Renderosity Forums / Poser - OFFICIAL
Welcome to the Poser - OFFICIAL Forum
Forum Coordinators: RedPhantom
Poser - OFFICIAL F.A.Q (Last Updated: 2025 Jan 24 6:22 pm)
Subject: My Poser 6 opinion
I always love people that try new software for a couple hours then lable it crap... try waiting to figure out how everything works before condeming it, it's possible that the scene you've set up isn't optomized for P6. The way morphs and textures are handled is different and your old scene very well could be bloated compared to the way new scenes will be created.
My opinion: Your demands are too high. Every tool has its power parameters and you are expecting too much from these incredibly inexpensive programs. The scene is not problematic....your personal approach is. Plenty of people are accomplishing plenty...they respect the limits of the tool, make what greatness they can from the tool while it is in the 'sweet spot' of its intended use. They then make/purchase other tools to expand to fulfil their needs. Even though your language in your last sentence is over the top, and you have an attitude, here is a free tip anyway. Take it or leave it: Why don't you use the new area render and or shadow catcher, then composite in post. ::::: Opera :::::
I don't think I've every loaded that many characters in a scene, but when I couldn't render a large scene (over 150 mb file), I usually take it into Vue 4 and have less of a waiting period to render it. One thing I wish Poser had was layered scenes like in Vue. That way instead of hiding every character and piece of clothing you could hide layers and render certain parts of a scene especially when the file is large. Generally all of my Poser files are large.
Poser 2, 3, 4, 5, 6, 7, 8, 9, 10, Pro 2014, 11, 11 Pro
I think you are missing my point in this. I agree that the file is not optimized and it is large and yes there is a lot of other tools and many ways around any problem. It is not like I have not managed to do anyting with Poser 5. I am sure if I rebuild the file from scratch I may have more luck. The point is that I payed a lot of money for Poser 5 even if it was cheaper than anyting alse and Curious Labs gave us a great upgrade deal but, by and lage I still have nothing for my money. The free DAZ studio beta is now more sucessful at taking this file and rendering it than Poser it self. Yes DAZ studio has a long way to go still. The point is Poser still does not understand what needs to be accomplished. I would have settled for active suport for EMC and the likes, MAT file creation and a easy to browse library. Right from the begining Poser 5 ran into problems for not creating a program that works but, adding all maner of new things. Poser 6 is supose to have taken this advise. I see no sighn of that. DAZ studio has a better idea of wich way to aim. I just hope it gets there. Poser 6 is not it.
Extract: "I would have settled for active suport for EMC and the likes, MAT file creation and a easy to browse library." Comment: EMC and the like was based on a bug in preexisting version of poser that was fixed. A known, stated bug -- as in, a flaw of the program. They took the code and fixed it. You are now asking them to break it again. (Note that they did appear to have enhanced the internal code that the bug was around, as JCM seems to work better in Poser 5 and 6 than in 4.) MAT file support is actually built into Poser 6. So are several other features that were specifically asked for prior to Poser 5's release. Strikes me that they are indeed listening. Extract: DAZ studio has a better idea of wich way to aim. I just hope it gets there. Poser 6 is not it. Comment: Then it appears that DAZ Productions is indeed going in the direction you wish to go. In my opinon, you should stick with DAZ Studio. Learn it. Master it. Push it beyond what the designers intended it to be. Share what you learn. Increase the sum total of knowledge about the program, and listen not to those who would defame your choice. However, to be equally fair, do not do so where it is unwelcome. Do not walk into the poser forum and proclaim it the second coming, nor other forums, for that matter, as some have done. DAZ Studio has a forum for itself here. Let the battle commence. Daz Studio 1 versus Poser 6. It will split the community and fracture the market, but what the hell. It's always more fun to have to pick a side.
Well, there you go. :) My download, installation, and registration went flawlessly. After a bit of orientation, I was able to complete a render I'd not been able to complete before using Firefly in P5. I'd have to say my experience was the opposite- and I'm old (55), grumpy, and a diehard cynic who doesn't expect to accomplish anything (by anyone else's standards, anyway).
Ok it looks like I am not able to make my point. If I realy did not think Poser has any value I would not have botherd using Poser 5 past the first week. I will be using Poser 6 from now on. I have used Poser 5 every day since its release. I just see so much more potential considering the size of the comunity around it and the content available. I don't think anything can match this. I just wish Curious Labs could see it as well. DAZ saw the potential and is trying to cash in on it in it's own way. I am sure I will come to terms with Poser 6 as I did with Poser 5. I am just very disapointed with what they acomplished between Poser 5 and Poser 6. I expected more. I expected to at least be able to load a Poser 5 created file and render it even if it took 2 days for a single frame.
