Jump to content

hvanderwegen

Limited Member
  • Posts

    593
  • Joined

  • Last visited

  • Days Won

    19

Everything posted by hvanderwegen

  1. I thought so too, and adjusted or deactivated the ones I could find. Made no difference at all. Did notice that a dark coloured transparent water would even introduce a noticeable anti-aliasing on the left side of that dark line. To me this seems to be a bug. Or unintended behaviour of materials interaction. Doesn't happen in Physical or other render engines I tried (Cycles, Eevee).
  2. How odd. When "Show Samples" under the Sampling tab is activated, it shows a clear increased number of samples across that line of intersection. I went through a lot of RS settings, activating and deactivating them, even resorting to trial and error, and that line remains stoically present in the render. It definitely is caused by the intersecting polys. After a while I got a bit frustrated, played around with some of the assets in the debug scene and instead wasted an hour of my life on this (not RedShift): Sigh. Maxon's support agent "fix" is not really a fix and a rather silly one, in my opinion.
  3. hvanderwegen

    Inkscape 1.2

    Well, in my opinoin it makes more sense to feed a (semi)final render as a live referenced file in an image editor or comp app like Fusion, Nuke, AFX or Blender compositor. For example, PhotoLine, Photoshop, and Affinity Photo all allow for referenced file layers to be added, which then can be adjusted with live adjustment and filter layers. Both PhotoLine and Photoshop use a live smart layer (PS: Smart Object | PhotoLine: Placeholder Layer) which reference the render as an (for example) EXR file. Affinity Photo also offers an option to place a live referenced link to a file. The difference between these three: when the render file is updated with a new version, only PhotoLine will update it in real-time behind the scenes. Photoshop requires the user to double-click the smart object and re-save, and close it, while Affinity Photo user must update the live link in the resources panel/dialog. PhotoLine is much more flexible in cases when multi-pass render layers are saved as individual files and live-linked: no need to manually update the render layers. And in PhotoLine z-depth passes may be cloned and re-used throughout the layer stack - something not possible in Affinity Photo and very awkward to do in Photoshop. So re-render one or more passes, switch to PhotoLine, and everything is automatically updated. Very convenient 🙂 Every click saved is one more second in my life 🙂 Interestingly enough I mentioned this years ago to the PhotoLine developers. They implemented the next best thing: live round-tripping to InkScape. It means it is possible to send a vector layer / group to InkScape from PhotoLine, and edit it, then save, and PhotoLine updates the layer. This link remains live for the session as long as InkScape and the layer contents is open. Or place an InkScape SVG as a live referenced linked layer, and when updated in InkScape, the contents updates in PhotoLine. It is not a perfect link: certain live effects in InkScape are not compatible with PhotoLine, but for vector editing it works perfectly well. And btw, PhotoLine has quite adequate vector editing tools as well. I do use it for the vector plugins available for InkScape, though. And some of the other options such as live tracing (which PhotoLine can do, but InkScape's black and white vector tracing is just better). PS the live link also works with other apps: send a layer to Krita, apply G'MIC filters, save and PL updates the layer. Add more effects, save, and PL updates it again. 🙂 Corel! I used to teach Corel products a long time ago. I am finding the PhotoLine / InkScape / Krita triad to be an excellent replacement with additional features that Corel is missing. Much less destructive, superior bitmap editing workflow. Although I concede that Corel still has a few tricks on its sleeves that are missing in other apps. In the end it all depends on one's particular workflow and job requirements, of course. Agreed! Same here. I also use Affinity Designer for a couple of things, but generally I find Affinity's range of apps to have too many rough edges, silly GUI decisions, and paper cuts.
  4. Right-mouse click on the move, scale, rotate, and combined transform tool, and choose "Assign Shortcut". Then type the letters you'd want to assign to those tools. The shortcut will then work as in C4D and other DCCs.
  5. ByGen is free via their GitHub account or GumRoad. https://curtisholt.online/by-gen But it is kind to support them.
  6. The newer tools allow for fall-offs to work while changing the radius and tilt - which is not possible with the control point parameters.
  7. Don't forget that the latest versions added new tools to control these intuitively as well! (as well as the shortcut keys) (These happen to be part of the toolbar in your last screenshot on the left)
