Created on 2004-12-20

Title: Apache-Havoc: Razorback Interview
By: Len 'Viking1' Hjalmarson
Date: March 04, 1998 1110
Flashback: Orig. Multipage Version
Hard Copy: Printer Friendly

Razorback was founded by former Digital Integration employees whose credits include Tornado, Apache Longbow, and Hind. Yes, these guys have some solid simulation experience! In fact, Tornado is still considered one of the best simulations ever made and maintains a solid following.

Here are new shots from their first project, helping to relieve the pain of FN2s loss. Razorback’s first title is ‘Apache Havoc’, a combat flight simulator featuring two adversary attack helicopters, the American ‘Apache Longbow’ and the Russian ‘Havoc’. Razorback also intend to create a follow on sim titled ‘Comanche Hokum’ which features two scout/attack helicopters, the American ‘Comanche’ and Russian ‘Hokum’. Both titles will be fully interconnectable creating an awesome combat helicopter package! Here is our interview.

Csim: Tell us a bit more about Razorback. What do the team bring to sim design?

RZK: Razorback Studios was formed by myself, Matt Smith, Todd Gibbs and Dave Proctor. We are based near Oxford in the UK.

At some time or other we have all worked for Digital Integration on their titles 'F-16 Combat Pilot', 'Tornado', 'Apache Longbow' and 'Hind'.

With the help of Empire Interactive, Razorback and the 'Apache Havoc' project began in February 1997. The team has now expanded to 8 people (4 programmers and 4 artists).

We are a small but focused team. Everybody has an input into the design of the simulation as opposed to one person dictating it. This means that we are all driven to produce the best work that we can.

We view Apache Havoc as a platform to build from. We started a year ago with zero lines of code and no graphics. In a sense this helped because we had no legacy code to hold us back and we were starting afresh with DirectX. Apache Havoc has been designed as a multi-player game from the ground up and this will be self evident in the finished product.

Attention to detail is an important issue for us. For instance, we are determined to bring a product to market which has the best looking cockpits to date. Our Apache and Havoc cockpits have been painstakingly constructed as 3D models (we even built a 3D pilot to sit in the cockpit so that we knew the cockpits were ergonomically correct and so that we could position the viewpoint exactly where the pilot’s head is). From the 3D models (which have in excess of 100,000 polys each) we have rendered 30 different views per aircraft. With all the instruments backlit for two different levels of night lighting that adds up to 180 rendered cockpits.

Csim: Out of all the possibilities, why was Apache-Havoc chosen as the first project?

RZK: Firstly, we have a preference for helicopters as opposed to fast jets so it was always going to be a helicopter sim. Secondly, we wanted to create a real multi-player game and that meant that we needed two adversary aircraft for competitive games.

The given choice was 'Apache Havoc' and 'Comanche Hokum' as these pair better with the Apache and Havoc being attack helicopters and the Comanche and the Hokum being attack/scout helicopters.

We went with Apache Havoc as the first project as at the time we had more information available. However, we have now compiled enough data to go ahead with the Comanche Hokum project.

The great thing about pairing the Apache and the Havoc is that there is so much contrast between them. For instance, the Apache has the glass cockpit whilst the Havoc has more traditional analogue instrumentation.

In Comanche Hokum we are actually simulating the two-seater Hokum, the Ka-52 Hokum-B, sometimes referred to as the 'Alligator'. Again there is plenty of contrast with the Comanche. In the Hokum the pilot and CPG sit side-by-side as opposed to tandem so this should make the cockpit graphics interesting having to include people! Also the Hokum has ejector seats. The co-axial rotor blades are blown off and out come the crew. Could be interesting if your flying in tight formation!

Csim: Screen shots look very, very good. Can you compare the graphics engine of A-H to other sims we are seeing this year: DiDs Total Air War, Janes Longbow, Falcon 4?

RZK: Our graphics engine has been specifically designed for low-level helicopter combat where ‘line of sight’ tactics are essential. Therefore our terrain is smothered in dense forest canopy with roads, rivers and lines of electricity pylons cutting through it. All of the fields are encompassed by hedges and brick walls so ‘hedge hopping’ is possible.

With fast jet sims the action usual takes place at higher altitudes so the priority is different. Recent fast jet sims I’ve seen use large textures to cover the terrain which works fine unless you are flying very low then you notice that the roads and rivers lay at an angle on the terrain. In our system roads, rivers, etc. are ‘cut’ into the terrain with terraced banks so that they lay flat.

Sense of speed is another important issue. The ‘texel’ size needs to suit the speed of the vehicle. The large texture map approach means that you either have a blocky low level visual or you are forced to have small maps.

Our maps are 256km by 256km in size and currently use about 40Mb of data per map so at least changing combat zone doesn’t involve going away and making coffee!

Csim: Apache-Havoc will model forests and trees. Tell us more about this, how much tactical significance will this have in gameplay?