This is going to sound like a rather unusual question. Can you describe the mesh sizes, texture sizes, amount of transparency, amount and number of ray tracing capabilites, and complex shader settings used in the scene? There is a very simple reason for asking. I do not, myself, use Poser at this time. I'm actually here to laugh at the absurdities. However, I have a very close friend who does, and they have a fair knowledge of the inner workings of it. It is my understanding that if during the loading of or calculation for the scene, you exceed the 2GB cieling imposed by the operating systems bit structure (such as windows or macos 8, 9, and 10), then the render will not finish. Poser does a great deal more than DAZ studio does. D|S, therefore, has a much smaller memory footprint than Poser does, which means that there is more "space" available for it to work in. That, in and of itself, could actually be the reason that Poser cannot render the scene while D|S can.
There is a known problem with P6 memory handling.
Unfortunately by the time we had discovered this on the pre-release it was too late to make any further changes. Curious Labs, however, have been notified of the problem and have told us that they will address it in a future update.
The general memory requirements for storing morph data in memory are somewhat larger (~33%) than in P5 and as with complex figures such as the DAZ Unimesh line the majority of data storage (pre-render) is actually the morph targets.
On top of this files (cr2 or pz3) that are loaded with external morph data stored in the new pmd files actually seem to consume double the amount of memory, or to put it another way you will only be managing to get about 40% of the figures in a P6 scene than you would in a P5 scene.
This additional memory footprint does disappear when the figures are deleted, so its not leaked memory and won't affect stability, but is currently a major limitation and one that we consider to be serious enough to count as a bug.
If you are stuck trying to get scenes to work in P6 it may be a good idea to configure the General Prefernces to store internal morph targets as per P5. This will increase loading time but should alleviate memory problems when they are loaded.
The good news is that, as the morph loading and storage processing has been changed in P6, it can now be considered to be "live" code and it should therefore be open to further improvements. They may well want to consider Studio's lead in resource handling.
The actual changes needed to substantial improve memory handling are extremely trivial. I spent about 4 hours today writing an improved loading routine for the pmd files, one that recognises identical morphs and stores them without duplication.
CL are due to get the code tomorrow along with our test results, whether they decide to run with it or do there own thing is their business but it is worth keeping some polite pressure on them to improve the situation as it is an area that promises great improvements at very low cost.
Bill
Message edited on: 03/22/2005 14:51
Hi Opera We have an independent business. We just managed to wrangle a late pre-release copy from CL for compatibility testing but unfortuanetly too late for any credible beta testing. Overall though I should add that P6 seems extremely stable and I was one of those who found P5 to be seriously problematic in it initial release. Even with the resource usage issues I would still recommend P4/P5 users picking it up earlier rather than later, if there are large scenes that P6 currently chokes on you can always revert to your previous version, but in the meantime on lighter scenes you can certainly spoil yourself with the new features. BTW [as an aside] I forgot to mention that personally one really promising new feature that I don't think has had a mention yet is the new Gather node in the material room. I plug it into the raytracing input (there isn't a wacro) for quite credible fake radiosity effects. Bill
Attached Link: http://www.renderosity.com/messages.ez?Form.ShowMessage=2172519
Bill, would the problem you are reporting cause the big hit (way slower render times) on the Jim Burton benchmark, as reported on another thread (attached link).I have already ordered, just awaiting Box. CL is getting very good reviews for stability in general, and the goodies all seem delicious.
