin-frame <begin-frame> (optional) the 1st frame to be process into output video. Only valid with video inputs. Do not use for image sequence inputs.
-e, --end-frame <end-frame> (optional) the last frame to be process into output video. Only valid with video inputs. Do not use for image sequence inputs.
Use this to specify a subsection of the input video that you want exported to image sequence.
Scale Flag
-s, --scale <scale> scale the size of output video [1.0 - 6.0] where 1.0 = 100% and 6.0 = 600%
The scale parameter is used to set my how much the input video should be scaled by. The width and height of the input video will be multiplied by this value to give the output dimensions.
Dimension Flag
-d, --width:height <width:height> output video frame size, must be in W:H format
For manually specifying the height and width of the output video. You can use this paired with --fill-mode to specify if the video should be cropped to fit the new size with the output ratio, or if it should fill the space with black using the letterbox option.
Fill Mode
-l, --fill-mode <fill-mode> [fill|letterbox] when -scale is not specified, use fill|letterbox mode to format output frame. default: fill
When used with the dimension flag this specifies how to handle if the input video is a different aspect ratio than the output.
Model Selection
Next we have the option for specifying the AI model you want to use.
-m, --ai-index <model shortname> the shortname of the ai model
thd-2 - Theia Detail (selected by default)
thf-2 - Theia Fidelity
gcg-1.0.1 - Gaia CG
ghq-1.0.1 - Gaia HQ
ahq-1.0.1 - Artemis HQ
alq-1.0.1 - Artemis LQ
If nothing is entered you will process using the Theia Detail model by default.
CUDA Device Selection [Windows Only]
Note: On MacOS all devices are used by default.
-c, --cuda < 0 | 1 | 2 | 3 | 10 > set cuda device index. 10 will set toCPU. Choose 0 to run on card 0, 1 to run on card 1 and so on.
You can process on multiple cards at once by opening two separate CMD sessions in and running one with -c 0 and the other with -c 1 for example,
veai.exe --input C:\Users\Topaz\Downloads\file_example_AVI_480_750kB.mp4 -c 0
veai.exe --input C:\Users\Topaz\Downloads\file_example_AVI_480_750kB.mp4 -c 1
Using -c 10 will set it to process on the CPU. Keep in mind that CPU processing is nearly always slower than GPU processing.
Theia Model Parameters
-p, --sharpen <0-100> [Theia models only] - Amount of sharpening for output video [0-100] Defaults to 15.
-g, --enhance-detail <0-100> [Theia models only] - Amount of compression, affects the amount of detail that can be kept. [0-100]. Defaults to 50.
-r, --reduce-noise <0-100> [Theia models only] - Reduce noise. Values can be from 0 to 100. [0-100]. Defaults to 0.
You can only use these if your ai-model index is 0 or 1. (ie. if it is a Theia model)
These set sharpness, detail enhancement, and noise reduction respectively.
Troubleshooting and Notes
If you see errors like the one below you can safely ignore it.
<FileAppender::FileAppender> File name is empty. The appender will do nothing
The program will launch the GUI and you may have to login if you are not logged in already
After login is done, you may have to click process to start it. However, if no login is needed, the program will start processing automatically, and then quit automatically when it is done.
You can see in the command line output where it is saving the finished image sequence.
Saving: C:/Users/Topaz/Downloads/file_example_AVI_480_750kB_2.00x_960x540_Theia-Detail_50_15_0_tif16/
By default the program always saves to the same folder as the input.
Note: the command line app still launches the GUI, unlike many command line apps.
Comments
0 comments
Article is closed for comments.