Mash

Is this a new machine youve recently bought or have you already had it a couple of years? I guess my main question is why a pair of expensive, energy sucking xeon server chips and not a single more modern efficient threadripper? They're 14nm 200+ watt chips when you could have a slightly slower 32 core threadripper (but faster for everything else outside of rendering old projects) on much newer 7nm tech or a 64 core chip which would be faster and still cheaper than a pair of xeons.

I couldn't recommend either of those specs to anyone as a 3d workstation personally. The budget build blows its own foot off with a shotgun by using an intel arc gpu. Immediately most of the gpu render engines are off the table due to no cuda support, but beyond that I would never recommend a system with V1.0 gpu drivers for a work machine, its just begging for something to not work. Streaming is broken, video encoding is broken and you cant reliably render with them. Maybe further down the road when the drivers are sorted out it might be a viable choice, but not today. For the high end machine, what reason would someone pick 2 xeon cpus and then pair it with an A6000? This system makes a significant single threaded speed sacrifice in order to get those 48 cores, you would only pick this system if you are doing heavy 3d cpu renders, but if thats the case then why blow half the the system's budget on an A6000 card? If you're deep into CPU rendering territory then you would be much better served by a 64 core threadripper. If youre doing gpu rendering then you would have far better performance with an 8-16 core cpu with higher clocks and a pair of 3090 or 4090 cards. It isnt a bad machine, its just a machine which wastes half of its budget no matter which way you slice it.

I think you probably need to go spend some time researching the topic as you have some misconceptions about what they will offer and the main reasons for using one. First of all, there are no render engines where you will be able to open an old c4d project, press render and have a better looking image, it simply isn't going to help, and even if it did work, it would be missing the point. Whilst most engines will have a basic conversion function, that conversion is only a starting point to begin moving things over. There are several reasons to use another render engine, but you need to strike "it looks better" from the list, what you're really aiming for here would be "it looks better with less effort and in a shorter amount of time" All render engines have their own materials, their own lights, their own cameras and their own render settings, you will need to get familiar with these to make any progress. I'll use octane for my examples because its the one im most familiar with. In physical I have to spend time messing about with the clunky multilayered reflectance channel, dialing in obscure conductor and dielectric settings to get a realistic finish. In octane, just pick the starting material you want and 90% of the work is done (metal, glass, diffuse, glossy etc) In c4d I have to spend ages adding polygon bevels to my models to get nice realistic edges, In octane I just tick the "round edges" setting in the material and it does a high quality bevel at render time. In physical I have to restrict usage of area lights, area shadows, soft reflections and frosted glass because it destroys render times, in octane I turn on whatever I want because it makes no real difference. In physical I avoid GI because it adds a zero to the end of the render time and flickers if I get the wrong setting. In octane, GI is on by default and it makes no significant difference to render times. In physical I render out depth passes so I can apply DOF in after effects. In octane its so fast and looks perfect, so I just render DOF in the renderings. In physical I click render, go make a sandwich and 5 mins later I can see enough of the image to make a judgement call about whether my light has the right brightness, is in the right position etc. If not, I adjust it, hit render and browse reddit for another 5 minutes. Basically when it comes to the final look, Im making about 10-15 decisions an hour to get it looking right. In octane, the change is instant, maybe up to 5-10 seconds of rendering before I can make a decision. I can get 100's of adjustments done to my scene per hour. Regarding hardware and render speed, it opens up a whole world of opportunities. If you took a single machine and put a single 4090 gpu in it, you would possibly have all the render power you would ever need all in a single system. You would be able to churn out thousands of 1080p animation frames overnight, or 100's of high res 8k stills. In octane a full production quality 1080p animation frame takes us about 10-30 seconds to render with all the bells and whilstles on. That same frame with AO, GO, blurry reflections, best AA, motion blur, DOF in physical, would take an hour on a 16 core Ryzen system. If we're doing 8k stills then it might be 2-10 minutes depending on complexity compared to a couple of hours for the same still in physical. TLDR; You need to look at alternative render engines as an opportunity to get better looking images in a fraction of the time, and to be able to do all renders in-house with no more render farms. They're not a simple way to click a button and everything looks better. PS. just to say clearly what Srek probably isn't allowed to say. Redshift in CPU render mode is complete garbage and should be used by nobody. It has no reason to exist.

To add to the above, don't be afraid to cut, either in the 3d timeline or in post production. One pit that many 3d people fall into is animating one single never ending camera movement. In the real world they would have filmed the left to right pan with a camera on a rail or crane. Then they would switch to another camera shot for the closeup. It will help you to get a more natural animation if you do the same. Either animate the camera suddenly jumping to a new location over the span of a single frame, or render out 2 individual sequences and stick them together afterwards. Considering your shot, go watch some lord of the rings and find a scene where they do some ring closeup shots

Make sure "View > Use as render view" is enabled on your main camera's viewport.