  8. To demonstrate Blender's curve options, check out the official demo file. All curves: (ps this is a screenshot of the viewport) https://www.blender.org/download/demo-files/
  9. I opened the object in Blender to double-check, and it seems to have hundreds of loose vertices as well, even after doing a quick Mesh-->Clean Up-->Merge by Distance to connect the geometry. I deleted these with Clean Up-->Loose Geometry, and then recalculated the normals (just in case). But there were still normal issues with a number of faces: Which were resolved with Mesh-->Normals-->Average-->Corner Angle. One of the ends was closed with a N-gon, and I deleted that because it caused shading issues. It needs to be closed manually. How was this object generated?
  10. One of the reasons why I work in PhotoLine for comping: any layer (or group of layers) may be "virtually duplicated" - or cloned/instanced and reused anywhere in the layer stack. Even adjustment and filter layers! This means one mask is reusable as a layer mask as well, and cloned/referenced layers can be further adjusted with other adjustment and filter layers, because unlike Photoshop, in PhotoLine a layer mask is treated like a regular layer. This allows for tremendous freedom in the layer stack, and understandably may reduce a typical comp file size considerably: a single mask is recycled rather than copied. Krita also allows for layer clones, but it is not nearly as versatile.
  11. I take it then that you've never opened Krita before? 😁 Layer blend modes, anyone?
  12. The funny thing is that PhotoLine is only a few years younger than Photoshop - the first version appeared on the Atari ST! It is developed by two German brothers ever since. They never put any effort into marketing at all. I first happened on PhotoLine while looking for an alternative for Photoshop when Adobe went down the rental route with CS6. At the time, I had tried everything: from Gimp to Corel to Paintshop Pro and any other image editor I could find. Nothing quite fit my workflow as Photoshop did. One link to an obscure website mentioned PhotoLine, and I visited their page. The site looked terribly outdated (it has since been updated): I almost closed it, but after spending weeks of testing, what was one other badly functioning image editor? Downloaded it, opened it, and was pleasantly surprised by the layer stack: I literally laughed out loud at the genius behind allowing a layer, filter, or adjustment layer to have its layer opacity to go beyond 100% (up to 200%) and actually allow for negative value down to -200%. And as many layer masks as you wanted, great vector support, and so on. It had some unique features I had never seen anywhere before (it still does!). The GUI was pretty outdated looking, though, and barely configurable. And there were a lot of paper cuts. The devs have been incredibly receptive and responsive to requests over the years, and while the GUI may still not be the prettiest, PhotoLine has become a damn fine workhorse. Some things are just utter genius. It actually outperforms Photoshop in some key areas: for example, image mode and bit depth are layer dependent, rather than file dependent. And it even allows for layered 1bit bitmap layers, which is truly helpful. It is also possible to set it up for round-tripping layers and files to external software. Send a vector layer to Inkscape, edit, save, and PhotoLine automatically updates the layer. Or send a layer to Krita, paint, save, and again the layer is updated. Or send a layer for export directly to another external app. It's also (in my opinion) the most powerful alternative for Photoshop in terms of raw graphics editing capability, yet ironically the one least known about 🙂. And to be fair, most users would probably be attracted more by Affinity Photo's GUI look, and granted, Affinity Photo is a great alternative for many. I have Affinity installed as well, but mainly use it for things like its focus stacking. Affinity has too many paper cuts for me and a number of show-stopping limitations (for my work, at least). PhotoLine is my main hub for image editing and still comps.
  13. Agreed, I do this all the time between apps. Btw, are you aware of this? Simple copy and paste between DCCs. https://github.com/heimlich1024/OD_CopyPasteExternal I used it before, but I am unsure if it still works for the latest C4d release - no updates in a long time ;-( Still works for Blender, Houdini, and LightWave. Unsure about the other apps.
