Virtual Car Reviews - Ginetta G55 GT4 2017 V1.1
This is hopefully the first of many detailed and thorough reviews of car mods for Assetto Corsa. I decided to start with the Ginetta G55 GT4 by Shaun Clark as we were thinking about the possibility of running it in a league with the Porsche and Maserati. Below is my thoughts:
-----
3D Model:
Perhaps the highlight of this mod is the 3D model, specifically the exterior and a few select areas in the interior.
Exterior - Polyflow looks extremely professional, use of ‘fake-chamfers’ is comprehensive, and density seems spot on. This would be 5/5 in isolation.
Interior – Unfortunately this is where things start to take a turn for the worse. It appears that whoever was responsible for modelling the exterior so beautifully suddenly became unavailable, as the interior is in a different league. The centre console, dash and door frames/sills seems to be the only redeeming areas of the interior model. Elsewhere the detail is extremely low, lacking chamfers, lacking in accuracy (pedals, roll cage welds non-existent, steering wheel etc are all prime areas very visible from the driver seat).
3D Model Result: 3/5 (Great exterior for the most part, poor interior)
Textures/Shaders:
Exterior – For the most part this is ok, with the skins looking reasonably decent. Some small issues like not enough margin/padding on the AO bake (most visible at the rear edge of the doors), as well as the fact the exterior shader does not make use of a proper RGB map based on the AO bake – which means you have a flat amount of reflection throughout which leave darker areas too shiny.
Few small areas such as the exhausts are lacking AO altogether to my eyes.
Interior – Much like the 3D model, the quality of texturing is significantly worse on the interior. Most objects lack any proper AO bakes or RGB maps, which only emphasises the lack of detail in the mesh.
Parts such as the metal foot board really suffer from lack of AO, and are extremely bright in VR, and the steering wheel/column look very flat which is a shame given they are so close to the drivers view. The bright coloured buttons/cabling next to the fire extinguisher are also very bright in VR and lack any real textures.
All of the above texture issues would almost be forgivable if it results in a very small, optimised file size, though the next section shows this unfortunately is not the case.
Elsewhere there are a few priority issues such as the black window borders looking through the door glass.
Textures/Shaders Result: 2/5 (Acceptable exterior but poor interior)
Optimisation:
File size - 70MB/60MB for the LHD/RHD respectively – This is not great considering the level of texture detail is low, ideally to justify this the file size should be significantly lower than the Kunos standard of around 42MB.
LODs – Only 3 LODs (A,B,C), which would be fine if they were especially efficient whilst still to keeping great visual continuity. Unfortunately this is not the case, in fact these LODs are very poor indeed.
LOD A from a tri count perspective is fine – well within the limit, though at ~240k tris, I do wonder where they are being spent, as the interior mesh is severely lacking in detail.
The LOD B and Low Resolution interior mesh is particularly troubling, as instead of reducing the density of the polys, is seems the only effort has been deleting a few objects (brake calipers for example on the LOD B) The LOD B sits at around 125k tris (should be around 30k).
However the LOD C is by far the worst, coming in at around 75k tris (!) with almost no interior (making it possible to see the track through the windows). A typical LOD C should be less than 15k tris, and with the lack of a LOD D we have nothing that comes close to a low object 3k tri model for long distances.
Combine that with the fact the LR interior doesn’t appear until around 40m (LR Interior should be around 7m with the B coming in at 12m), you end up with a grid of full fat LOD A’s, and the ones that have switched to LOD B’s are not much better for performance.
Frankly this is unacceptable for a V1+ mod, as performance is very poor as a result.
In order to test this I ran a full grid at Barcelona with the Porsche Cayman GT4 as the benchmark. The overall tri count on the grid was ~8million vs the ~3.7million of the Porsche. FPS was around 20% lower for the Ginetta.
Optimisation Result: 2/5 (1 would be no LODs at all, at least that effort has been somewhat made).
Physics:
Ideally this car should match the laptimes of the Porsche and Maserati GT4, and thankfully it does (within reason, track depending etc).
I do however feel that despite overall laptimes being very similar, the ease in which you can reach them is significantly easier in the Ginetta than the other two rivals. The Ginetta seems to have endless longitudinal grip, making traction and braking zones far easier to perfect, whilst the much lower power seems to be the only thing keeping them balanced.
I don’t necessarily dislike that, having cars that excel in some aspects and struggle in others is fine, though the disparity seems too large for my liking (though that is understandably subjective – I don’t have definitive data that says otherwise).
Aero seems well balanced and offer realistic levels. Similarly the setup options seem good and exactly what I’d expect from a car of this nature.
Elsewhere there seems to be a few small issues, namely the fact tyre wear in the Ginetta is not dependent on load like every Kunos car is these days.
Autoclutch seems to not only be implemented poorly (race starts are much slower than the other 2 GT4 cars), but seemingly unnecessary given this has a clutch pedal (either the model or the physics is wrong in that case).
Physics Result: 4/5 (Performance seems realistic though the manner in which it does so seems suspect).
SFX:
I’m not particularly well qualified to judge SFX – usually most sound fine to me. In this case I have no issues, nothing outstanding but perfectly acceptable.
SFX Result: 4/5
Other Comments:
Steering animation could use some work, appears as if the driver has learned how to use ‘the force’ at points.
Pit stops seem incredibly unbalanced compared to the Kunos GT4 cars, with just 7 second tyre changes vs the Porsche and Maserati’s 20 seconds. Additionally body repair takes half the time, and engine repair takes twice as long, meaning longer races can very quickly become very unbalanced.
-----
Conclusion:
Whilst I had high hopes that this would at least provide variety/filler for the GT4 class, I’m afraid the poor optimisation and unbalanced pit options put a stop to that, which is a great shame as there are some very promising aspects of this mod – notably the exterior 3D model is high quality and the overall laptimes are in the right ballpark for the class. Unfortunately a few major and many minor faults have soured the overall taste.
It remains to be seen whether the Supercup version resolves these issues, being much more popular I have hope that is justified – I’m expecting to test and review it soon.
3D Model: 3/5
Textures/Shaders: 2/5
Optimisation: 2/5
Physics: 4/5
SFX: 4/5
Overall Result: 3/5
-----
Next Up:
The recently released Dodge Viper by A3DR from the NFS Tournament Class B pack.
Sponsored Pimax Crystal Light Giveaway: Win A Freshly-Unveiled High-End VR HeadsetTwo new VR headsets join the Pimax lineup: The Chinese manufacturer revealed both the Crystal...
Wisniewski Wins Twice at DTM Esports Pro’s Penultimate RoundFive of the six DTM Esports Pro rounds are complete for 2024, with Nikodem Wisniewski winning...
Luke Whitehead: Meet The 24-Hour Solo Sim Racer Who Finished 3rdTo celebrate the launch of the Nürburgring 24h pack on Assetto Corsa Competizione, publishers...
Could you provide a replay file or something that I could watch?