• VP Land
  • Posts
  • Puget Systems' Matt Bach on NVIDIA's New GPUs and Practical AI Applications

Puget Systems' Matt Bach on NVIDIA's New GPUs and Practical AI Applications

NVIDIA's new GPU lineup is expanding hardware support for video formats and bringing significant VRAM increases that enable more practical AI applications in media workflows.

Matt Bach of Puget Systems shared insights on how these new graphics cards are helping professionals work with a broader range of content types and enhance their production capabilities.

Hardware Decoding Breakthrough: New NVIDIA GPUs restore support for legacy formats professionals still rely on daily

One of the most significant but understated improvements in NVIDIA's latest GPUs is expanded hardware decoding support, particularly for 4:2:2 10-bit in both HEVC and H.264 formats.

  • The return of H.264 hardware decoding support is particularly valuable for professionals working with archival footage or content from older cameras

  • Broader codec support eliminates the time-consuming need to transcode or create proxies for incompatible footage

  • The performance gains from hardware acceleration are substantial, especially when dealing with multiple video sources

Memory Matters: Increased VRAM capacity opens new possibilities for AI experimentation and high-resolution workflows

NVIDIA's RTX Pro lineup will top out at an impressive 96GB of VRAM, bringing capability that was previously limited to expensive server-class hardware.

  • The expanded memory allows professionals to work more easily with 8K footage, virtual production assets, and AI models

  • For AI image generation, more VRAM enables multiple simultaneous outputs, facilitating a more efficient creative selection process

  • The upcoming RTX Pro cards make on-premises AI training more accessible, with a typical ROI of just six months compared to cloud services

Beyond the Hype Cycle: AI tools finding their place in practical production workflows rather than replacing them

The industry is moving past both the over-enthusiasm and fear surrounding AI to find practical applications that enhance existing workflows.

  • Metadata tagging, localization, and automated dubbing with lip sync are emerging as valuable AI applications rather than full content generation

  • Adobe's Media Intelligence for content searching and automatic transcription features represent practical integration of AI

  • DaVinci Resolve 20 release includes AI-powered script sync and audio track splitting that address specific production pain points

The Production Reality: The most valuable technology developments are the ones that integrate seamlessly into existing workflows

As the industry moves beyond the initial excitement of technologies like virtual production and generative AI, professionals are identifying which elements actually improve their day-to-day work.

  • The conversation is shifting back to core production needs rather than speculative applications

  • Cloud-based AI users are increasingly moving to on-premises solutions for both cost and data privacy reasons

  • The most successful AI implementations address specific workflow challenges rather than attempting to replace creative decision-making

Reply

or to participate.