  14. @zedenIn both PhotoLine and Krita one or multple layers may be fed into multiple cloned masks. In PhotoLine any layer mask may consist of an entire dedicated layer stack, including live effects and adjustment layers. An entire comp file may serve as a re-usable single source for anything really. Anything can be cloned/instanced for re-use as masks. Similar to After Effects. That said, it is possible in Photoshop as well with clipping masks/layers - up to a point, and I find these rather awkward to work with. Agreed! If only selections would support more than 8bit in Photoshop...
  15. A fellow PhotoLine user! Don't be surprised that Photoshop is somewhat more limited in terms of non-destructive editing. I have been a Photoshop user since version 3.5 myself, although I no longer use Photoshop for my own personal or professional work if I can help it. The main issue with Photoshop is its legacy code base. Layer transformation is a good example. It is possible to non-destructively transform a layer, but only when the content is placed as -- or converted to -- a Smart Object (similar to Placeholder layers in PhotoLine). The drawback is that (just like After Effects and its pre-comps) the content can no longer be directly edited: to edit a Smart Object's content in Photoshop or a comp in AE, these must be opened individually in a new window. Even for something as simple as a simple pixel edit. In PhotoLine (as you are aware) this is not necessary: a transformed layer remains editable and the transform itself is non-destructive as well. Affinity Photo also allows for non-destructive "image" layer transforms, although those layers can again NOT be edited until those are rasterized to editable "pixel" layers. After that conversion the transforms are non-destructive. In short, Photoshop relies heavily on Smart Objects to do anything non-destructive. Any layer or groups of layers must be converted to a SO before a filter (outside the standard adjustment layers) such as a gaussian blur can be applied non-destructively. Both in Affinity Photo and PhotoLine this is not the case, of course. It is for example possible in PhotoLine to convert a layer or group of layers (or import an entire file either as linked or embedded) as a Placeholder layer and then apply almost any type of filter as a non-destructive one. And that includes many compatible third-party Photoshop filters! Just like Photoshop! So in PhotoLine the user is given a choice, and often it is not even necessary to convert a layer to a Placeholder (SO) layer. The main reason why Photoshop limits non-destructive filters via its Smart Objects is simple: performance. The senior developer (Chris Cox) at the time decided that non-destructive filters (unlike adjustment layers) would become too slow and degrade performance too much. (Which did make sense at the time.) And while this may still be true when working on heavy and complex layered comps/projects, for most regular jobs live effects like the ones in PhotoLine and Affinity Photo live effects work just fine nowadays with existing hardware. For very heavy compositing with high resolution files and high bit-depths (32 bit EXRs anyone?) it becomes essential to "pre-comp" stuff using Placeholder layers in PhotoLine as well. Also something to consider when working with Photoshop: the so-called "16 bit" image mode is actually a 15bit one. This is again a legacy core code part that was never updated after the 16bit mode first was introduced in the nineties of the previous century. The implication is that when a full-range 16bit render is opened in Photoshop (or a 32bit image downgraded to 16bit) that only half the values are retained of a full 16 bit mode. This is problematic for HDR work, for example, or scientific purposes. Even more problematic is that Photoshop does NOT warn its users against this. So a 0-1-2-3-4-5-6-7-8-9 ~ 65535 range becomes 0-0-2-2-4-4-6-6-8-8 ~65536 (the infamous "15bit+1" programming trick to simplify calculations and speed those up). Photoshop throws away the rest. I kid you not. Selections are always done in 8bit space. Which means your lovely 32bit or 15bit image part that was just selected, copied and pasted is severely downgraded to an 8bit range. This remains unfixed til this day. Many filters still refuse to work in 32bit in Photoshop. No such issues in PhotoLine or Affinity Photo. (Yes, the user must take responsibility for some of the filters downgrading the range of values, of course.) Which is all the result of Photoshop's aging legacy core. And there are quite a few features missing in Photoshop compared to PhotoLine, Affinity Photo, or Krita: multiple layer masks per layer? Sorry! A -200 up to +200 layer opacity value range (only PhotoLine does this!). Freely combining layers with content at different bit depths and colour modes? Sorry, need a Smart Object in Photoshop (PhotoLine allows the user to combine these freely in the same layer stack). And so on. Obviously I could point out a few things that Photoshop does better, but it's mainly down to its way of handling channels and of course plugin support. 3D is deprecated in PS and no longer counts as an advantage (in spite of the horrible implementation - good riddance, I say! 🙂 ) All in all, I found PhotoLine to be the most flexible of image editors in terms of workflow and adaptability. Its layer stack is arguably one of the best. It has a number of tricks up its sleeve that no other image editor offers.