::::: Opera :::::
Actual counts is a little dificult to get. I have 6 copies of V3 no morps 1 copy of V2 base preteen boy 1 copy of V2 based preschool boy 1 copy of David no morphs 1 Copy M3 no morphs about 40 or so props from Poser primitives(ball, box, torus and cylinder 3 copies of lowres DAZ wedgecut 5 copies of a cutom prop with just over 300 polygons each Lowest posible res textures in each case 6 infinite lights now shadows for ambient light 1 infinite light raytrace shadows. In Firefly I have limited the texture res to 1024. Only hair has transpareny. No bump or displacement. Over time I have tried a number of settings and eliminated any fancy shader nodes. Just tex map and solid colors. My computer specs: 3.6Ghz P4 with hyperthreading. 2Gb DDR2 memory. Nvidia PCX5900. 700Gb SATA drive space over 4 physical drives. Windows has it's own 10 Gb partition. 6Gb swap file fixed size on it's own partition. 200Gb Poser partition with 100Gb free space. I might include this is my 25th year living infront of my always self built computer. First as programer then CAD user in the electronics industry and the last 5 years or so ever increasing 3D animator. (Yes I used Windows 2.0 and Micro Grafix desighner. LOL)
I never did get what a fake radiosity effect is... what is it? Is it lighting as though the source is inside the mesh with shadow based on how oblique the normals are or something? I have never seen it... could you show an example... do people like it because it looks cool or for "historical" reasons or what? Thank you.
Oliver, that is a scruptious render.
May I/we ask on behalf of us "medium intermediate" node warriors this:
the new Gather node in the material room. I plug it into the raytracing input (there isn't a wacro) for quite credible fake radiosity effect. <<<
You would do a great service if you understand Bill's method and if either of you could post a Poser6 Render using this, and also post a screen shot of the node connections. Thank you.
::::: Opera :::::
Message edited on: 03/22/2005 23:31
Having returned from my daily joy (I make people feel small for a living), I am able to look closely at your scene as described. I note you use the phrase "Lowest posible res textures in each case" in your list. According to the person from whom I am getting my information, that means that those textures are roughly 2048 by 2048 in memory, and effectively 16 million colors at the minimum, for each figure. That means 10 textures for the people, alone. Even with the texture resolution limitation (which only applies at rendertime and only to calculations performed), those items are going to occupy between 1 and 1.5 GB of memory all by themselves (images are stored uncompressed in memory). That does not give you much play room for shadow maps, calculations, mesh storage, file storage (10 CR2s with all the joint morphs is still a lot, and the joint morphs are not injected), and similar. It is suggested (I am discusssing this via IM at present) that you open the textures for each of those figures that are in use and save out a copy of them at a 1024 by 1024 size. The squaring will not affect the texture. Then apply those resized versions and attmept your render again. Also, check the size of shadow maps if still in use -- there are some lightsets that use extremely high numbers that are a holdover from earlier versions and are poorly suited to the lighting features in Poser 5 and 6. Shadow maps above 1024 should be used with extreme caution and only if the shadow cameras are properly set for the scene, and shadow maps of 2048 are a waste of resources. After that, give it a shot at rendering.
Me? I have no clue, truly. I do not use Poser. I have a very smart friend, however, who is very much into using it, and those were her suggestions. However, on a system with specs of 512MB of Ram, a 2GB VM, and a 2G CPU, the person I've talked to has rendered 15 unimesh figures in a raytraced environment in about four hours. This same person also noted that the major problem with the unimesh figures' mesh density (what the hell is mesh density?) is their eyes and mouth parts, which combined have a greater polycount than the rest of the figure. She figure it was an oversight and error on the part of their makers. Outside of that, the largest issue with the unimesh figures is the injected cr2s. Far too many people inject the full morph sets, needlessly increasing the sizes of the cr2s. This was not the case in his example.
This same person also noted that the major problem with the unimesh figures' mesh density (what the hell is mesh density?) is their eyes and mouth parts, which combined have a greater polycount than the rest of the figure. She figure it was an oversight and error on the part of their makers. No, that's intentional. DAZ increased the polycount in those areas that need it most (particularly the face) for better morphing potential. Mesh density generally refers to the polycount or resolution of a mesh. Detailed areas of the mesh need a denser (higher) concentration of polys to define that detail.
stewer, I didn't mean advantage over three poser lights and depth map shadows, I meant advantage over some of the OTHER new approaches now open to us such as SoftRayTrace and IBL and AO, or are those considered entirely different from radiosity as targeted by Poser's method of unbiased single bounce monte carlo path tracing (had to say that at least once!). ::::: Opera :::::
Gather is not faster than AO, because Gather needs to evaluate the full surface shaders on every ray hit, where AO just needs to know whether or not it encounters a surface. OTOH, Gather provides actual light transport like color bleeding, which AO doesn't do. For most cases, I think AO is the better solution as it's easier to set up, faster and more predictable.
