Joe Moya
Forum Replies Created
-
When you export and then import the letters into a single layer within ProAnimator (I assume you mean within PA), will this combining of letters free up memory demands? And, how will this effect the textures and material used for each letter (specially with regards to mov files being used as a texture/material)?
At this point, I have made this project so complex that it appears AE can not render the project no matter what I do (including my old standbye of the “secret” preference change or using Cineform codecs). And, when I say the project does not render… I mean it “fails to render” on the first frame because of a lack of memory. I am almost certain that using 32 bit XP which limits the use of RAM memory to 3G’s is my problem.
………………..
As a side note, my computer is not a slacker…but, it is running 32bit XP with 6G of RAM. I am working with the fastest chipset available 3.2Ghz 7i chip on a dual quad mobo with the fastest 4 x 15k rpm hard drives in a dual stripe Raid 0 setups available (including an very fast Adaptec Raid controller). The downside (because of AE’s limits) is that 32b can only use 3g of executable RAM – max.
………………..I am really struggling to figure out a way to reduce the size of RAM needed to render this HD project… heck, I even tried to render it at SD and have not been able to render the project… EVEN with Cineform codecs. Rendering with uncompressed files worked (albeit very slowly) until I added the shadows…now, I can not render at all – it “fails to render” in AE on the first frame. I have done every trick in he book that I know of to reduce the RAM demands on this project and have failed (i.e., “secret” preference option tweak, min. size background layers, taking opacity to 0 as soon the object leaves the view, not activating layers that can be rendered seperately, not using ANY blurs or color correction, etc.). It appears the RAM demands by ProAnimator “shadows” are making the render fail.
Will combining the individual layers into one layer with on shadow wall reduce the RAM needed to render? And how does this effect the materials (…specially with MOV files currently being used as materials on the letters). How does multi-processing fit into this equation… I currently have AE using max. memory available with most RAM allotted to the single processor being used.
-
Yes… I tried that… and, it works with about 20% of the project.
The problem… and, trying to keep a long story short… is this…I have a pretty complex 3d text “typography-like” setup where that each letter (aprox. 16 letters) is a layer. Applying a shadow catcher to each letter/layer produces a bit of a problem with shadows cast by the letters as a group.
In hind sight, creating the each letter as a layer provided for a lot of flexibility in design, motion and effect… but, as I start to fine tune the details (such as, shadow)… I started to run into problems creating individual shadows for each letter.. and, how the shadows cast for each letter conflicted with the letters as a group.
This individual conflict of each layer of letter’s and their shadow issue is why I wanted to render the entire letter/layers before I added shadows. I was hoping I could create shadows from a combined rendered video with all layers.
At least now I know I need to go back to the drawing board and start solving this problem using shadow catcher and give up on trying to create a faux shadow with a pre-rendered matte.
……….
BTW, I have another issue with regard to rendering texture layers that use a movie file. But… I might save that for another thread if I don’t find a solution… Until then, let me just say… I can not seem to render into ANY compressed video file format (no matter what kind of draft, size or length of video I use) when I use a movie file as a texture. As a reult, I am looking at 16 hr. long render times for a 1 min. 20 sec long HD uncompressed video project (…and, this is without any blur, atmosphere or a few other final touches I know will lengthen the render time even moreso). I haven’t determined if this is a ProAnimator issue or AE… at this point, I am thinking it is AE’s render capability limitations. -
I think I have a solution…
Render as a PNG animation/pictures video file (or any alpha available pic file) and matte the PNG pics to produce shadows… I haven’t tried this – yet… maybe someone has tried this and tell me if it will work before I spend lots of time rendering to a PNG animated pics.
… On second thought, this may NOT work if (as in my case) I am working in 29.97 where frames are matched to specific and precise music cues… my guess (and based on past experience)… 29.97 to a 30 fps PNG videos looses sync with sound as the video progresses…Heavy sigh…
-
If looking for speed… you should probably use a dual Raid 0 setup using the 15k rpm hard drives (…that’s a total of 4 HD’s working with 2 x I/O’s Per Raid) with an Adaptec Raid Controller…. plus, a dual quad (8x) CPU. You’ll see a significant boost with two things: 64bit with the max. amount of RAM you can afford…PLUS… a dual Raid 0 harddrive with really fast raid controllers. The speed bottle neck is your hard drives more so than your cpu/gpu or ram.
A good HD render setup will run you at least 800 US dollars… but, for only about 400 dollars more you can get larger HD’s… which can be a signficant advantage if you are working with HiDef video.
All that being said… the best way to improve the speed of your machine is work with compositing software that is not so much RAM dependent… but rather…uses HD caching. And, that would exclude AE as a compositing application…. at least until CS5 comes out (based on what has been recentlyl released by Adobe)…. which brings me to my next point… if you are trying to get ahead of the curve… your video card will not be the best choice for what is seems to be on the horizon for CS5- you need the 4800).
In short…you have plenty of room to make your suggestions for a “really fast” render machine… specially with regard to your HD setup, CPU speed and video card.
-
Have up ever been working with AE’s camera and ProAnimator and get completely lost and not able to find the 3d object because the camera position is totally out of whack? Well, I have many times and here is a solution I found to help get me back on track.
After MANY hours of struggling with this problem… I think I found a solution to effectively use an AE camera with Proanimator…a solution I have not found noted anywhere. Hopefully others can test it out more thoroughly and possibly avoid some of the major headaches I have always experienced when using ProAnimator with AE camera’s.
How to approximate the center in ProAnimator and use camera’s in AE:
I am assuming you are using Use Comp Camera, Use Comp Lights, Composite On Original (i.e., these boxes are checked in ProAnimator – Checking the …Use Comp Lights and Composite on Original box is optional). AND (…VERY IMPORTANT…), the AE solid layer containing ProAnimator is NOT set to 3D.