  16. No, as I said: curve profiles only work for curves. Any curve object in Blender may be extruded and/or beveled and/or tapered with the built-in curve parameters. I failed to mention that no modifiers are required in this case 🙂 It's actually very flexible. The newer Blender versions also add a nice "Add curve as bevel" option in the Add Curve menu (which I only noticed today 😉 ). Ensure to activate all the extra curve object addons, because it allows for a plethora of different curve object possibilities.
  17. Curve objects also offer the profile option in Blender. Absolutely agree! Performance (or rather the lack of it) is currently its main challenge. It's at a state that it is actually driving users away to switch to other DCCs. Scenes that are no problem at all in other DCCs lock up Modo and take ages to load. It is do-or-die for Modo now: fix these issues or become a footnote in DCC history.
  18. This is a worst case scenario for banding. When I open your file in C4D, I actually do not notice any visible banding on my screen, but when I save out a PNG8 or PNG16 and open the files in PhotoLine, the banding becomes very apparent. If I export as EXR32 it looks just fine in PhotoLine. I notice RedShift's noise doesn't seem to work that well in this particular case, and areas with and without noise are introduced. That, in my experience, may cause unwanted banding when opened in image editors: There is much more to be taken into account, though: the reason why you are experiencing banding in C4d on your system and I am not is probably caused by the interplay of our screens, the graphics card, and the software/driver which may introduce visible banding - I am not too up to date with the technicalities here, but your screen may or may not exacerbate the visible banding. Not even mentioning whether your screen is calibrated, or what colour gamut your screen supports. Too many variables. But in short, it is impossible to avoid banding between very low colour tint transitions such as in this case when working in 8bit or even 16 bit - and screens cannot deal with those transitions without adding some noise to break up the visual banding. I applied some noise to the EXR32 version in PhotoLine and saved it as an 8bit png - and presto: no visible banding (ps these two images were grabbed by zooming out to 50% - the RS version looks as bad as 100%, the PL version looks better at 100%): When I compare the noise patterns between RS and PL, the more uniform noise pattern produced by PhotoLine just works better. Ergo: do not rely on C4D and RS to produce an acceptable result in this case. Save as a 32bit EXR and open in an image editor that can deal with that. Then apply a noise layer and adjust and export an 8bit PNG. This step requires a bit of trial and error until you arrive at a version that doesn't look noisy and still removes (most) of the visible banding.
  19. @luchifer I have noticed quite a few users are interested in creating (as they call it) a retro render look from the nineties. And I have to admit that it can be relatively hard to achieve a similar look in unbiased render engines (which makes sense, of course). Last week one Blender user complained about that Cycles and Eevee are incapable of rendering that old-fashioned monochrome green CRT wire effect (which, granted, was very easy to do in the old biased renderer). After some tinkering around for a few minutes, I proved them wrong (Eevee real-time): wires0001-0180.mp4
  20. In my own experience, unless you need light linking / grouping, e-Cycles is overrated and expensive. With Cycles X the speed differences are marginal, so I've noticed that he e-Cycles (and k-Cycles) devs tend to market the extras such as light linking. k-Cycles also includes these things, and is only $49. Also, the upcoming 3.2 release will feature light grouping as standard. The important ones are free, though. 🙂 Tex Tools, UV Packer, Magic UV (built-in), UVSquares Yeah, these are more than awesome. I can't live without. These (and the ones listed in HardOps's preferences that it seamlessly integrates with) take hardsurface modeling to the next level and beyond. Ridiculously affordable too (even though all Blender addons are GPL, I prefer to support the devs).
