DAZ Speed - Any plans on improving this?
jdavison67
Posts: 650
Playing in Unreal 5, then moving back to DAZ, I find myself less tolerant of waiting 30 mins for a figure restore function to complete, or the 10 seconds it takes for DAZ to let me move from one body part to the next, just to pose it..
Is there anything in the pipeline to make these actions happen any faster?
JD
Post edited by jdavison67 on
Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
I would wait until we get closer to DS5's development, (I'd say DS4.75) that way we'll have a better idea of the overall development of DS5, and it's future within the 3D industry.
In the meanwhile, I would cut the character/scene's rez to base 0
I honestly don't ever see DS getting as fast as unreal or unity, they are just two seperate animals. I feel they would have to totally redo Daz Studio and base it's core on a game engine and then find a way to add all the features in DS that make everything work like they do in DS already
I am kinda in the same boat though. Been using a VR app (based on Unity) that allows for the creation, posing and animation of DAZ G2 figures in real time and life size and I gotta admit I am spoiled by it and wish DS had the same features.
I have a ton of morphs and it takes forever to load figures. Feels like the more content you buy, the more you get punished with a slower gui. It can really wear thin at times and breaks the artistic workflow (all the time I have to kill while something is loading). I really hope they focus on optimizations, but as long as things are selling well I'm not optimistic. The one hope I have is if more and more customers are using the Unreal and Unity bridges, they'll be forced to improve perforamnce or risk losing those customers.
As far as the morph problem, I have several products that help with this:
https://www.daz3d.com/turbo-loader-for-genesis-8-and-81
https://www.daz3d.com/turbo-loader-for-genesis-3
https://www.daz3d.com/turbo-loader-for-genesis-and-genesis-2
https://www.daz3d.com/turbo-loader-booster-utilities
I don't see how that is hope, if anything it makes it worse. those are game engines, very different from what DS is and to compare them is pointless IMO. They won't lose customers since even if a user decided to do everything in unreal or unity, they still need the content from DAZ. They just now have to jump thru more hoops getting content into a 3rd party app instead of a native one.
With all due respect and appreciation for your skills and efforts (I have a bunch of your scripts), it really shouldn't be necessary to have to buy something to solve a problem that should be addressed during DAZ Studio development and testing. But that seems to be the way DAZ works - don't worry about the design shortcomings, just let some PA come up with a work-around that we can then sell.
The "hope" is that super slow DS + fast Unreal/Unity will convince DAZ they need to improve performance vs. lose customers fed up with the slow performance (now that they are exposed to other store fronts/ecosystems). Or that to continue to grow the DS user base, they need to improve performance because coming in from a fast engine and trying DS out for the first time can't be a great experience.
Far too much truth in this, and it's going to make a DS 5.0 migration a real pain -- add-ons for 4.x that won't be available and repurchasing DS5 versions for ones that are. It's a terrible way to develop a product. No control over a consistent user base, and a bunch of features that the DS team cannot use because not all users have those features.
This is true. The more characters and morphs you install, the more burdensome your G8 figures become on your memory recources. My G8f became so impracticle, that I began converting my figures to props in order to compensate for the memory bloat. When that became tiresome, I started creating more content directories to try and compensate even more.
Do yourself a favor and purchase, at least, https://www.daz3d.com/turbo-loader-for-genesis-8-and-81 (or whichever, depending on your figure choice). It disables all the morphs and characters that you dont need to load resulting in far better memory costs than just loading your default g8.
@marble, yes you are correct. We should NOT have to purchase something that daz3d should have included as part of its native app. But how many times have we said that before about so many more missing daz3d features?
I swear by this product. It is worth the purchase, full price or not. Especially if you have large collections of characters and morphs.
Yes, this thread raises a number of speed/performance issues with current versions of Daz Studio, as well as a number of questions about the future of Daz's apps and content.
1. I'm sure Daz thinks about problems of lag and sluggishness, and they want to balance the desire to have Daz Studio be a flexible, near-state-of-the-art application with the necessity to have Daz Studio function in a time-efficient way. To me, one of the surest ways to get consumers to buy more content is to speed up the whole process of using Daz Studio, from opening the app to finishing renders. Waiting while the app is unresponsive isn't fun or productive. Why buy more content if you haven't made use of what you already own? Smoother operation of the app itself, improvements to the clothing fit systems without the need to wait for a simulation stage, and faster rendering are probably priorities. We can't be sure because Daz is unusually reluctant to discuss future plans.
2. Given how frequent the complaints are about the slow load time of Genesis 8, I'm sure Daz is planning on making changes to how characters are loaded into the scene, and what options are available to the user without taking special steps. The current loading system is a little frustrating and actively discourages buying additional figure models because of their negative effect on performance. Most Daz users are hobbyists, and they will find other things to do if the frustration level is high. They'll buy more content if they're finishing projects quickly with results they find satisfying.
3. The animation/game engine users vs. stills artists issue. It's a basic question that needs to be resolved at some point. Is Daz planning a completely different app for animation & action game use? Daz hasn't given us a clue about their thinking on the topic.
