Apple hardware recommendations for Reeltime: 3 setups under the mircoscope

With the recent release of Reeltime Lite, we introduced a more compact (and more affordable) version of our full-featured video playback software, Reeltime Pro. Our goal: to make professional video assist accessible for productions of any size, scale, or budget. But while choosing the right software is a key part of building your setup, it’s only one piece of the puzzle. Your Mac (and the Apple silicon chip it runs on) plays a significant role in determining what your system can actually handle. At this point, you might be asking yourself: Is my current hardware powerful enough for my next video assist job? Should I consider an upgrade? And if so, which Mac is the right choice? 

To help you make an informed decision, we headed into the Pomfort test lab to assess how Reeltime Lite and Pro perform across a range of Apple hardware configurations. In this article, we’ll give you a behind-the-scenes look at how we actually test the performance of our own applications – and, more importantly, provide a clear overview of what’s possible with Reeltime in three setups of increasing complexity: We’ll cover a basic setup for essential video assist needs, an advanced setup for your average production, and finally, a maxed-out setup that pushes Reeltime to its full potential.

A few notes on Reeltime and Apple silicon

The Reeltime project began shortly after Apple launched its first self-developed M1 chip, which introduced groundbreaking possibilities thanks to its advanced architecture and integrated media engine. Notably, the unified-memory architecture, powerful on-board GPU performance, and ProRes hardware acceleration significantly boost the performance of our video pipeline. With these new opportunities, our development goal was clear: To create a native Apple silicon application that fully harnesses the power of the new processors, delivering the cutting-edge performance that today’s video assist workflows demand. 

Thanks to a video processing pipeline optimized for the latest Apple hardware, users can now apply multiple video filters without introducing any additional signal delay. Reeltime is designed with the fast-paced environment of film sets in mind, with a lightning-fast, intuitive interface that remains smooth and responsive – even in complex multi-cam setups. Furthermore, the application relaunches where you left it within seconds.

So, while Reeltime was specifically designed to leverage Apple’s latest processors, the question remains: Which piece of Apple hardware is best-suited for your video assist needs? Let’s find out!

Test setups

Testing every possible hardware configuration would be excessive. To keep things manageable (while still providing meaningful insights for productions of varying size and complexity), we defined three representative test setups: Basic, Advanced, and Max. Each setup reflects a typical production scenario and is paired with either Reeltime Lite or Reeltime Pro, depending on the assumed feature requirements:

Please note: These recommendations focus on HD signals. As soon as Reeltime’s 4K support moves out of beta, we will add 4K setup recommendations.

What is performance?

In general, we tested a range of hardware configurations for each of the setups above to identify the most cost-effective options that still deliver the desired performance with Reeltime. However, the term performance can be broad and ambiguous, especially in the context of software. That’s why, before diving into the testing details, we want to clarify what we actually mean by “Reeltime performance” in this context:

Reliability, video latency, and process parallelizability are three important key factors for good video playback performance. The most obvious point is Reeltime’s ability to process an incoming video stream fast enough so that no frames are dropped, and all frames are delivered in time to the SDI output. It gets more interesting when multiple streams are being processed simultaneously. Then, it’s necessary to parallelize these processes to spread the load across all GPU cores. The goal here is to utilize the entire chip. Especially for video assist software, it is important that the performance is reliable over a long period of time and is not affected by other processes. It’s also important that the video pipeline doesn’t add latency to the signal, even when multiple filters are added to the image. Reeltime’s UI response time is also critical to the user. Does the application start up quickly, does the user interface respond immediately to any user input? These issues determine whether the software feels “fast” to the user. Only when these three conditions are met we can speak of good performance. 

How do we test performance?