  21. @BabumbolVizRT acquired NewTek (the company behind LightWave) in March/April 2019. A very vaguely worded statement was then issued that (to me) indicated that LightWave would no longer be actively developed. Others disagreed. One more update was released after that in April 2020 with a final bugfix release in July 2020. In the period leading up to and following the 2020 release and the final bugfix the entire LW team left NewTek or were reassigned to other projects within VIzRT/NewTek. Management involved with LightWave moved to other opportunities as well, even though some (like Dr. Andrew Cross) kept stating that "a path" for LightWave to survive would be in the cards. Cross is also no longer with NewTek, btw, since earlier this year. In the past two years all sorts of rumours about a potential third party buyout buzzed around the remaining LW community. In leaked emails it seems NewTek wants 1 million dollar and assurances in place that LightWave will be in good hands for continued development. It seems that third party withdrew its interest at some point. (Again, these are all rumours!) Some said the company (forget the name) behind Octane supposedly offered a deal, but were rebuked by VizRT. Octane's support for LightWave stopped last year, which was a big blow since LightWave's new unbiased render engine, while of a high quality, is CPU only and cannot compete very well with GPU ones (it is rumoured again that the LW devs were working on integrating GPU support). LightWave users are (understandably) incredibly frustrated, and some have even gone to the lengths of calling out VizRT managers during public online VizRT events and on Linkdin. These were ignored and removed by VizRT. The saga continues with Kelly "Kat" Myers (from liberty3d.com) now attempting to purchase LightWave from VizRT/NewTek to ensure its survival. He says he has a plan to make it successful once more, but my opinion is that that ship sailed a long time ago. Kat's been talking to VizRT management at NAB last month to acquire the LightWave code. He is very VERY passionate about LightWave, a very experienced LW user who used/s it in production for decades (Iron Sky comes to mind), but a bit of an outspoken and divisive character in the LW community. That's pretty much what I know. NewTek/VizRT have been (as per usual) stoically silent about LightWave. Amazingly enough NewTek still sells LightWave on their website, which I find rather offensive, because it is no longer developed or even bug fixed (and the last 2020 release introduced a number of bad bugs which are still unresolved). From a nostalgic perspective I would like to see LightWave make a comeback (I've used it on and off since my Amiga times!), but the realist in me knows better: LightWave needs a LOT of financial input and development time to be brought up to modern standards. And who is going to fund that? In the meantime, LightWave's community is dwindling, although a dedicated core of experienced older users keeps creating very good interesting art and content. Many though have left at this point, or are transitioning to alternatives.
  22. Well, the old Physical is part of a group of legacy ray tracers that produce a certain look quite fast. C4d users are not the only ones finding themselves in a state of render engine transition 🙂 LightWave users complained when LW 2018 removed the old render engine and replaced it with a modern up-to-date unbiased one. Old scenes had to be converted, but since the new renderer's material system is quite different users had/have to convert these manually. You can imagine many people were not that amused and stuck with LW 2015 or kept it around to open those older scenes. And currently a fairly large share of LW users still prefer 2015 or earlier. It was a mistake of Newtek to remove it altogether without allowing for a transition period. Blender made the switch to Cycles a while ago and removed the old Internal render engine - but this took a while so that users could adapt and smoothly transition. Eevee was introduced as a replacement aside from Cycles, but still there remains an ardent small group of users who insist on bringing the old Internal back to life. Likewise, Houdini is in the middle of a render engine transition as well. The new Karma GPU renderer is nice, but it still does have a tendency to crash for me on occasion. Mantra CPU doesn't cause issues (aside from taxing/heating up my CPU). It's a good thing that C4d management has decided to keep Physical around for now. But nothing lasts forever. At some point they'll remove it (I'd hazard a guess, and I expect it to be gone by S28). Btw, I do agree that Redshift GPU should have just been the standard for all C4d editions. I tried using the CPU version, and it is... SO.. SLOW... I can imagine anyone using standard issue C4d without a RS/Maxon One sub thinking that RS is a mediocre alternative to Physical. It renders quite well with GPU, though.
  23. @HippoDasTamus This is the one: https://www.rendereverything.com/groups-for-blender-addon/ Look for the new entries under Object-->Parent.