4. Hoping or relying on the end user buying a high spec machine to solve software performance issues isn't a viable strategy in the 2020s.
Yup, been using blender, UE4/5 and iclone8, Daz is the slowest of all. on the latest i9 CPU, M.2 SSD and RTX3090, it takes 5 mins to load a scene, 2-4 mins for a single character, and even longer to rig when a character have dforced cloth, hair and heavy morph, doing animation is not unsustainable with the current pipeline.
Hope that changes with the next big update
edit: mod please delete other 2 posts, I think the forum forzen when I clicked post comment, so I clicked 3 times, my bad
That has always been the way DAZ works. A PA will come up with something and quite often it either gets bought out and re-labeled as a DAZ Original product, or something extremely similar will come out as an official DAZ version for the next generation. There have also been features that have only continued to work properly because some PA or Freebie artist took the time to come up with a workaround for when DAZ dropped the ball, as when DAZ stopped creating new templates for Puppeteer for three whole generations of products. In the end, PAs tend to sit down and anticipate needs to create products that either earn them money or fill a use they've found themselves, whereas DAZ makes it's money by brokering and selling accessories, not off sales of DS itself.
Thanks for starting this thread JDavision.
Rendering speed is the number one improvement Daz could make to Studio. Real time rendering should be the target. No, we're not going to get there immediately, but it should be the target. Focus on tear duct geometry instead? C'mon now. Speed is the reason why I'm using an older version of Studio. The newer ones are slower.
Yes, game engines are different, but they are converging. Some can handle high density models with millions of polys and in real time automatically optimize the mesh depending on distance from the viewer. Real time ray tracing has gone from a novelty that kills your frame rate to an expected feature for certain situations, with steady advancement towards full scene ray tracing. Physically based materials have been around for a while and continue to be improved.
We are not far from "game engines" that can render the same quality that Daz Studio can, but in real time. Somebody is going to spot the market opportunity to merge these two streams. If not Daz, then somebody else. Look at the example FSMCDesigns mentioned, the VR app using G2 figures. That's a start.
This is not a new discussion. The problem is that what Daz Studio needs to do - be able to load additional morphs and properties with links to others properties - requires a lot of processing as the number of properties grows, solutions like Zev0's (or just uninstalling products you are not using in the current project - the product is a convenience, not the only solution) can be used to control the load. There is a tradeoff between flexibility and performance, no one can do anything to eliminate this completely and the solution currently used has worked well over four geenrations of figuers, but it is stressed by large product sets (Genesis 3 Female had similar, though less severe, issues towards the end of its shorter product cycle).
There are some wonderful posts in this thread.
I'm encouraged that there is a lot of concentration on this issue, and I am hopeful that a sollution presents itself.
I have come to love DAZ for all it's strengths.
I just wish for it to overcome it's weakness, as well.
I will continue to create, using whatever tools are available, and hope for the best.
Thanks for the amazing feedback!
JD
One of the big problems with the large number of morphs is that the whole process is single threaded. This was probably fine back when the code for loading morphs was written for Genesis 1 and they probably never thought that people would have thousands of morphs including all the hidden JCMs and expressions. Now this design needs a total revamp for the modern era where users have large content libraries and multi core CPUs. It's so depressing watching Daz Studio stuck on the UI thread using 6% of my total processing power because of this ancient design. A better design would give the user the option to load Genesis figure in asynchronous mode where the base figure instantly loads and the UI remains usable and the morphs load in a background thread. This way a user could begin to move and pose, dress, and set materials on the figure while the morphs load in the background. Once the morphs finish loading, then it could apply active morphs to the figure and update the viewport.
Given the nature of the taks I am far from sure it could be multi-threaded, but I would hesitate to act as a back-seat driver over something about which I know very little of the fine detail.
I don't pretend to know how all of this works and I am not a programmer so I can't suggest solutions but I can compare - perhaps not exactly like-for-like but close. I have been playing with diffeomorphic to transfer G8F including all morphs (shape keys) to Blender. Blender does not seem to have a problem loading thousands of shape keys. I think a similar point was made early in the this thread. Making excuses for the shortcomings of DAZ Studio does not encourage users to stick with it. Perhaps that's the reason things like Diffeomorphic exist? Perhaps DAZ have given up on development and are encouraging users to export to other applications by providing their own bridge add-ons?
The issue isn't the number of morphs that requires the time, it's the links between them - does Diffeomorphic preserve those?
You'd have to ask the developer. As far as I understand, the purpose of the shape keys in Blender is to give a similar experience to that of being able to adjust the morphs using sliders as we do in DAZ Studio. If you look at the Diffeo web site you will see that the export includes JCMs as well as shape morphs but I don't know if that is what you are talking about.
Maybe DAZ should just make DAZ Studio Open Source and give it to the Blender Development team.
Some recent developments:
--Filament. Not sure how many are using it. I'm not seeing a lot of finished renders.
--MetaMixer. When I see Daz renders, I see far more Daz users are customizing characters with non-character specific sliders than with hybridization. I suspect that some users think it's the best thing since sliced bread, but I have little interest.
--Volumetrics. Of little use to me, but it's a real thing.