Now that we’re on the same page about what performance means in this context, let’s take a look at how we actually test Reeltime’s performance. At the core of our process is the Pomfort test lab, where we try to replicate as many different workflows as possible. The focus is on testing all the features with real hardware and real camera data to better understand how the software will behave on a real set. For our three setups, we primarily used hardware that our users work with, and that fits the budget of each setup. We tested the following combinations:

a) Basic setup:

• Software: Reeltime Lite
• Macs: Mac mini M4, MacBook Pro M2
• Video I/O device: AJA IoX3 (2x inputs, 2x outputs)

b) Advanced setup:

• Software: Reeltime Pro
• Macs: MacBook Pro M2
• Video I/O devices: AJA IoX3 (3x inputs), Blackmagic DeckLink 8K Pro (3x outputs)

c) Max setup:

• Software: Reeltime Pro
• Macs: Mac Studio M2 Ultra
• Video I/O device: AJA Kona 5 (2x inputs, 2x outputs), 2x Blackmagic DeckLink 8K Pro (4x inputs, 4x outputs)

all setups:

• For each setup, the camera signal came from an ARRI ALEXA 35, and we used Flanders DM240s and XMP310s as reference monitors. Signals were routed through an AJA Kumo 1616 12G router.

Once we had built the setups in our test lab, we began evaluating their performance based on our previously defined criteria for what “performance” means in the context of Reeltime.

Testing reliability

The most crucial aspect of Reeltime is the video playback reliability. That’s why it is essential to ensure that every single frame of the incoming camera signal is either recorded or displayed on the reference monitor without loss. To achieve this, we developed several tools to detect frame drops and pinpoint exactly where they occur within the video pipeline. But we didn’t stop there: We went a step further by capturing the outgoing SDI signal using test footage that shifts a detection line every frame. With the help of our drop detector tool, we are able to count every frame leaving the SDI output. This testing method allows us to identify every single frame drop and determine precisely when it occurs. This approach is used to test all our recommendations.

Testing latency

Another important part of the video pipeline is the signal latency. This is the time it takes for the signal to travel from the camera’s sensor to the output device, such as the reference monitor. Keep in mind that different cameras themselves add latency to the outgoing SDI signal, and different video capture devices add delay to the signal. If quick response time is important, check whether your monitor has a fast-mode feature like the Flanders DM240. Given the significance and complexity of latency, we will cover this topic in a dedicated blog post soon.

Testing endurance

Days on set are long, so Reeltime has to be reliable and stable for a full day’s shooting. Hence, each setup is tested for several hours with a mix of idle video playback and heavy use, including recording, filter settings, and slot changes.

Testing responsiveness

As previously mentioned, ensuring a smooth user experience when interacting with the UI is just as important. When pushing your machine to its limits, it becomes crucial to manage how the chip’s performance is utilized. To address this, we implement options for reducing load where it is not crucial. An option in the setting is to reduce the slot preview frame rate without affecting the output signal to SDI. This setting is especially useful when working with multiple slots, as it keeps the UI responsive even under heavy system load.

Disclaimer: Most of the time, Reeltime is used with a variety of peripherals, such as video capture devices. We work hard to provide seamless integration with external hardware, but it’s always important to test your setup regularly. Firmware or macOS updates or even just changing devices can cause problems. Our goal is to fix as many of these issues as possible as soon as we notice them. But we can’t recommend a complete device setup configuration that we haven’t tested ourselves.

Practical tips for users

During our testing, we paid close attention to a few key factors that can also impact system performance. Here are some practical considerations to keep in mind when building your own setup:

• Avoid bottlenecks: Video assist setups are a combination of many components and unfortunately one part can affect the whole performance. External drives or Thunderbolt enclosures can slow down overall performance. Make sure all components (including cables) can handle the data rates being generated.
  
• Invest where it matters: Reeltime benefits above all from the GPU cores and the media engines of the Apple silicon chips. In contrast, more CPU cores or more unified memory do not necessarily lead to better performance.
  
• Optimize your Mac: If you want to use all of your Mac’s resources for Reeltime, make sure that no other applications are running in parallel if possible. Check the macOS Activity Monitor to see if any unnecessary tasks are using computing resources. Additionally, Activity Monitor is helpful for checking whether your Mac’s GPU might be coming to its limits (Tip: cmd+4 opens the GPU History).
  
• Use what you need: During a production, it’s easy for many library slots, composite slots, and filters to accumulate. To optimize performance, make sure to keep only the slots that you actually use. Also, a small preview UI uses significant system resources.
  
