The order...

 At 2pm, SCAN.co.uk made the new generation of GPUs available. The NVIDIA 50-series. I struggled through a glitching website while configuring a PC with an RTX 5090. No luck! Only one model of the 5090 was available, apparently.

Then, when I'd almost given up, the website refreshed, and some systems appeared with the RTX 5090. I swooped in and stumbled a bit, then panic struck -- this is quite a lot of money. Order. Submitted. Payment authorised! Now, I wait to see if I am one of the lucky few. A lot of people were made aware of the 2pm release. If we can be cynical for a minute, we can recognise how you can overcome a customer's sensibilities by creating a sense of panic rush on an apparently scarce resource - driving nervous fingers to twitch on the buy button.

So now we ponder whether, if they've already run out of stock when they get to my order, should I leave the order in place and put my faith in a restock in the not-too-distant future? I don't know yet, but probably.

[Update: Order received and Payment Authorised. Payment taken. I received an email congratulating me on securing a 5090 machine but then asking me to stand by while they sort out the build and delivery date. Hmmm.. They have received a massive spike in orders and are a few days behind, or they sold out of stock in seconds and are now madly scrambling to organise a follow-up consignment. I expect to get an update on Monday, although I fear the update will be that my order has been delayed by a lack of GPU stock for weeks or maybe months. ]

More playing with World Creator and playing with the results using True-Terrain 5.

I compared the difference in apparent detail when using 16k vs 8k height maps. Hardly a sausage of difference! 

You might see some tiny areas where the 16k map introduces extra edges. I was surprised by the lack of difference. I can't see the value of 1GB maps if you can't tell the difference. I'll need to consider whether this is an expression of detail levels.

Calm before the order

 In two days time, the 500-series GPUs will be released. I planned to get a new system that carries the new flagship 5090 card. However,  it now seems likely that their supply will be super-low and their cost super-high. I came up with the idea of getting a machine that runs with two 4090s instead, but SCAN, the company that usually builds my machines, confirmed that they can no longer supply machines with 4090s. This is pretty terrible! NVIDIA has stopped making the cards, so their value has gone through the roof, which conveniently removes them from running should people decide the 5090 is just too expensive.

So we'll see what happens on Thursday. I'll need to order quickly to get a chance to get one. Even then, it might be that I have to wait several more weeks or months for restocks. That's if I decide to go ahead and order one.

What's pressing is that my current machine has a serious problem where turning it fully off leaves it in a state where it can take me several hours to coax it back to life. I should send it away to be looked at, but I need an exact fit box in which to put the case. Buying a new machine means that I can use its box and I won't be left without a PC.

Terrain work

I created more height maps in World Creator... since the Scout ship is on a break...

Islands

This exercise aimed to get an expansive ocean around an island using a subordinate terrain that sits beneath the one carrying the terrain. That worked. You often get an ugly line visible beneath the ocean where the underlying terrain's edge.

Project: Traveller Mk25 Scout Cleaned up

 

I cleaned up some of the Scout's worst issues and built landing gears. After applying some Sanctus metal materials, I felt I needed a break before deciding whether to continue the build. My feelings are that it's okay. Just okay. It sometimes happens: you sketch something that you really like. You do a build that you thought captured the sketch, but in the end, it's not got "the thing".

I landed the scout ship on alien terrain, but it's not working.  Maybe it just needs a final layer of detailing and better materials.

Project: Traveller Mk25 Scout Further Progress

 The latest on the Mk25 Scout:

Overall, I still like the design, although some elements have diluted the original feel and flow. This is a tad more streamlined, and the engine nozzles have shifted inwards from the winglet tips. The biggest problem, as always, is the loss of design coherence as small details break the design purity and inconsistency of things like panelling height and width start to make things a bit messy.

The build continues, although it gets slow as I add fine detail then delete it, repeatedly.

Project: Traveller Mk25 Scout Progress

 This morning, I restarted the construction of the original sketch design. I may return to the partially completed traditional design, but not yet.

By holding back on the sub-D step and getting more controlled angles in earlier, I got a lot closer to the feel of the sketch. Onwards!

Ad Astra! This is a good mid-point

Lots of cleaning and tightening, but I'm reasonably happy with this as a start.

