Exploring Heatblur's New F-4E Phantom II Simulation Platform: A Technical Report

[Super Grover]

Hello everyone, 

My name is Krzysztof Sobczak (some of you may know me as “Grover”), I am the Technical Director at Heatblur. Although I haven't been very involved in direct communication with our users because I have been focusing on working hard on our projects, today, I am excited to share some technical insights about our latest release. As some of you may read from me for the first time, I'll start with a short introduction: I am originally a physicist holding a PhD in particle physics. However, most of my professional life has been dedicated to developing flight simulations. I have had the privilege of working with companies like A2A Simulations and Metrea. In 2018, I joined Heatblur Simulations to work on the F-14 Tomcat. During this project, my areas of responsibility included the communication and navigation systems, the LANTIRN targeting pod, and JESTER LANTIRN. In 2021, I became the Technical Director to create a new core system - codename Anvil - a genuinely new-generation simulation platform.

Having been involved in creating several flight simulation engines throughout my career, I have gained a deep understanding of the requirements for such engines and the shortcomings of existing solutions. This extensive experience has been invaluable in shaping the form of our new flight simulation platform, which we present today through the release of the Heatblur F-4E Phantom II. Our users may already be familiar with the innovative approach we applied in the F-4E Phantom II, particularly the concept of the components system. For those with programming experience, this might initially sound like a marketing trick, similar to object-oriented programming and using objects through composition in different contexts to build complex structures. However, there is much more depth to the system we created. Our approach goes beyond simple object composition and reuse. What sets our system apart from other solutions and programming techniques is its orientation around data flow. This focus is crucial for creating a system that is both reusable and granular enough to build entire aircraft and reuse components in other aircraft or systems. The challenge in making such a system lies in the uniqueness of many aircraft devices, so to achieve true reusability, we had to reach deep to a level where components can be generalized - the level of individual switches, relays, lamps, amplifiers, actuators, valves, and all other kinds of low-level components. By focusing on the components and the data flow between them through the connections, we managed to move all non-simulation elements outside the components and let Anvil handle them.

The first core aspect of our F-4E Phantom II, driven by our engine, is its multi-crew capability. The component connection approach allows our developers to focus entirely on creating the simulation part of each component. At the same time, the system handles the communication between different components and distributes it over the network. From the developer's perspective, there is no difference between developing an aircraft simulated entirely on one computer and an aircraft whose simulation is spread across multiple computers.

The second feature unlocked by the component system is multi-threading. By strictly separating components from each other and not allowing them to access other components directly, we ensure that the simulation part of each component can run concurrently. This approach means that developers do not have to worry about the complexities of multi-threaded development and can work as if the simulation of each component were entirely single-threaded. The connection system governs the rest — it de-conflicts multiple simulation threads seamlessly. It is important to note that while our system is designed to support multi-threading, this feature is not enabled in the version that users are receiving today to limit the risks of the adverse effects for the premiere of such a complex aircraft. However, multi-threading will be enabled in future updates, offering a significant boost in performance for everyone. Also, note that this is not the same as using the multi-threaded version of DCS - the multi-threading part of our component system adds another layer of parallelization, effectively offloading the aircraft simulation from the main DCS thread.

The component system is oriented toward providing developers with the tools to create new components quickly and efficiently. With those tools, we can focus on recreating components' real-life properties and giving them an organic feel by adding imperfections, individual characteristics and failures that can occur as a natural result based on the treatment of the components. Later, we use those components as bricks to build the aircraft in a fully synthetic way, while Anvil makes each airframe unique and each flight a new adventure while giving virtual pilots an incentive to not mistreat the aircraft.

Finally, we plan to use Anvil and the same library of base components in all our future products, retrofitting them to the already published products where applicable. Sharing the same code will ensure that all our future aircraft will be automatically updated with our latest achievements and changes introduced to the system while developing any new product. This means that from now on, our simulation can only get more profound and more realistic.

After introducing key features of the component system, let's delve into some statistics about DCS F-4E Phantom II. The total count of components used in our simulation, as in the final build provided to Eagle Dynamics, is 6,389. We used 14,627 component connections. Out of these, 2,297 connections are synchronized in multi-crew. The total number of properties across all components is 34,964, with 25,032 of these properties synchronized in multi-crew. Many of these properties are randomized and depend on the wear and condition of the aircraft. Finally, the total count of all damaged states in the aircraft is 1,562. While this number is already high, we plan to implement many more new failures during the early access period.

Some of you might be concerned about the costs of creating detailed products like our F-4E Phantom II, which could negatively impact the entire market. We never intended that, and we are confident this won't happen. First of all, our goal was to make the development of complex modules less expensive by improving the reusability of the code and offloading most of the low-level and repetitive development to Anvil, which will handle it automatically. Furthermore, we are part of the flight-sim community and want the study-level flight simulation to thrive. From the first days of this new platform, we shaped it in the form of SDK, which we want to offer our partners. This includes the entire library of components and JESTER. We believe that cooperation and fostering synergies is more creative and profitable than non-amicable competition.

