DRUM ultrafast camera: how diffraction-gated imaging threatens $100k high-speed systems

DRUM Camera Makes $100,000 High-Speed Cameras Obsolete

Estimated reading time: 10 minutes

Key Takeaways

• Researchers from the Institut National de la Recherche Scientifique (INRS) in Canada, working alongside Concordia University and Meta Platforms Inc., have built a DRUM camera — a diffraction-gated real-time ultrahigh-speed mapping system — that hits 4.8 million frames per second using parts you could order from a commercial catalog.
• That means a research-grade DRUM system could, in theory, be built for under $10,000 USD — compared to six-figure systems on the market today.
• One of the most forward-looking applications mentioned in the original research is lidar — the laser-based ranging and detection technology that sits at the heart of most autonomous vehicle sensor systems.

For decades, capturing the world at millions of frames per second was a privilege reserved for well-funded labs and deep-pocketed corporations. If you wanted to film a laser pulse splitting a water droplet or watch a cavitation bubble grow in real time, you needed a camera that cost as much as a house. That world is changing fast.

Researchers from the Institut National de la Recherche Scientifique (INRS) in Canada, working alongside Concordia University and Meta Platforms Inc., have built a DRUM camera — a diffraction-gated real-time ultrahigh-speed mapping system — that hits 4.8 million frames per second using parts you could order from a commercial catalog. The full technical findings were published in Optica in 2023, and the ripple effects are still spreading across the imaging, biomedical, and autonomous vehicle industries.

This is not a minor tweak to existing technology. This is a fundamental rethink of how high-speed imaging works, who can afford it, and what it can be used for.

What Is the DRUM Camera and How Does Ultrafast Imaging Work?

To understand why the DRUM camera matters, you first need to understand what ultrafast imaging actually is — and why it has always been so expensive.

High-speed imaging means capturing events that happen far too quickly for a normal camera to record. We are talking about things like:

• Laser pulses interacting with biological tissue
• Shock waves moving through fluids
• Plasma channels forming in a fraction of a microsecond
• Cavitation bubbles expanding after a laser fires into water

To freeze these events, a camera must open and close its shutter millions of times per second. Each opening is called a “frame.” A camera that shoots 4.8 million frames per second is opening and closing its shutter nearly five million times in the span of a single second.

Traditional ultrafast cameras — often called streak cameras or compressed ultrafast photography (CUP) platforms — achieved this using highly specialized and expensive optical components. Their price tags routinely exceeded $100,000 USD. Some of the most capable systems cost significantly more. That made them accessible only to elite research institutions and large industrial operators.

The DRUM camera changes this equation entirely. Researchers on the team described it plainly: “Our camera uses a completely new method to achieve high-speed imaging. It has an imaging speed and spatial resolution similar to commercial high-speed cameras but uses off-the-shelf components that would likely cost less than a tenth of today’s ultrafast cameras.”

That means a research-grade DRUM system could, in theory, be built for under $10,000 USD — compared to six-figure systems on the market today.

Source: Lab Manager

The Digital Micromirror Device — The Key to Affordable High-Speed Lidar and Imaging

The most important component inside the DRUM camera is something you have likely already owned without knowing it: a digital micromirror device (DMD).

DMDs are the tiny arrays of microscopic mirrors used inside projectors to create images. Texas Instruments manufactures them at commercial scale. Each mirror in a DMD array can tilt to one of two positions — typically ±12 degrees — and switch between those positions at speeds measured in the tens of kilohertz. Because they are mass-produced for the consumer electronics market, they are comparatively cheap, reliable, and widely available.

The INRS team recognized something clever: if you control the tilt angle of DMD mirrors across an array in a precise, time-varying sequence, those mirrors can act as a diffraction grating — a surface that splits incoming light into multiple replicas, each traveling in a slightly different direction.

Here is how it works step by step:

1. Light from a fast event — say, a femtosecond laser firing into water — hits the DMD array.
2. The DMD’s mirror pattern changes rapidly, creating what the team calls a “sweeping optical diffraction envelope.”
3. Each sweep of the diffraction envelope acts as a time gate, capturing light from a specific moment in time.
4. Each time-gated slice of light is sent to a different region of the camera sensor.
5. Those slices are then assembled into a sequential ultrafast movie.

