Novastar H15 In-Depth Review: Specs, Real-World Performance, and Setup

The H15 is a 15U rack mount video wall splicer and LED sending controller built into a single modular chassis. Older NovaStar setups chained a separate video processor (a VX or J6) to a sending card in some external box. The H-Series stuffs everything into one enclosure: input cards, output cards, processing engine, control interface.

It’s built for fine pitch LED. We’re talking sub-2mm panels in control rooms, broadcast environments, and corporate installs where image quality actually gets scrutinized. The H15 handles 4K@60Hz natively across its pipeline, with 10-bit color depth and HDR10/HLG support that isn’t an afterthought bolted on through some external converter.

The modular card system is the real headline. Up to 30 input card slots and 10 to 16 output card slots depending on whether you spring for the Enhanced variant. You spec the chassis once and configure it per project: SDI for broadcast, HDMI 2.0 and DP 1.2 for presentation inputs, IP decoder cards for surveillance feeds. You get the idea.

AV pros spec’ing a processor need to know whether it can drive the wall the CAD drawing promised. Here are the numbers.

The H15 is not small. At 482.6 × 683.0 × 533.0 mm and 41 kg net (closer to 62 kg fully loaded with cards), it’s a two-person lift and then some. You’re not throwing this in a shallow rack. It needs full depth mounting and proper ventilation. Power draw sits around 900 W with dual supplies standard, and you can add two more backup PSUs if uptime is non-negotiable.

The modular approach pays for itself on the input side. NovaStar offers something like 15 different input card types: quad DVI, quad HDMI (1.3 and 2.0 variants), HDMI 2.0 + DP 1.2 combos, 12G-SDI, quad 3G-SDI, quad VGA, dual CVBS + VGA, HDBaseT, IP decoder cards, even audio I/O cards if you need embedded audio routing. Pick the cards that match your signal environment and hot swap them without powering down the chassis.

One thing worth knowing before you plan your card layout: the HDMI 2.0 + DP 1.2 cards are limited to slots I-1 through I-8. Each card handles 4× 2K×1K, 2× 4K×1K, or 1× 4K×2K at 60 Hz.

The standard Novastar H15 takes up to 10 output cards, supporting 40 output channels max. The Enhanced version bumps that to 16 cards and 64 channels. That’s meaningful if you’re driving a very large or unusually shaped wall.

Pixel loading depends entirely on which output card you choose:

• The H_20×RJ45 card handles roughly 13 million pixels per card
• The H_4×Fiber card pushes about 20.8 million pixels per card on the standard Novastar H15, or 26 million on the Enhanced
• Max out a standard H15 with fiber cards and you’re looking at roughly 208 million pixels total

That’s more than enough for most 4K and even 8K class LED installations. The Enhanced can stretch past 400 million with the right fiber cards, though I’ve rarely seen anyone actually need that.

Other processors make you power down to swap a failed card. The Novastar H15 doesn’t. In a 24/7 control room or during a live broadcast, that matters. Pull the bad card, slot the replacement, the system keeps running. The chassis doesn’t even ship with cards. Everything is ordered to spec, which means you only pay for the I/O you actually need.

The Novastar H15 handles 4K×2K at 60 Hz across its pipeline with 4:4:4 chroma sampling and 10-bit color. With the HDMI 2.1 + DP 1.4 input card, it stretches to 8K. HDR10 and HLG are processed natively. No external HDR converter sitting in the signal chain adding latency and another potential failure point.

You can stack, scale, crop, and position up to 160 independent layers across your output canvas. Each output card handles its own layer budget (16 layers per card on the standard H15, or 10 per card on the Enhanced), so the compositing load distributes across the chassis. For a multi screen installation with PIP windows, lower thirds, and live feeds coming from half a dozen sources, this is table stakes.

NovaStar moved to a browser/server architecture for the H-Series. No proprietary software install required. You connect to the H15’s IP address from any device with a browser and the full control interface loads. Works on Windows, Mac, iOS, Android, Linux.

In practice, this is great on a laptop or desktop. Mobile screens get cramped fast when you’re managing 160 layers. For quick preset recalls or status checks on a phone it’s fine. For actual configuration work, use a real screen.

The 10-bit pipeline is noticeable when you feed it good content. Gradients render smooth without banding. HDR content pops without looking overcooked. Latency sits at roughly one frame in the standard processing pipeline, which is fast enough for live IMAG use in most corporate and broadcast scenarios. If you’re doing esports or anything needing sub frame latency, test it in your specific signal chain. For 95% of applications, nobody will notice or care.

One thing that catches people: the H15 processes at the bit depth you tell it to. If your source is 8-bit and you leave the processor at 10-bit, you might see banding or artifacts that aren’t actually the processor’s fault. Match your bit depth settings to your source.

Dual PSUs come standard. You can add two more. There’s device-to-device backup, input card failover, and optical port redundancy. For a control room or broadcast facility where a black screen means someone important gets very angry, this is the kind of thing you spec the H15 for in the first place.