As you will be able to experience the F-4E Phantom II for DCS yourself, I hope you will enjoy every minute spent in our recreation of this magnificent aircraft. On behalf of the entire team, thank you for the support and excitement you express daily through all communication channels.

Have fun!

Edited May 22, 2024 by IronMike

[G.J.S]

A pleasure to hear from you.

Some incredible insights there, ground breaking. 
Thank you for the post. 

[Talisman_VR]

Thank you Grover for making contact and posting such informative and inspiring information.

Good luck with the release.  

Happy landings,

Talisman

[VR Flight Guy in PJ Pants]

A simple like could not express my appreciation.  BTW, I am going to count how many of these components are shaking in flight.

[Hiob]

Interesting read! Much appreciated! 

[Maverick87Shaka]

26 minutes ago, Super Grover said:

and offloading most of the low-level and repetitive development to Anvil,

Really good deep-dive of your Phantom, Love it!
What exactly is this "Anvil" @Super Grover?

[Qcumber]

Sounds amazing. If something stops working in the cockpit can you fix it with a good thump? 

[Alicatt]

@Super Grover I see a lot of black text on a dark background, can you please reformat that for those of us that use the dark version of the forums?

 

Thanks 

Edited May 22, 2024 by Alicatt

[average_pilot]

11 minutes ago, Qcumber said:

Sounds amazing. If something stops working in the cockpit can you fix it with a good thump? 

I was actually thinking about creating a thread asking for that feature to be implemented. Some poking on the gauge to make sure it's not stuck, or if it is, to un-stuck it.

Edited May 22, 2024 by average_pilot

[Super Grover]

1 minute ago, Alicatt said:

@Super Grover I see a lot of black text on a dark background, can you please reformat that for those of us that use the dark version of the forums?

I'm sorry, I hope it's better now.

 

19 minutes ago, Maverick87Shaka said:

Really good deep-dive of your Phantom, Love it!
What exactly is this "Anvil" @Super Grover?

It's the codename for our core library/tech managing all components and their communication.

[okopanja]

On which bases do you intend to provide SDK to 3rd parties?

[Grundar]

I think for 3rd party development, access to an SDK as you have described is excellent, but it is also a big boon for DCS players. There are so many multicrew aircraft out there (some DCS appearing already and many many more which have not) that would benefit.  Jester, for example, certainly makes the single player experience a lot better and I am all for airframe accurate simulated systems that can fail (or behave) as it adds that bit more complexity to an already studious game.

[Glide]

This is great news!  Go team!

[IronMike]

39 minutes ago, okopanja said:

On which bases do you intend to provide SDK to 3rd parties?

We recommend that any interested 3rd party contacts us via our official channels, and we will be happy to discuss the particularities with them.

[Raven (Elysian Angel)]

Very interesting to read!

thank you!

 

[gnomechild]

1 hour ago, Super Grover said:

From the first days of this new platform, we shaped it in the form of SDK, which we want to offer our partners. This includes the entire library of components and JESTER. We believe that cooperation and fostering synergies is more creative and profitable than non-amicable competition.

This is amazing. Absolutely love to see this. A rising tide and all. Great work and thanks for the detailed write up. Super interesting stuff!

[pimp]

As a Software Engineer myself, I think I found a new role model. Anvil sounds amazing. I wish I was this smart.

[LizardKingYB]

Nice article!

[polosat1y]

1 hour ago, Super Grover said:

total count of components used in our simulation, as in the final build provided to Eagle Dynamics, is 6,389. We used 14,627 component connections. Out of these, 2,297 connections are synchronized in multi-crew. The total number of properties across all components is 34,964, with 25,032 of these properties synchronized in multi-crew. Many of these properties are randomized and depend on the wear and condition of the aircraft. Finally, the total count of all damaged states in the aircraft is 1,562. While this number is already high, we plan to implement many more new failures during the early access period.

Testing all of it must be hell. Kudos to the QA team and all involved in the project.

[Scatlet-565]

Please ED, get in touch and send George to Heatblur College XD

[flameoutme]

I knew, based on the F-14 which was superb from day 1 (with minor exceptions), that behind the curtain at Heatblur were some extremely talented individuals.  Extremely. 

Now we get to meet one of them, Grover. PhD particle physics!   Welcome Grover to the board and thank you for making my flight dreams (while comfortably snacking in my seat), a reality.  

I personally hope you can find time (in your spare time!) to tell us more about the F-4E.  But even if you can't, may the force be with you and thank you!

Flameoutme

[Katsu]

2 hours ago, Super Grover said:

inally, we plan to use Anvil and the same library of base components in all our future products, retrofitting them to the already published products where applicable. Sharing the same code will ensure that all our future aircraft will be automatically updated with our latest achievements and changes introduced to the system while developing any new product. This means that from now on, our simulation can only get more profound and more realistic.

So in a future Viggen and F-14 will be "Moved" to this new "Engine"? 

Looking forward to this. 

Thanks.

[Lime_Acid]

I hope all this crazy amount of synchronization will work better in a network game than on Tomcat

[Super Grover]

It's "smart" , so it should only happen when a state change needs to be synchronized.

[Aqil Huseynov]