The result is a single-shot movie with a sequence depth of seven frames, each separated by roughly 0.37 microseconds, meaning each complete DRUM sequence captures approximately 2.6 microseconds of real time. At 4.8 million frames per second, that is an extraordinary level of temporal detail packed into a single exposure.

One of the most significant design decisions the team made was to ensure that DRUM is compatible with standard CCD and CMOS sensors — the same sensor types already found in millions of scientific and industrial cameras worldwide. This means existing camera hardware could potentially be retrofitted with a DRUM diffraction-gating module rather than replaced entirely. That drastically lowers adoption costs for labs and companies that already have imaging infrastructure in place.

Source: Texas Instruments
Source: Lab Manager

Stop Running Your Business on Manual Work

If the DRUM camera proves anything, it is that doing more with less — smarter tools, lower costs, faster output — is how modern teams win. The same principle applies to your business operations. At Flowgrammer, we help businesses automate their workflows so they stop wasting hours on repetitive tasks and start scaling without adding headcount.

👉 See How Business Automation Works for You (https://flowgrammer.ca)

What the DRUM Camera Was Tested On — Real-World Validation

Publishing a camera specification is one thing. Proving it works in demanding, real-world experimental conditions is another. The INRS team did not simply claim performance numbers — they demonstrated DRUM on some of the most challenging ultrafast phenomena in current scientific research.

Femtosecond Laser Breakdown in Liquid

The first major test involved firing femtosecond laser pulses into distilled water and recording what happens. When an ultra-short laser pulse hits a liquid with enough intensity, it triggers laser-induced breakdown — a rapid, complex process that creates a plasma channel (a streak of ionized matter) followed almost immediately by the formation of a cavitation bubble.

The DRUM camera captured the full evolution of this plasma channel and tracked the growth of the cavitation bubble frame by frame. Crucially, the bubble radii measured in the DRUM footage matched theoretical predictions from established cavitation physics models. That is a strong validation signal — it means the camera is not just fast, it is also quantitatively accurate.

Carbonated Drinks and Bubble Dynamics

The team also used DRUM to image bubble dynamics in carbonated beverages — a useful proxy for studying microfluidics, aeration processes, and liquid behavior under rapid perturbation. While this may sound like a novelty experiment, bubble dynamics in liquids have serious implications for drug delivery, industrial chemical processing, and food science.

Laser Ablation in Biological Tissue

Perhaps the most significant test was imaging laser ablation in single-layer onion cells. Laser ablation — the process of using a focused laser to remove or alter material — is foundational to surgical techniques, including nano-surgery on biological tissue. Capturing these interactions at millions of frames per second, in a biologically relevant sample, opens direct pathways to surgical planning and real-time procedure monitoring.

Lead researcher Jinyang Liang was specific about the long-term vision: “In the long term, I believe that DRUM photography will contribute to advances in biomedicine and automation-enabling technologies such as lidar, where faster imaging would allow more accurate sensing of hazards.”

Co-author Xianglei Liu added: “DRUM photography may even be applied to nano-surgeries and laser-based cleaning applications.”

Source: EurekAlert!
Source: Lab Manager

DRUMS — The Schlieren Extension That Expands DRUM Even Further

The DRUM platform did not stop at direct imaging. Researchers have since introduced a follow-on variant called DRUMS — Diffraction-gated Real-time Ultrahigh-speed Mapping Schlieren photography.

Schlieren imaging is a technique used to visualize refractive index gradients — in simpler terms, it makes invisible differences in density or temperature within a transparent medium visible on camera. This is the same technique used to see heat shimmer above a hot road, or to visualize airflow around aircraft in wind tunnel tests.

By building schlieren contrast into the DRUM framework, researchers have unlocked new application domains:

• Aero-optics: Visualizing how high-speed airflow distorts optical paths, critical for aircraft and missile guidance systems
• Combustion diagnostics: Capturing the early stages of flame ignition and fuel-air mixing at sub-microsecond timescales
• Shockwave studies: Recording the propagation of pressure waves through gases and liquids — relevant for both defense research and industrial safety

DRUMS is a single-shot technique, meaning it captures these complex, non-repeatable events in a single firing — exactly the constraint that makes ultrafast imaging so difficult and so valuable.