• 6G and 12G video I/O cards: As mentioned above, 4K support is still in beta. However, if you want to work with 6G or 12G signals and use an appropriate 12G-capable video capture card (e.g., AJA Kona 5, Blackmagic DeckLink 8K Pro) in a Thunderbolt enclosure box, keep in mind that current enclosure boxes with Thunderbolt 3 or Thunderbolt 4 ports transfer about 2800 MB/s. Four 6G signals are right at or above that limit. For large multi-cam setups with 6G and 12G signals, we recommend waiting for Thunderbolt expansion systems with Thunderbolt 5 ports, which are expected to be available later this year.
  
• Streaming performance: The streaming performance of local and internet streaming primarily depends on the network’s speed and robustness. Check the data rate, latency, and stability of your network to ensure optimal performance. In addition, you can inspect the connection quality per (local streaming) client in the slot manager and modify the target bit rate of a stream if needed.
  

Test results & hardware recommendations

Alright, without further ado, let’s dive into our test results and the hardware recommendations for each setup! After conducting extensive testing, we’ve created an easy-to-read table outlining which Apple silicon chips are best-suited with each setup.

a) Basic setup

This setup utilizes all features of Reeltime Lite and is the perfect fit for on-location or smaller productions. With the Mac mini M4, you are able to run this setup with the cheapest Mac currently available. For mobile setups, the MacBook Air and Pro M4 are also viable options.

b) Advanced setup

The advanced setup uses the full feature set of Reeltime Pro. Even with demanding green-screen compositing and multiple filters in use, you can get this setup up and running on a MacBook Pro with an M2 Pro chip (or better).

c) Max setup

Now if you ask yourself: “What Mac do I need to use all 6 input and 9 output slots of Reeltime Pro and additionally use complex compositing settings?” The short answer is a Mac Studio M2 Ultra (or better).

A few further notes on the results table:

Recommended Apple silicon: Apple differentiates its M chips via generation and size. The different sizes are M, M Pro, M Max, and M Ultra. In 2024, the current M4 generation was released. Not all chip sizes are launched at the same time. In the past, the Ultra size was the last launched chip of each generation. So currently, the most performative chips are the M3 Ultra and the M4 Max. To prevent confusion, this means that when we recommend for the Basic setup a base M4 chip (as in the Mac mini M4), the term “or better” means all larger chips (M Pro, M Max, M Ultra) of the current and previous generations. In most cases, the chip’s size does have a bigger impact on performance than the chip’s generation.

Tested Mac models: We tested the setups with several Macs but only list the cheapest option we recommend in the table.

Note 1: None of the recommended hardware setups poses a hard limit. The software does not limit its capabilities depending on the hardware, so you are free to test your specific use cases beyond the given feature scope or with even smaller hardware. Our recommendations propose the “safe” option for the mentioned features.

Note 2: If you want to test Reeltime with an available, maybe older “M” processor, please keep in mind that some of the M1 processors (i.e. in the first Macbook Airs) have no ProRes acceleration and thus will get to their CPU limits much earlier than other machines. We do not recommend to use the M1 Macbook Air notebooks for production where recording of more than one HD streams is required. 

Outlook & Conclusion 

With Apple’s move to Thunderbolt 5 ports on the new M4 and M3 Ultra Macs, and the fact that capture card enclosure manufacturers will also be switching to Thunderbolt 5 in the near future it becomes way easier to build a 4K multi-cam setup. Until then, we’ll add fitting 4K-setups to our list of recommendations. We’ve set up a place in our Knowledge Base where we’ll publish our latest hardware recommendations on an ongoing basis. To make your choice a little easier in the future.

The proposed performance recommendations give a good idea of how versatile and scalable Reeltime is. The basic setup can run on the baseline M4 chips, so it’s perfect for both beginners and veterans who want to use a lightweight and cost-effective setup on their film sets. From our experience we know that most users are using a MacBook Pro with at least an M Pro chip or faster, so they are all able to use the full feature set of Reeltime Pro and handle a complex multi-cam setup. It’s also great to see how the possibilities increase with more processing power, and the Max setup shows what Reeltime is capable of.