RZK: As I’ve mentioned above, the dense forests make ‘line of sight’ tactics a real part of the gameplay. The player will have to learn to use the terrain for cover.

Also, there are hidden dangers in the forest ...

Csim: What resolutions will be available to the player?

RZK: Apache Havoc is fixed at 640x480x16 but we may include higher resolution graphics later.

Csim: Will we see light source shading, smoke and fog, dynamic lighting?

RZK: All of the above. Our artists and the 3D engine programmers work hand in hand to develop special effects. You won’t be disappointed.

Csim: The modeling of wind and weather is a real growth area, probably because of the horsepower required! Tell us about what is being done for A-H.

RZK: Weather effects are essential as they add to the atmosphere. Imagine flying a difficult mission in fine weather and then flying the same mission in poor lighting and adverse weather conditions...

We have dynamic time of day and weather effects. This means that the lighting and sky textures are constantly changing as opposed to just changing the effect when the mission is completed, the curtains close and the player goes into the menu screens.

In our multi-player dynamic campaigns, you don’t have to wait for all the other players to get back to base before the next mission starts so the time of day and weather must be dynamic too.

In fact, in our campaigns you are always active as a component in the game. Just because you got back home safely doesn’t mean you get sanctuary whilst you’re rearming and refueling. Oh no.

At the last Paris Airshow myself and Matt were in an Apache. It was there that we noticed the wiper blades are so prominent. That inspired us to make the wipers fully functional in the sim.

We may introduce seasons into the sim. We’re still working in this.

Csim: We've seen some very, very good flight modeling this past year with Janes Longbow and the design goals for MiG Alley look very ambitious. What can we expect to see in A-H?

RZK: Our flight models use a full force model which exhibit all of the required artifacts of helicopter flight. We know some Apache pilots who are coming to give Apache Havoc a test flight soon so hopefully we should be able to get things just right.

Csim: Some sim design firms have forged ties with para military organization like Janes and World Air Power. These ties provide a flow of information that contributes to realism in design of flight models and avionics. How have you acquired the info you need for accurate design?

RZK: Basically from researching the subject. Getting books, magazine articles, videos, the Internet, speaking to the manufacturers (amazingly we got some information from Mil at the last Paris Airshow). We have amassed more than enough information to do the job.

I’m not sure how much information that Janes and World Air Power give to the other sim companies that they don’t already publish?

Incidentally, World Air Power is an excellent and affordable journal. Get Volume 29 for the 60 page article on the Apache.

Csim: Systems modeling has likewise been a growth area for sim design. How much detail will we see in avionics for the Havoc and Apache compared to past sims like HIND and Janes Longbow?

RZK: We will make the systems as authentic as we can. We also want to make the game accessible to the novice player as well as challenging for the accomplished pilot. So there may be some realism options.

Csim: Weapons and physics sort of fit together in sim design. Tell us about the goals in this area.

RZK: We’ve modeled the weapons systems so that they behave as you would expect. The algorithms take account of weapon/target aspect etc. so that if a weapon has to maneuver heavily it’s range is reduced.

We have a good intercept point algorithm so the guided missiles move quite convincingly.

Csim: What will we see for AI opponents in terms of aircraft and helicopters?

RZK: Our current vehicle inventory includes 50+ models. There are 12 different helicopters and 8 aircraft.

All of our models are built accurately and have many forms of articulation. For example, on all the helicopters, the rotor blades start off drooped and then straighten out as the rotor RPM increases and then become motion blurred. It looks great especially on the helicopters that have tandem or co-axial rotors.

Csim: Will AI opponents obey the same physics and human limitations as the virtual pilot must face?

RZK: All AI vehicles move according to a force model so they have the same physical limitations. AI pilots have parameters which affect their performance.

Csim: How many AI team mates will we normally fly with? Will AI pilots grow in skill over time?

RZK: We're not sure about the number of team mates yet. AI pilots will improve over time. You can accelerate this by flying as that pilot and ‘training’ him. I’m not sure if you can make him worse though!

Csim: Comms and interaction with other assets has been a huge growth area this year, with DiDs ADF and even Innerloops JSF breaking new ground. How important is this element in A-H?

RZK: Very important. After reading some accounts by combat helicopters pilots I know how much communication goes on in real life. We really want to capture the essence of that.

Csim: Will we interact with FACs?

RZK: Should be comms with FACs. We’re planning for targets of opportunity to be included so some missions may get diverted on route. The FAC will be involved here.

Csim: Tell us about tasking/mission planning. In Janes Longbow we could construct our own flight and choose from various missions. Is this the way the campaign works in A-H?

RZK: Basically our whole AI system is task driven. For instance, if a base requires resupply then a request is made and if a suitable vehicle is available the task will be picked up and executed.

So as tasks (or missions) become available the player may choose to accept one or plan a task of his own. Route planning will also work in multiplayer mode.