Much better 🙂

You have several keyframes all set to disable your arrow extrudes.

All Octane dcc plugins are a single dev afaik.

Use pcpartpicker.com to check your system for compatibility Between the gpus, theres not a massive difference, theyre just different manufacturers sticking the same parts together into a product. The only differences are the fan cooling system and the warranty. https://www.techpowerup.com/review/gigabyte-geforce-rtx-4090-gaming-oc/38.html

I have no idea about the financials behind it, but if I had the choice, I would go back to opening up the forum for all visitors. you can ignore ads, you can ignore hit and run questions, I dont think theyre a significant detriment to the site.

Not sure. If you can whack that material on a cube and share it then we can take a look though.

Whats the texture tag settings next to the object?

Good catch

Whats a few decimal places between friends? 😉

For a 5 grand machine dedicated to computer graphics, 512gb ssd is a very odd corner to cut. I would personally be looking at 2-8Tb So it can act as os, cache, storage etc. theres no need for a ton of different os, app, data, cache drives if you just get one decent sized one. A nice setup is a 2-4Tb main drive to work from, then a 8-12Tb spinning drive for archives, backups, asset libraries. edit: gb -> Tb For the ram, aim for 3600mhz if you can, this will allow the cpu bus to sync at its maximum speed, will give 4-8% more performance. 64gb is a good amount, but 128gb removes it as a possible bottleneck when working with large after effects comps. cpu cooling, either a 240-280mm aio radiator. Avoid enermax, they skip on the liquid. Or a nice air tower like a noctua d15

I would start by just sticking a few boxes of wind along the tube to just try blowing the object through.

Why was my post deleted, which pointed out that your "female voice artist" was just a text to speech generator, and that you should probably tell your customers this before they buy the product?

If an 8gb scene file goes down to 12mb then not to put too fine a point on it, your projects a bit fecked. If I had an 8gb c4d file, I would expect it to have 100's of millions of polys. If it saves down to 12mb, then it cant have more than 100,000 polys or so. Given you have the project, which of these is most reasonable? My first guess would be that you have a plugin which has generated a ton of junk data. Or, that maybe you have some particle plugin like xparticles which had an internal cache, but the 2023 version of c4d you used didnt have the plugin, so that particle cache data is now gone.

Maybe contact one of the existing plugin devs and see if theyll extend the functionality for you? https://www.3dtoall.com/asset-juggler https://nft-randomizer.com/index.php/product/nft-randomizer-plugin/

Im still not sure thats right. The anamorphic setting by itself doesnt change the render time. If anything it halves your render time because you only need half the horizontal resolution. The important question here, is whats your final output? Are you delivering square pixels 9216x3164 to the client or do they want 4608x3164 anamorphic delivery? If its square, then, yes, maybe best to render to the full square pixels. Its the longest render time but will be the highest quality. If they want anamorphic then you may as well do the same in c4d and save half the render time. At no point does a 2304x3164 render make any sense if the final square pixels output is 9216x3164. I hope you didnt switch c4d to anamorphic mode just for the sake of loading in background footage? Just pre-unsquashing it to square pixels would have been a much simpler solution.

It depends what you want to happen. When you set an anamorphic pixel aspect, it changes the composition. Did you set the anamorphic ratio at the start of your project and have framed up everything you want with it enabled, or did you do the entire project and then flick it to 2:1 at the last moment before rendering? The number you see in brackets is the equivalent size that it would be if you werent rendering anamorphically. ie. your 4608x3164 anamorphic project would take up a canvas in most editing apps of 9216x3164. It wont double your render time, if you look carefully in the c4d picture viewer as it renders, youll see that the horizontal resolution is only half of the vertical resolution. So, my question is, are you really intending to output a mega wide 3:1 image from c4d? This would only really be used for some sort of imax/planetarium or other unusual project, no movie is going to be produced with such a wide aspect typically. I suspect what you might actually be looking for is a 2304x3164 render which then stretches out to a 4608x3164 square pixel canvas whilst editing. In short, rendering anamorphic should halve your render time, not double it. eg if I take a sphere and crank up my render settings to slow it down. If I render 1:1 pixel ratio at 1920x1080, it renders in 18 seconds. If I switch to 2:1 pixel ratio and then halve my horizontal resolution to squash the oval back into a sphere (960x1080) then it renders in 9 seconds as expected. tldr; halve your horizontal resolution

Ahh, I saw the lines as part of the original design 😮 Clearly the only solution is to change the maxon logo to fit the background image

but please what?

Are there any samples to view?

"problem: [R21] slow saving of .c4d file to network share" This was improved in R26, network saving can be >10x faster due to the new file handling code. Users report that on for example a 10gb network, saving speeds are around 500 mbyte/sec