NEW PODCAST: Can Anyone Shoot an IMAX Film? From Production to the Big Screen – Focus Check ep142 →🎙️ WATCH/LISTEN Now
Watch/Listen NowCan Anyone Shoot an IMAX Film? 🎙️NEW PODCAST
Education for Filmmakers
Language
The CineD Channels
Info
New to CineD?
You are logged in as
We will send you notifications in your browser, every time a new article is published in this category.
You can change which notifications you are subscribed to in your notification settings.
Aiarty has updated its Aiarty Video Enhancer desktop software to version 3.9. The new version lets users pause an export and check the frames that are already processed, and it follows version 3.8, which added recovery of unfinished exports and MKV support. The application runs offline on Windows and Mac and combines AI denoising, deblurring, upscaling to 4K, frame interpolation, audio cleanup, and SDR to HDR conversion in one interface. A Black Friday offer currently reduces the lifetime license by 45%.
We have previously reported on the Aiarty Video Enhancer at its introduction and after an update with new AI models and control options, and we also covered the Aiarty Image Enhancer, which uses a similar approach for still images. Aiarty is a division of Digiarty Software. The company positions the Video Enhancer as a separate enhancement stage that sits next to the editing application, not as a replacement for it: footage is cut in the NLE, and clips that need more resolution, less noise, or smoother motion go through the Video Enhancer before final delivery. As a local video upscaler, Aiarty keeps footage on the user’s computer, helping protect client privacy and avoiding transfer delays. Its lifetime license includes free lifetime updates and commercial use, with no recurring subscription or processing-credit fees, potentially lowering ongoing software costs for frequent users
The software uses three AI models, and the user selects one per clip. moDetail-HQ v3 is a diffusion and GAN model for fine texture such as hair, skin, foliage, and fabric. Smooth-HQ v3 is a diffusion model for faithful restoration, with gentler deblurring and denoising and a focus on color fidelity. superVideo vHQ, also a diffusion and GAN hybrid, is for very noisy, low-light, and dark interior footage.
moDetail-HQ v3 and Smooth-HQ v3 can upscale by 1x, 2x, or 4x, or to fixed 1K, 2K, and 4K targets. superVideo vHQ works at 2x only. A Strength Control slider sets how strongly each model processes the footage, so users can choose between a subtle, natural result and stronger sharpening. According to Aiarty, the models were trained on 2.38 million images and 120 hours of HD and 4K video.
The Denoise Video function targets luminance grain, chroma noise and high ISO artifacts. Aiarty names typical use cases such as concerts in dim venues, indoor scenes and night shots. The same models also handle softness from slight focus errors, compression blocks and pixelation from heavy digital zoom or cropping.
With Upscale Video, sources at 360p, 480p or 540p can be upscaled to HD, 2K or 4K. For productions, this is relevant when cropped shots, archive material, or older camera footage must match 4K material on the same timeline. Because the models are generative, the added detail is reconstructed by the AI and does not come from the source file.
The frame interpolation module generates new frames to bring footage to 60 or 120fps, or to create 8x and 16x slow motion. Aiarty says that it uses AI-trained flow estimation and motion compensation to avoid warping, smearing, and the jello effect. The company lists live action with fast motion, time-lapse, animation, and 3D renders as suitable material.
An Audio Denoise module removes background noise, clicks, echoes, and wind noise and makes voices clearer, with a real-time preview. A color panel adjusts white balance, exposure, contrast, saturation, highlights and shadows. The experimental HDR function converts SDR 8-bit footage to HDR 10-bit in three modes (Neutral, Bright and Graded), with its own strength slider. Basic edit tools for trimming, cropping, rotating, flipping, and forced deinterlacing complete the feature set.
Version 3.9 adds pause and resume for export tasks. When the user pauses an export, the software writes a playable clip of the part that is already processed, and a “View Clip” option opens it. If an export is paused after ten seconds, for example, the clip holds those ten processed seconds. Together with the trim function, this gives a quick way to compare enhancement settings before a full-length export. The update also adds a “Start All Tasks” button to the task queue and support for copying image-based subtitles in MKV files.
Version 3.8 focused on reliability for long clips and batches. Unfinished export tasks can now be recovered and resumed after a restart of the software, instead of starting again from the beginning. Aiarty also optimized CUDA encoding memory management to reduce VRAM usage during demanding exports, added MKV import, improved decoding and export logic for some formats, and added Spanish and Italian interfaces.
The Video Enhancer imports a wide range of formats, including MP4, MOV, MXF, MTS, M2TS, MKV, AVI and WebM. Output is MP4, MOV, or MKV with H.264, H.265, ProRes, or AV1. Depending on the codec and profile, the export settings offer YUV 4:2:0, 4:2:2, or 4:4:4.
On Windows 10 or later, the software uses DirectML on AMD, Intel, and NVIDIA GPUs and TensorRT on NVIDIA cards from the GeForce 10 series to the RTX 50 series. Aiarty asks for at least 8GB of RAM (16GB recommended) and 4GB of VRAM on dedicated GPUs (8GB recommended). On macOS 13 or later, it supports Intel Macs with a dedicated GPU and Apple Silicon, with CoreML acceleration.
Turbo Mode prioritizes processing speed. According to Aiarty, upscaling a 25-minute 2K clip at 30fps to 4K takes 25 minutes with Turbo Mode on a Windows PC with an NVIDIA RTX 5090. Step Mode uses tiled AI inference to lower VRAM usage on long, high-resolution clips, at the cost of speed. Batch export processes many clips in one session, and all processing stays on the local computer, with no upload to a cloud service.
The free version processes clips of up to 120 seconds each and adds a watermark; batch export requires a full license. superVideo vHQ, the model for the most difficult low-light footage, is limited to 2x upscaling. The HDR conversion is marked as experimental and needs an HDR-capable display for correct viewing. The speed figures above are Aiarty’s own, and we have not tested them.
Aiarty Video Enhancer is available now for Windows and Mac, and a free trial can be downloaded from the Aiarty website. With the current Black Friday offer, a lifetime license for three Windows PCs or three Macs costs $129 instead of $235, with free lifetime updates. A one-year license for one PC or one Mac costs $79 instead of $99. The Media Enhancer Bundle with the Aiarty Image Enhancer is $199 instead of $390, and the 3-in-1 bundle, which adds Aiarty Image Matting, is $259 instead of $499. All prices are in US dollars, and Aiarty gives a 30-day money-back guarantee.
For more information, please visit the Aiarty website.
Do you use AI enhancement as a separate stage after the edit, or do you keep noise reduction and upscaling inside your NLE? Don’t hesitate to let us know in the comments below!
Δ
Stay current with regular CineD updates about news, reviews, how-to’s and more.
You can unsubscribe at any time via an unsubscribe link included in every newsletter. For further details, see our Privacy Policy
Want regular CineD updates about news, reviews, how-to’s and more?Sign up to our newsletter and we will give you just that.
You can unsubscribe at any time via an unsubscribe link included in every newsletter. The data provided and the newsletter opening statistics will be stored on a personal data basis until you unsubscribe. For further details, see our Privacy Policy
Johnnie Behiri is a documentary filmmaker with more than three decades behind the camera: he learned the craft in daily television news gathering, filmed for Japan's public broadcaster NHK, and spent eight years shooting and editing for BBC News. He joined cinema5D in 2011 and has co-owned and co-run CineD and MZed ever since. Johnnie writes many of CineD's camera reviews, co-hosts our weekly Focus Check podcast, and heads CineD's manufacturer relationships and product development.