They are developing features. We'd all like our particular issues and problems catered to, but my individual wants and needs won't coincide with Daz priorities much of the time. And the rarity of useful communications about strategic vision and planning can be frustrating for Daz users.
Bridges/converters are probably intended for a real, but nonetheless limited, fraction of the customer base. I don't see any reason to believe that Daz plans an about-face in regards to being a content provider without its own proprietary app. We will see a Daz Studio 5.0.
What Richard means by "links" are the 1600+ formula you find in each character morph set, they are also known as ERC links, in DS4 we use them to adjust the position of the bones so they work "better" with the morphed shape.
Odds are those ERC links aren't being exported from DS.
Yeah! It´s more than a year now since we´re waiting for it.
I know things take time and that´s ok. But we haven´t yet seen any screenshot or at least a list of planned features. Even a simple information if it´s still beeing worked on would calm down people.
We all love DAZ Studio for it´s easy to use features but as long as we see literally every other 3d application on the marked beeing faster in almost every aspect, the questions beeing asked in this thread are absolutely reasonable.
Explanations why this is the case don´t solve the problem for anyone unfortunately.
If multiple characters are being loaded, I don't see why they couldn't all be processed on separate threads. If multiple clothing items are being fitted to a character, I don't see why they couldn't be processed on separate threads. I don't see why multiple dForce simulations couldn't be run on separate threads. All of these things appear to be independent operations. All of these things happen before the rendering starts, and before the rendering of the preview window starts.
I am pretty sure I saw a comment from a DAZ person suggesting that DAZ 5 is a long way off - at least a year, maybe several. In the meantime there are issues that DAZ Studio lovers (and I guess that includes me as I play with it every day - it is the most-used toy I've ever had), issues that are soooo frustrating. The slowness of loading and closing scenes, the slowness of dForce simulations, the slowness of rendering (and I am fortunate enough to have a 3090 but I would not like to try to render a 1 minute animation), the lack of a truly functional animation toolset that is common in most other 3D applications ... the list goes on.
Filament is a half-baked solution that could perhaps be useful if DAZ had included the all the features. As it is, it is of little use. I use it for making sure my characters' feet are on the ground and not through it but little else.
No idea what MataMixer is for. My main pleasure from DAZ Studio is in creating my own characters from the morph dials. If there is anything approaching the artistic in what I do it is that.
I don't have much use for Volumetrics but I thought that it had been possible for years. Perhaps I'm wrong.
I hope I'm wrong about the potential wait for DAZ Studio 5 but slow seems to be the operative word when it comes to DAZ.
You're quite right, developing a certain feature to add onto Daz Studio may be the greatest thing ever for some users, while unneeded and ignored by others. But working to improve the speed of Studio will benefit *ALL* users. That's why I'm glad JDavison made it the focus of this thread.
That's pretty much the typical life cycle of software. It starts off with certain technical decisions that made perfect sense at that time. Then as improvements are made over the years, features are added, or user expectations change, that framework gets steadily more creaky.
At some point the hard decision has to be made for a major ground-up overhaul even if it breaks a lot of backward compatibility (one of Daz's strengths) often annoying customers who have large content investments, or risk a competitor coming in and taking over. We can all think of companies who dominated a certain market and had huge mind-share, but who failed to keep up and are now just footnotes.
Thank you for posting this comment.
I stopped shopping and purchasing Daz content about 21 months ago because the load time problem was so bad, and purchasing more content just made it worse. After a few months of being away, I posted this:
https://www.daz3d.com/forums/discussion/478476/just-in-time-morph-loading
and don't think I'd been back shopping or purchasing since.
So, these Turbo Loaders are going to be my first DAZ purchases in a long time. I know they are not the solution in code that most people want, but the DAZ experience has been unbearable if you own a lot of content, and it looks like these Turbo Loaders can help.
I am glad I can help. It would be great if you can share your experience (before and after times) in the official thread: https://www.daz3d.com/forums/discussion/557186/released-turbo-loader-for-genesis-8-and-8-1-commercial#latest
Perhaps DAZ Speed is like Light speed, a fundamental constant of the Universe, but slower.
This would not help Daz as commercial enterprise at all..
And there are already several external Character solutions for blender along with diffeomorphic.( HUMGEN CC4 pipeline, Autorig pro etc)
We prefer the Blender foundation focus on needed Blender improvements such as a usable animated hair system and improved cloth simulation, instead of trying to recreate Daz studio inside of blender.
Hmmmm....
I gotta wonder though. I would have hoped that speeding up the morph loading solution and/or moving to a multi-thread approach would be a top priority for Daz, but I am getting more skeptical....
I would like to think that when RiverSoftArt approached DAZ with his pitch for the Turboloader suite, that Daz would have told him... "We don't feel that that would be a good use of your time and effort, because DAZ 5.0 is going to solve all those problems and your product will be redundant in 11 months..."
Maybe that conversation happened...RiverSoft?
But it probably didn't, so either DAZ 5 fixes these issues and Riversoft will have an unneccessary product, or DAZ 5 Doesn't fix these issues at all.... No offense to RiverSoft, but I am hoping that it is the former.