Csim: I understand that the campaign settings are not final. What are the likely settings for the campaigns?

RZK: We’re just sorting this out at the moment. I can’t give you any locations as yet but we will have 4 combat zones two of which are land based and two of which involve an amphibious assault. We really wanted assault ships in the game and as it turns out the Russian Kiev is a real center-piece object.

Csim: I understand that the campaigns will be fully dynamic. Tell us more about campaign AI and structure.

RZK: As mentioned above the dynamic campaigns are task driven and should never play the same twice. The underlying structure is that there is a front line, predominantly a tank battle, behind that are the infantry and support vehicles and behind that the artillery.

Advancing forces try and capture ‘key sites’. The player may get involved in the front line action or elsewhere.

Csim: Will there be a Theatre Commander mode of play where the player controls all assets in the field?

RZK: Not in the first release but this could be included in an upgrade.

Csim: How much control over other assets will we have. Will we be able to call in back up or artillery?

RZK: The player can request artillery and air strikes. This will go through the tasking system so you may have to wait a while until the task can be executed.

Csim: Will resource management be a factor in the campaigns?

RZK: Yes, it’s an integral part of it.

Csim: How are downed pilots dealt with? Will we be able to perform rescues or will we have to call in SAR?

RZK: This idea has been around the loop a few times. Yes, it would be great to create SAR tasks for downed pilots. Whether human players want to hang around waiting to be picked up is another matter ...

Csim: I understand that troops will be modeled. What can we expect to see, and in what way will ground troops be tactically significant?

RZK: There will be ground troop insertion and extraction missions and definitely ground troops with shoulder launched SAM.

Csim: Some sim designers have been attempting to integrate other factors into AI, like morale, fatigue etc. Will we see any of this in H-A?

RZK: Yes, the AI pilots will be affected by morale and fatigue.

Csim: Damage modeling is another growth area in sim design, moving from probability models to actual ballistics and sub system modeling . What will we see in Apache-Havoc in terms of damage resolution?

RZK: There will be various forms of damage which can be incurred to the airframe and avionics. I particularly want to implement undercarriage damage so that the player is forced to make a difficult landing with say one wheel blown off!

Csim: How sophisticated is the AI that goes into the ground war? What tactical options will be available to a SAM unit. Will we see hide and evade AI, for example?

RZK: AI movement is controlled by a ‘route planner’. Basically if a vehicle wants to go from A to B all of the possible routes are considered and the best path is chosen. The factors which are assessed include whether any bridges are impassable, line-of-sight exposure, radar coverage etc.

Csim: What about multiplayer support? How many players will be able to connect in one mission? Will there be a "capture the flag" option as well as coop missions? Will we be able to fly coop in campaign mode?

RZK: We’re not committing ourselves to a maximum number of players. We’re leaving that open ended because we don’t want to limit the game for players that have more bandwidth available.

Any campaign can be flown co-operatively or competitively. People can join at anytime as long as there is helicopter available. This is already working in our system. Apache Havoc was designed from the ground up as a multi-player game.

Csim: When Janes Longbow added the CPG role it really added depth to the gameplay, but multiplayer for both positions won't be in place until Longbow 2. Will H-A allow the mplayer dual roles in one chopper?

RZK: We have not included the CPG seats in this release. We opted to concentrate our effort on the pilots cockpits. We have built hooks into the program to make this facility available in an upgrade.

Csim: Razorback have already stated their intention to release a second simulation based on the Comanche and Hokum. Frankly, this is sounding VERY cool. How long will we have to wait to see the second release? What will Comanche/Hokum add to the fun?

RZK: You will definitely see the second component within a year. Comanche-Hokum will add two new helicopters to master and they are more suited to attack/scout missions. C/H will be interconectable with Apache Havoc and they will be able to share combat zones. In multi-player games you can make ‘hunter-killer’ teams.

Csim: What is the direction for 3d support? Will we see proprietary support for 3dfx as well as other chipsets?

RZK: We’ve gone for Direct3D support. This should make Apache Havoc compatible with most 3D accelerators.

There’s seems little point in backtracking and producing a 3Dfx Glide version. The D3D drivers seem are very optimized now. We have been working very closely with the 3D chips manufactures and Apache Havoc already works on most cards that we’ve tested it on.

Csim: Will there be any support for new hardware like force immersion, virtual reality, multiple monitors, surround sound?

RZK: We are looking at Apache Havoc as a platform to build from so if we don’t support any particular hardware on the first release we have the opportunity to do so at a later date. We really need some customer feedback to know what is popular.

Csim: What do you expect will be the minimum system requirements?

RZK: This is very difficult to speculate. I am currently developing on a 200MHz Pentium Pro machine and that is more than adequate. I use a 133MHz Pentium machine with an early 3Dfx card for testing multi-player code with and that runs ok too.

We will probably have some options to vary visual quality and frame rate.

Look for Apache-Havoc in October of this year.