Source: Researching

Why the DRUM Camera Threatens the $100,000 High-Speed Camera Market

The commercial implications of DRUM are significant and increasingly recognized across the imaging industry.

The high-speed camera market currently relies on a small number of premium manufacturers — companies that produce cameras for automotive crash testing, scientific research, defense applications, and industrial inspection. Entry-level professional high-speed systems start around $10,000 to $20,000 USD. Systems capable of millions of frames per second routinely sit at $100,000 and above.

DRUM disrupts this structure from multiple directions:

1. Commodity components replace custom optics.
By using DMDs from the projector industry and standard CCD/CMOS sensors, the DRUM architecture avoids the custom-built optical components that drive costs up in streak cameras and other ultrafast systems. Mass production economics apply from day one.

2. Retrofitting is possible.
Because DRUM is compatible with existing CCD and CMOS sensor platforms, system integrators can potentially add diffraction-gating capability to cameras labs already own. This lowers the barrier to adoption further — it is not always a full replacement purchase.

3. New markets open up.
The biggest impact of a low-cost ultrafast camera may not be in displacing existing buyers — it may be in creating entirely new markets. Researchers at universities without large imaging budgets, startups working on autonomous vehicle lidar, medical device companies developing laser-based surgical tools — these groups currently cannot afford ultrafast imaging. DRUM could change that.

As one technology commentator noted, DRUM’s architecture could “erode the premium segment of high-speed camera markets — where $100,000-plus systems remain common — while opening new demand in autonomous driving, robotics, and nano-surgery.”

Source: Fanatical Futurist

High-Speed Lidar, Autonomous Vehicles, and the Bigger Picture

One of the most forward-looking applications mentioned in the original research is lidar — the laser-based ranging and detection technology that sits at the heart of most autonomous vehicle sensor systems.

Lidar works by firing rapid laser pulses and measuring how long they take to bounce back from objects. The faster the system can fire, detect, and process those pulses, the more accurately it can map a three-dimensional environment in real time. Current lidar systems face trade-offs between range, resolution, and cost.

An affordable, ultra-high-speed imaging platform based on DRUM principles could allow lidar sensors to:

• Detect fast-moving hazards — a child running into the road, debris falling from a truck — with greater temporal precision
• Operate at higher frame rates without requiring expensive custom components
• Be integrated into more compact and cost-effective sensor packages

This puts DRUM-based technology directly in the path of one of the fastest-growing hardware sectors in the automotive and robotics industries. Government-backed programs for autonomous vehicle development and biomedical imaging diagnostics — both of which depend on high-speed optical sensing — are natural early adopters.

The fact that DRUM is explicitly built to scale using DMD hardware that already ships by the millions each year means manufacturing risk is low. The same industry infrastructure that produces projectors for classrooms and conference rooms can, in principle, supply the core components for the next generation of lidar sensors and ultrafast diagnostic cameras.

Source: EurekAlert!
Source: Fanatical Futurist

What Comes Next for DRUM Technology

The INRS team has been transparent about both the current limitations and the near-term roadmap for DRUM.

Current limitations:

• The sequence depth is seven frames per single-shot exposure. For many applications, this is sufficient — but some complex, longer-duration events would benefit from deeper frame sequences.
• The system currently captures images in grayscale. Color imaging has not yet been demonstrated in the DRUM architecture.
• Temporal resolution between frames sits at approximately 0.37 microseconds, meaning very fast events that require sub-100-nanosecond discrimination are still better served by streak cameras in some contexts.

Planned developments:

• Increasing both imaging speed and sequence depth beyond current benchmarks
• Developing color imaging capability within the DRUM framework
• Expanding lidar integration for autonomous sensing systems
• Applying DRUM to nano-surgery research and laser-based industrial cleaning

The DRUMS schlieren variant is already in active development, and the broader DMD ecosystem — with kilohertz-scale switching speeds and proven MEMS reliability — provides a solid foundation for commercial scaling.

From a research infrastructure standpoint, the combination of commodity components, sensor compatibility, and a growing library of validated use cases positions DRUM as one of the more credible paths to mainstream ultrafast imaging adoption seen in the last decade.

Source: Lab Manager
Source: EurekAlert!

Frequently Asked Questions (FAQ)