I am aware of all the solutions to rendering this scene that I was complaining about and I have managed to render it in Poser 5. The point is still that without any intervention from me it did finaly render in DAZ studio but, to get it to render in Poser I have to resort to all kinds of tricks. I would have expected Poser 6 to handle it. I realy don't care if it took 7 days to render because data has to be swaped in and out of memory, or what ever it had to do to make it work. The problem is that Poser6 once again just died a silent death without a hint. Poser 6 certainly has inprovements. The hair room impressed me. The walk designer is still hopless though. The larger undockable library is nice. Tree structures that stay open and do not have to be re-traversed every time is nicer. so the list can continue endlessly. For every positive there is still one or more negatives. I am not just talking about heavy technical stuff, I am talking about every day common sense. I use to program I have a pretty good idea what goes into it. I personaly believe if I can not program it, it does not exist. Every program I see causes me to spend time trying to figure out how it is done and how I would have attempted it. Every thing I see makes me try to figure out how would one get a computer to do the same. Much of the things we finaly have on computers, I foresaw as 'n 18 year old boy in 1981.
I see. What you have run into then, is an issue of presumed expectations. That's a tricky and unsatisfying habit. It means that, ultimately, any program you ever deal with will be, to some extent, displeasing. It would be comparative to my walking into a call center or a restaurant and looking at how they operate. No matter what they do or how they do it, I will always see things therein tht are done poorly, improperly, or without regard to the nature of the business in the manner in which I see it. For this reason, I do not dine out very often, and there are exceedingly few call centers that have my number on their dial logs. You, however, do not have that option available to you -- unless, that is, you were to put effort into writing your own program. Which might be fruitful, and then again might not be. The methods by which Poser and D|S work are very different, and the approach that each company has taken to creating them is differnt as well. Neither are likely to change their methodology in the near term. "Common Sense" is a mythical creation that varies between location and individual -- there is no universal common sense. I will refrain (this once, I'm feeling strangely benevolent) from noting the inherent flaws in your programming worldview. As a side note, I am not a fan of tree structures -- they hamper my workflow. Evidence, perhaps, of not being able to please all the people all the time...
Hi Ultrop Your scene setup is interesting as there doesn't seem to be much in the way of morph target overheads so the pmd problem may not be a factor. We haven't as yet run a definitive set of tests for the rendering costs although it may be worth tuning down the maximum texture resolution to 512 or 256 to see if that makes a difference. If you are suffering from general problems loading the scene it may also be a good idea to switch the display mode to Sree3D (SW) rather than the OpenGL display in the General Preferences. There may well be some as yet untested limits in the use of the OGL buffers that could impact performance or even prevent complex scenes from loading. Stewer - interseting image. I hadn't actually thought about using Gather to illuminate the entire scene as a replacement to IBL/AO, rather to pick up colour bleed from nearby objects. Long rays will, of course, increase rendering times due to more intersections being tested across greater distances. Another problem with using long rays is that as the rays spread throughout a scene they will start picking a diverse scatter of different illumination values from different areas of the scene, which may end up causing random artifacts in the final colour. To compensate you could increase the number of rays, but at that point you may as well be using IBL/AO or light probe as the result is going to end up being an averaged out measure of the incident light. The speckling effect in your image is something that I found as well and couldn't really shift except by turning the intensity down, which is probably how it should be used anyway. Bill
Privacy Notice
This site uses cookies to deliver the best experience. Our own cookies make user accounts and other features possible. Third-party cookies are used to display relevant ads and to analyze how Renderosity is used. By using our site, you acknowledge that you have read and understood our Terms of Service, including our Cookie Policy and our Privacy Policy.
Hi, I am not happy. To get it running I had to disable all my other programs before installing and then access each room and close the pop-up help before I could run and access the rooms with all my other programs runing. Makes no sense, but it works. Took several install sessions to figure this out. Now that it is runing, I decided to try my 3D render killer file. It is a large file with 11 diferent characters an numerous simple props as well as 6 more complcated props. I have never been able to render this file in Poser or DAZ studio till the latest version of DAZ studio. I can now import and render it in DAZ studio but not in Poser 6. Not good. In actual fact none of my problimatic files render in Poser 6. Poser 6 has som nice inprovments and they did not try to re-invent the wheel but, they just do not seem to get it yet. The library for instance. It is better but, whether you like DAZ studio's library or not it is more effecient to use in a large library like mine and I do not see how Curios Labs could have missed this one. It would seem that once again we have a program that is very nice to use as long as you do not intend to accomplish anything.