Gigapixel AI for Video Help Document
If using via command line (videosrcmd.exe), please refer to the usage text for parameter information.
For non-command line usage, Gigapixel AI for Video is registered with many common video formats by default. File types supported include avi, flv, m1v, mkv, mov, mp4, mpg, mpeg, webm, and wmv.
To use the program, follow these steps for your video file.
1. Navigate to your file
2. Right click your file and find the Gigapixel AI for Video entry that is made after install.
3. Use the context menu to select your operation. It may take a few seconds for it to open.
Gigapixel AI for Video Command Line Mode Guide
1. Please open a command line terminal. You can type "cmd" in your run window to find it.
2. Run videosrcmd.exe "inputVideoPath" with the following parameters:
Print this help
Set output video path. Must have extension ".mp4" or ".png"or ".tif" or ".tiff"
"inputvideo_Nx.mp4" in the same folder of input video where N for upscale factor (1, 2, or 4)
Output 16 bit png sequence. Only effect for png output
8 bit pngs
Set upscale factor as 100%, 200%, or 400%
Explicitly set output video width
200% of input video width
Explicitly set output video height
200% of input video height
Explicitly set output video FPS
Input video FPS or 30 if input as image sequence
Crop output video if needed, otherwise fit video inside output size
Set processing start point (in second)
Set processing end point (in second)
End of video
0,1,2,3: explicitly uses nvidia GPU device #0,1,2,3 (same index as nvidia-smi)
10: use CPU/iGPU.
-vram 10-90 (in percentage)
Set available VRAM manually. Will use only a fraction of the VRAM (10% - 90%)
-l username password
Login with username(email address) and password
No preview display windows
1). videosrcmd.exe "C:\test.mp4"
Will process the video under path "C:\test.mp4" and save the output as a .mp4 file into "C:\test_2x.mp4". The upscale factor will be 200%. Whole video will be processed. Device will be automatically chosen.
2). videosrcmd.exe "C:\test2.mp4" -o "C:\test_output_pngs\test2_outputs.png" -s 4 -b 3 -e 6 -nvidia 1 -vram 70
Will process the video under path "C:\test2.mp4" and save the output frames as pngs under the folder
"C:\test_output_pngs\test2_outputs". If there is no such folder, the software will create one. The upscale factor will be 400%. Video will be processed from the 3rd second to the 6th second. The software will run by Nvidia GPU. If your device has two GPUs, please make sure to set the gpu number to the correct one (same index as nvidia-smi). In this example, Nvidia GPU device number is 1. The VRAM will be used is 70% of the total available VRAM.
3). videosrcmd.exe "C:\test3.mp4" -o "C:\test_output_pngs\test3_outputs.png" -16bit -w 1920 -h 1060 - c -nvidia 10 -nd
Will process the video under path "C:\test3.mp4" and save the output frames as pngs under the folder
"C:\test_output_pngs\test3_outputs". If there is no such folder, the software will create one. Output pngs will be 16bit with 1920 pixels in width and 1060 pixels in height. If the output aspect ratio is different from the input, it will be cropped. The upscale factor will be 200%. Whole video will be processed. The software will run by iGPU/CPU. No preview display windows will be shown.
4). videosrcmd.exe "C:\test_image_sequence\image_%06d.png" -fps 60
Will process the images under path "C:\test_image_sequence\" and save the output as a .mp4 file under the same path. Images should be in png format, and their names should be like "image_000000.png", "image_000001.png", "image_000002.png", and so on. Output video will be in 60 FPS.
3. You are all set! Please wait for your upsampled video and enjoy it!
Gigapixel AI for Video FAQs
1. What is the minimum hardware requirement for Gigapixel AI for Video?
We suggest you to run Gigapixel AI for Video on high end Nvidia graphic cards. The minimum requirement is GTX 1050 with 4 GB VRAM. Any graphic card later than GTX 1080 (included) will do a fairly fast job.
If you do not have a Nvidia graphic card in your device, the software can run on your Intel iGPU. Typically, it will be 5 times slower. If you do not have an Intel iGPU, the software will finally fall back to you CPU, which can be 10 times slower than the Nvidia GPU version.
2. Can Gigapixel AI for Video work on Macs?
For now, Gigapixel AI for Video cannot work on Macs. It can only work on Windows devices.
3. Can I do batch processing on several videos?
Sure you can! Please check the "Gigapixel AI for Video Command Line Mode Guide" to make sure how to write command line instructions. You can write all the instructions in separated lines into a .bat file and then run it.
There is an example .bat file in your video folder after installation. You can check and run the "test.bat" there. It will help you to generate 2 sample upscaled videos.
4. My machine cannot meet the minimum requirement. Is there a way that I can use this product?
We have a cloud service that can help you to upscale your videos online. Please check https://videoai.topazlabs.com/ for more details!
5. How long it will take to process a video?
We tested mainly on Nvidia GTX 1080. For a video from HD to 8K, it typically takes 4-5 seconds for each frame. If you upscale a video from SD to HD, it will take about 0.5 second for each frame.
6. What is the maximum size of an output video I can get?
Now we support up to 8K video output. Any size larger than 8K will cause a failure.
7. Can I process interlaced video?
Not directly. If your video is interlaced, please deinterlace it first before using Gigapixel AI for Video to upscaling it. We are working on a deinterlacer AI at this moment, stay tuned.
8. My machine has more than one Nvidia GPU, can I use all of them?
Yes and No. The program itself only use one GPU. However, you can open multiple command consoles and use videosrcmd.exe with -gpu 0 or -gpu 1 to run each GPU on a separate video at the same time.
9. Can I read in a sequence of image as input?
Yes but you have to set it up via command line. Please refer to example 4) in “Gigapixel AI for Video Command Line Mode Guide”.
10. My GPU utilization seems low in task manager. Is there something wrong?
In fact it is normal. The GPU utilization in task manager is not precise. If you want to track the actual GPU usage, please try nvidia-smi.