Project: Traveller Mk25 ships

 The idea is to build a model of each of the classic Traveller starships - Scout/Free Trader/Subsidized Liner, etc with a modern twist. I did a sketch of the new Scout ship, which I liked. Alas, I have repeatedly failed to get the form into a basic Blender model.

The ship I want to build...but haven't managed yet.

I wasted a lot of time with more traditional variants. They always reached a point where I felt I was no better off compared with the previous Scout build.

At this point, I felt like I needed to show something for my time, but I stopped hitting my head against the wall in terms of the more complex geometry of the planned design. So, I whipped up a traditional build in a couple of hours.

I'll run this to completion, but then I may restart the above design. I'll probably watch some training videos for modelling first.

Challenge 13: Submitted

 It's done. Time to move on.

Competition entry

Video slideshow of the development:

Next up - the first in a series of classic Traveller ship rebuilds. This time, for sale!! Hopefully.

Monday progress >

 First day of the week, first target: Winter competition entry. 

Almost complete. 

Starting notes

Just a couple more elements...

I unified the landing platforms, so the Mantas now sit on a single plane. I want a fuel van and pipes to clutter some of that empty foreground tarmac. I'm playing with the idea of a patrolling soldier in the foreground. I have just the asset, but I tested it for the first time, and it needs some adjustment.

BigMediumSmall's spec ops. On sale!

I lashed up an APU/Fuel unit. It needs another hour of detailing/Fixing. Note the wheels -- I stole them!

I will complete the final render tonight because I expect a multiple-hour render time. For final renders, I tend to ramp up the samples and render size to get the best from the new volume and possibly avoid the need for noise filtering.

Busy week ahead!

 Sunday night feeling, eh.

Next week should build on the momentum of previous weeks, which was lacking most of this last week. Better focus, clearer targets!

1. The Snow competition entry: evaluate. Fix it, or rebuild it.
2. Complete the Materials and Shader course. Complete a write-up and start the poly modelling course.
3. Build a Scout ship! This will be a modernised version of the Traveller Scout. Ideally, it should be printable and suitable for sale. I should write a detailed specification. It's really about delivering a modern version of the old Type-S, one that chimes on the classic but looks like a starship from today, not 1977.
4. Blow the dust off Substance Painter. Write up a basic guide. This cost a lot of money, and I only used it a couple of times. It just seems like learning more of the workflow.
5. Blow the dust off the HardOPs course. Restart it in the following week.
6. Start thinking about a Blender portfolio. Project work needs to be cranked up with two quality notches.
7. Prepare to make a decision about the new system. RTX 5090? What's its actual cost going to be? It will be exorbitant. No question. My 4090 was the cost they were estimating. This will be another 200-300 on top of the RRP. You know, I might consider looking into a dual-GPU machine with 4090s if I could get them cheap enough. This seems unlikely, and it might require two power supplies, which means custom work on a suitable case. Expensive in other directions...

That's a point for each day. Reasonable? Ha ha..

Playing with terrains

Today was supposed to be a rest from Blender. I did make some big terrains. Nothing noteworthy. I need to work out how to do scatters on massive terrains. I should rewatch Maarten Nauta's and Martin Klekner's YouTube videos again for revision.

A break from the snow...

 I've broken away from the competition entry for a couple of days. I'll return to it on Monday, hopefully to complete it. During this break, I've played around with stuff and progressed with my CG Cookie Blender training. I'm halfway through the Materials and Shaders course.

I tried to composite my lander into a photo I took in Sheffield city centre. It's extremely rough, but it proved the workflow. It is much more complicated than I realised with handling and converting between colour spaces, Gamma correction and all that compositor colour-grading Jazz.

Very rough, but it got me through the compositor workflow. Oh, there's so much I need to learn..

Terrain sketch

Tower on the glen

Winter Competition...too...many..elements..to fix

 Yesterday, I reached a point in the creation of my competition entry where I was making change after change, but instead of each change taking me closer to a happy place, they spawned another set of change requirements. Add more light into this area, and it fixed one problem, but I then realised the texture isn't good any more because I can see it too well. The lamps don't look like they are consistently lighting the right areas, but when that's fixed, they ruin the look I was going for.