If you’ve used NovaLCT before, the transition to the H-Series web interface is mostly intuitive. A few things moved around, but the logic is the same. If you’re coming from a different ecosystem (Brompton, Colorlight), expect a couple of days of genuine frustration.

NovaStar’s documentation exists, but it’s not always easy to find the right version for your specific firmware. YouTube tutorials from actual technicians tend to be more useful than the official manuals. NovaStar could improve this.

The H15 chassis alone runs roughly $5,500 to $5,800 USD. The Enhanced variant sits around $6,200 to $6,500. Neither price includes a single input or output card. I want to make sure that’s clear because the first quote always surprises people.

Cards range from a few hundred dollars for a quad VGA or CVBS card up to over a thousand for an HDMI 2.0 + DP 1.2 or 12G-SDI card. A fully loaded H15 with a reasonable mix of modern digital inputs and fiber output cards can push past $15,000 to $18,000 without much effort.

Is that expensive? Compared to a standalone Novastar VX1000 or similar all in one processor, yes. Compared to a Brompton Tessera SX40 with comparable I/O density, it’s actually competitive. In some configurations, notably cheaper.

If you don’t need 160 layers and 208 million pixels, the H9 or H5 is probably the smarter spend. The H9 is a 9U chassis with similar card compatibility but fewer slots. The H5 is a 5U unit aimed at smaller fixed installations. The H15 earns its premium when you genuinely need the I/O density and the pixel headroom. Buying an H5 and running out of card slots six months later is false economy.

Brompton is the benchmark for high end LED processing, particularly in film, broadcast, and virtual production. The Tessera SX40 and S8 processors are incredibly capable. Brompton’s Dynamic Calibration and HDR implementation are widely regarded as best in class. If you work on high end film sets, Brompton is often specified by name and you don’t really have a choice.

Where the H15 competes is on I/O density per dollar and the all in one form factor. Brompton setups typically separate processing, scaling, and sending into multiple boxes. That’s more flexible in some ways, but it adds rack space, cabling, and points of failure. The H15 puts everything in one chassis. For a rental house or integrator who values simplicity and rack density, that matters.

The trade off is ecosystem lock in. Brompton’s Tessera software and calibration tools are objectively more polished than NovaStar’s equivalents. But in corporate AV, control rooms, and live events outside the absolute top tier, the H15 holds its own just fine.

Colorlight and Linsn are the budget alternatives. Colorlight’s Z6 and X4e processors can handle similar resolutions on paper at a much lower price point. The software experience, documentation quality, and long term firmware support don’t match NovaStar’s level. Linsn is similar: viable for price sensitive installations where someone on staff already knows the Linsn ecosystem inside out, but not what you’d spec for a mission critical broadcast facility.

If you’re an integrator serving corporate clients who will call you the moment something flickers, the H15’s reliability and support infrastructure justify the price gap over Colorlight or Linsn. You’re paying for fewer 2 a.m. phone calls.

Once the H15 is racked and powered, connect a laptop to the management Ethernet port and navigate to the unit’s IP address in a browser. The web UI loads. No installer, no plugin, no license key.

Go to Configuration → Screen Configuration → New Screen. Name the screen something sensible. When you’re managing presets across multiple screens, “Screen 1” does not cut it. Set the rows and columns to match your physical cabinet layout, then drag the output card type onto the canvas. The interface shows you which ports map to which sections of the wall.

If there’s one step where installations go sideways, it’s this one. You must load the correct `.rcfgx` file for your specific LED cabinets. This file defines the scan mode, driver IC type, pixel order, and data grouping. Every LED manufacturer provides these. Never use a default template unless you enjoy troubleshooting at midnight.

In NovaLCT (or the web interface, depending on firmware version), load the RCFGX file to each receiving card on the wall. Then verify the cabinet resolution matches the physical panels. If your cabinets are 192×192 but the config thinks they’re 128×128, your image will be scrambled and nothing else you fix will matter. Took me a very long night to figure that out the first time.

1. Verify RCFGX file loaded and cabinet resolution confirmed
2. Check Ethernet port loading: keep under ~650,000 pixels per port
3. Match processor bit depth to your source signal
4. Set output resolution to match the LED wall’s actual pixel dimensions
5. Test at 60 Hz before pushing higher refresh rates
6. Disable advanced image processing features (gray compensation, temperature compensation) and enable them one at a time, testing each

The H15 isn’t the best at any one thing. Brompton’s calibration is better. Barco’s UI is more polished. Colorlight costs way less.

But none of those do everything in one box the way the H15 does. Hot swap cards. Redundant power. Web control that runs on any device. Enough pixel headroom to survive your client’s next wall upgrade. You configure it, you rack it, you mostly forget about it.

If your budget stretches past $10,000 for a configured unit and you need zero-downtime reliability, put the H15 on your shortlist. Just budget for the cards, not just the chassis, and triple-check your RCFGX files before you hit send. You’ll save yourself a long night.