  24. I don't really like threads like these, because all DCCs have their pros and cons. But I would like to respond to these things: Not quite so in my opinion. A Blender file is similar to a database in structure. This has its advantages and its disadvantages, and I do like collections and scenes to organize content, and empties to create hierarchies. The Collection Manager addon is a great assistant (turn off QCD) as well. An added advantage is that any part of a Blender file may be linked or appended easily, and I love how collections work for easy referencing/linking - not only in other files, but also within other scenes. That is something just not possible in C4d. If you are coming from an app like Maya or Max, group hierarchies are indeed something that is missing in vanilla Blender. There exist a number of addons to assist with this, however (for example, Hierarchy Tools or Group Pro). Not too sure about this. I find I hardly need to merge stuff, but I suppose it also depends on how a user approaches modeling. I use HardOps and Boxcutter for hard surface models, and I keep things non-destructive for the most part - even with quite heavy projects. I do agree that it would be awesome if modifiers would work on collections. In essence I work similarly in both C4d and Blender in this regard. That said, performance in Blender is generally just better. Blender easily deals with thousands of non-instanced "real" mesh objects in a scene and lots of modifiers, while C4d slows down rather easily. Scenes that Max and Blender have no issues with (dare I say even LightWave Layout), C4d still chokes on before even coming close to equivalent complexity. I don't share that experience, although I do agree Blender's units require more attention. Hardops does make a difference, though with its AccuShape options for more accurate modeling. ? I can drag images in the 3d view or node editors. Or drag and drop blend files. Or use the file explorer to dig into blend files and drag and drop stuff in my scenes. It is true, however, that importable 3d objects like *obj and Collada files need to be opened via file import. Would be nice if Blender allowed for drag and drop here. PS I find that for example obj import in Blender generally works better than Cinema4d. Fixing textures all the time is no fun. 😉 I'd rather have that work than the small niggle of a lack of drag and drop, because the time saved is pretty huge sometimes. To me that is an advantage, rather than a disadvantage. I prefer that over tabs. And C4d lacks the concept of scenes, which I use profusely in my projects. Different way of working 🙂 That depends: if the primitive is added through Geometry nodes, its parameters remain editable. I also use the free Wonder Mesh addon, which adds a bunch of useful parametric primitives to Blender which remain editable just like C4d. That was indeed one thing I missed from C4d. And Hardops of course offers some basic parametric primitives as well. Again, it depends where you look in each app. If you'd asked me about this 15 years ago I would have whole-hardheartedly agreed with you. In its current state I really can't anymore. C4d has a tendency to introduce half-baked features and which then see hardly any follow-up development. Sculpting is a good example. Bodypaint a train wreck. Particle systems? Another good example is Redshift integration as a renderer in C4d: I have been testing it, and I do like it, but Oh Ye Gods the implementation and integration is still rather shoddy compared to Cycles in Blender. I am aware this is not quite a fair comparison and we're comparing apples and pears, but still: after this many years of development, I would have expected a bit better. I won't go in details here. Anyone who is familiar with the two apps will understand what I am saying. In my view C4d used to be this solid piece of fully featured kit. It still is dependable and solid (up to a point), and by throwing enough money at it to amend its shortcomings with plugins, a really nice DCC - yet at the same time (to continue the house metaphor) it's like a villa with many half-built rooms and missing sections. The entrance and lobby looks great though with a double golden staircase leading up to a few beautifully finished bedrooms. Move beyond that, and dig a bit deeper, and you'll find that a lot of construction work is still either to be finished or hasn't begun yet and staircases and hallways that lead to deep pits. And a number of very basic things are missing compared to other DCCs. That is not to say Blender hasn't got its own share of issues, of course. Currently I use Blender, C4d, Houdini mostly in my work and for my personal projects, with a spot of Max at my workplace. None are perfect. Each has its own quirks. ...I wonder what kind of building metaphor would describe Houdini...
  25. By the way @BIgor what are your temps at idle?
×
×
  • Create New...

Copyright Core 4D © 2023 Powered by Invision Community