Cha..cha..changes. Too many.

What to do, where to go?

Option 1: Carry on. Go on, pull all your hair out.

Option 2: Restart from the basics and rebuild the scene elements. This has the benefit of opening up design and composition decisions that have been fixed in my head. The downside is that this will open up designed and composition decisions that may not be problems.

Option 3: Go away for a few days. Get a fresh eye. I have time! The deadline is the end of the month. This is the best option, but I worry that I won't return to it. I'll be too wrapped up in the next thing -- Traveller Ships ..for a publication. Honest, I'll finish it this time.

Winter Competition review point

 

The composition is mostly locked. I'm not building any more assets. So what's left to do?

Reflect and inflection

1. The dome wall texture is a bit mucky. I'll make it look more realistic in terms of casting patterns and stains. Maybe more lights and some garage-style shutter doors?

2. I'll put some unique decals on the two landers. Maybe I'll put some numbering on them. I might change their colour.

3. I should soften the silhouetting of the control tower and fuel tanks. They're almost black, but the tower is lighter. The tone should shift from lighter to darker, with the tower near black. I'll test that.

4. One of the lamps on the control tower is not lit, but it looks odd. I might need to make them consistent. There is not as much volumetric light being cast. I may tweak the volume fog a bit more to see if I can improve it.

5. I might darken the foreground snow bank. It has a blue cast on it to match the TT5 procedural snow.

Critique 1

Winter competition work-in-progress

I've just removed two attempts at populating the middle distance with a snow-covered forest. It just added clutter that didn't add anything, only distracted. I might try again near the end.
 

Two hours of control tower building...and ugh, its scale is off in key areas: Door, roof railings. What to do? I could scale those elements down, but .... <idea>

That works! Replicated the main building block. I now have a more slender, modular-looking building. I'll add lamps to that new middle tier. Progress!

Lander loiter or leave?

Manta Tactical Lift

 I'm in a good place, and the pressure to move on to other projects is too strong to ignore. Even if I felt I needed to stick with its development, I'd need a break at this point to cultivate a fresh eye. 

I'm getting on with either the informal Science Fiction scene competition with Berglmir or the True-VFX Community competition (actually set by Berglmir). The latter competition has a deadline, so I'll do that first. 

Sometimes, it just falls together

I started creating terrain heightmaps in World Creator. I didn't feel like I was getting anywhere, so I started to play with some of the results in Blender.

Starting point

A first brush at all elements

Of course, the lander was going to land on the scene.

I'm happy with this as a starting point. I need to add some additional points of interest. An antenna on the right-hand hill. So, snow clumps around the landing pads. Lighting and refuelling equipment. The concrete cylinder will need to be dressed up and lit. Maybe even a figure in the foreground. More atmospherics!

Lander: Even endier game: The final chapter

 Sarcasm aside, completing the lander model flew past me at high speed and is currently circling the landing strip and showing no indication that it intends to actually land.

Why? Standards, that's why. My standards levelled up about a year ahead of my actual skill level. So now I'm scrambling to fix the fixes that were originally fixes. This is good, though. This is me finally getting better because I'm making better judgements on what's good enough. When I'm back in a paid job, this will really slow me down. So, after getting better, I'll also need to get faster.

Not happy with the headlights.

Better headlights

Captain Leo is now in command

DecalMachine

Oh my god! DecalMachine, where have you been all my life? Answer: sitting on my storage drive waiting for me to test it, use it and fall in love with it.

What is DecalMachine?

The blurb says, "DECALmachine is a blender addon, which, through mesh Decals, allows for surface detailing in a very non-committal, non-destructive, UV-less way. As such it represents an alternative approach to hard surface texturing."

The big metallic pivots on the engine are not modelled, the're decals

The actual mesh. Note the wrapped plane from where the decal is projected.

Decals are incredible when used appropriately.

A quick test render

As I set up the duplicated Mantas, I realised that there were a lot of sloppy elements in play. Pivoting the engines and thrusters didn't work as it should. Some of the surface details looked sloppy. So, I stepped back from staging a scene to concentrate on a final modelling run.