MotionDSP vReveal Overview and Benchmarks
MotionDSP vReveal Overview
Main screen
Many basic consumers are using low-resolution sources like cell phones to capture video and share it with their friends and family on the internet. But, the outputted media usually has poor detail, shakiness, and noise artifacts. To help real end users make these low-resolution videos look their best, MotionDSP developed vReveal to enhance these poor-quality videos and make them a better product than what was outputting by the device.
General options menu
vReveal claims to dramatically improve the quality of problematic videos by combining advanced enhancement technology with the familiar touch-up tools found in photo editing software. This software can brighten dark videos, sharpen blurry videos, and upscale low-resolution videos. vReveal can also capture print-quality still images from enhanced videos.
Tools menu with “NVIDIA GPU On” enabled
vReveal features an adaptation of the unrivaled, CSI-style video enhancement technology behind MotionDSP’s Ikena, the high-powered forensic software used by law enforcement and intelligence agencies. This technology is called “super-resolution.” MotionDSP holds three patents related to its super-resolution algorithms. vReveal uses powerful super-resolution algorithms to analyze and combine information from multiple frames of video to produce output frames with the best possible details.
Enhance screen with basic video effects menu
This application has also been programmed to leverage the tremendous parallel processing power of CUDA-enabled GPUs and taps into their parallel cores to dramatically increase vReveal’s available computing power. Consumers will be able to enhance their videos in less time and have their CPUs available for normal everyday tasks like e-mailing and internet browsing.
Enhance menu showing “fine tuning” options
vReveal only supports AVI, MPG, ASF, and WMV formats. With Quicktime installed, it can also play MOV, MPEG-4, and 3GP files. One other important item to note is that vReveal can import HD video, but can only enhance videos with a vertical resolution of 576p or less.
Share menu with available resolutions displayed
This application seems to mainly be geared toward a niche market of consumers who use low-resolution video capturing devices and want to make them look better for their web sites or social networking web pages. I could also see people using this to enhance short video clips that can be e-mailed. The user interface is pretty well rounded, and I like how it initially searches your computer for compatible files and organizes them with topic header bars. It will be interesting to see how it utilizes CUDA with lower-resolution video because most of the other applications we are evaluating are using CUDA for HD video.
Share screen with original and enhanced comparison option enabled
This multimedia program has several options that can be configured like changing the color of the user interface, changing file associations, setting region areas, parental controls, and power settings for notebooks.
MotionDSP vReveal Benchmarks
Getting back to our transcoding benchmarks, we put vReveal against HandBrake to evaluate transcoding times and CPU usage. We also added a “clean” enhancement filter to make the system work a bit harder during transcoding. The outputted formats include 480p and 720p WMVs. We intentionally kept the resolutions lower because MotionDSP is targeting this application to people who want to improve lower-quality video from cell phones and other low-resolution video devices. We tested vReveal with CUDA enabled and disabled to monitor CPU usage differences as well as benchmark overall transcoding times for each application.
Overall transcoding times
Average CPU usage during transcoding
These CPU usage and transcoding times are very similar to Loiloscope, but since we didn’t benchmark vReveal at 1920×1080 resolution then they are actually a bit slower than the results Loiloscope scored. The CPU usage results are pretty good with CUDA enabled, but they get into the 90s with CUDA disabled. I really expected to see lower CPU usage numbers because the GPU should be able to handle alot more of the load, especially at these resolutions. Transcoding the 320×240 AVI files went extremely fast, but that was of course expected because of their file size and resolution.
The biggest takeaway from vReveal is that its really geared toward consumers wanting to use their existing low-resolution video devices, but have a way to spice them up a bit as well as improving clarity and overall video quality. This is definitely where vReveal excels and we hope they consider adding 1920 resolution support in future revisions.