First, you need to be using a camera Rig in AE.
I HIGHLY recommend you use Maltaannon’s simple camera rig… is is an easy and seeminly BEST rig to use. A full explanation can be found in detailed at Maltaannon’s site at https://maltaannon.com/. The video tutorial is called Simple Camera Rig.
In this camera rig tutorial, there is a crucial expression calculation that uses with the camera focal point as the tracking benchmark for the calculations in the is RIG… this is VERY useful and important when wanting to use different camera’s and/or video formats. I strongly suggest you look at this tutorial before you read on… take special note of how this rig uses the Camera ZOOM data entry for the rig. Plus, Also you need to have the Maltaannon rig in AE set up and built inside of a comp BEFORE you do the following (note: have camera rig layers on top of Proanimator Solid Layer).
The follow instruction will describe how to properly set a common centered position for a Proanimator 3d object using Maltaannon’s rig in ProAnimator.
In ProAnimator…you should be in scene preview – then, select existing, import or create a primative object in the drop down menu in OBJECT MODE.
Open Object Tab for the 3D primative.
Find the STATS button located in the Object Tab… open the STATS box by clicking the STATS button. You will see the Object Stats for this object… located in this box are the numerical inputs for the XYZ locations for Position, Rotation and Scale.
You will be transferring x,y and zoom data from AE into these boxes. The following instructions will describe which numbers (from AE) to put into these Object Stats inputs boxes.
Side note: Make certain the camera settings in AE is reset to it’s default setting – Camera>Transfor>(click)Reset.
In the camera layer open CAMERA OPTIONS take note of the ZOOM setting in pixels. For my camera setting is was 1866.7 pixels for a 35mm camera (54.4H). You will be using this ZOOM number in Proanimator Stats box – as Position Z input.
Now, take note of the AE CAMERA’s numbers indicated for the Transform>Position for x and y. Do NOT use the Z setting for the camera position if you are using Maltannoon’s camera rig (…his rig adjusts for issues that evolve around the tracking of a camera using the z position).
Side note: MAKE CERTAIN THE ProAnimator SOLID in not in 3D mode (i.e., the small box shaped icon in AE is not checked for this solid layer containing the ProAnimator effect)… as I noted earlier, you need to be in 2D mode in AE for the solid layer containing the ProAnimator effect.
Place the x,y position numbers in the Camera Layer AND the CAMERA OPTIONS ZOOM numbers in AE into the OBJECT STATS window…
…where X being the position number taken from AE for Proanimator’s “X” Object stats position X 1.5,
…where MINUS (very important to put a minus sign) “Y” position mumber taken from AE Camer Position for Proanimator’s Y Object Object Stat Position X 2.
…and (finally) “Z” is AE’s CAMERA “ZOOM” number taken from AE camera Zoom (i.e., Z = Camera Zoom not camera Z positon). STATS Rotation settings should be all 0’s and Scale should be all 1’s.Now apply these transferred X Y and Z numbers to ProAnimator. And, the object is centered OR ALMOST CENTERED… The multiplier numbers of 1.5 and 2 are only estimated center adjustment numbers. You will need to change the Object Stats to fit the size of comp setting you are using. I have found that the smaller the aspect ratio of the comp, the larger these adjustment numbers need to be… But,… the key is keeping the Z setting to the Camera Zoom setting the same and only adjust the X and Y position setting.
Shortened/Cliff Note version of aformentioned instructions below:
1)Build and Use Maltaannon’s camera rig in AE. (see video tutorial)
2)Add solid and place Proanimator into new layer.
MAKE SURE THAT THIS PROANIMATOR LAYER DOES NOT HAVE THE 3D BOX CHECKED… YOU NEED TO WORK IN 2D MODE IN ORDER TO MAKE 3D CAMERA MOVEMENT WORK WITH PROANIMATOR (and, yes that is strange)
3)Import, create or select existing 3D object in Proanimator..
4)Select 3d object in Proanimator.
5)Open Object Tab for the 3D primative.
6)Find the STATS button located in the Object Tab and open.
7)Take the following information in AE and place these numbers into ProAnimator’s STATS window (to be safe, reset camera defaults in AE):STATS X Position = AE’s Camera Transform X Position x 1.5
STATS Y Position = – AE’s Camera Y Transform Position (add minus sign) x 2.
STATS Z Position = AE’s Camera Options ZOOM (in pixels)
(Note: the 1.5 and 2 multipliers are approximate multipliers and the X and Y positions need to be changed to match your specific comp size – small aspect ratio numbers for the comp requires larger adjustment numbers). The key is keeping Z position constant.Once the center is fine tuned in Proanimator, Click OK in ProAnimator… now you are free to move the camera using AE’s camera rig.
Final Side note: You can only have one camera in AE when using this rig in conjunction with Proanimator… for some reason, when you have more than one camera… it sometimes confuses the camera rig and you get weird camera movements and you can screw up your common position between Proanimator and AE.
IMPORTANT FINAL NOTE: If for some reason you create a second camera or your center gets lost between AE and ProAnimator (…this has happen to me). Or, decide to change the comp size (which is a big problem for camera movements). Go back to Proanimators Object Stats Box and change the XY positions to equal the size of your comp. … Where X = comps X bit size… Where Y = MINUS comp bit size… and, Where Z still equals the Camera ZOOM number in AE. Ex. – 1440×1080 size comp would have a ProAnimator Stats numbers of – X = 1440; Y = – (minus) 1080 and Z = 1866.7 (for 35mm camera) This should give you a decent starting point to find the 3d object in the comp. No matter the problem… The key is keeping the Z positon constant (i.e.= AE’s Camera Option Zoom) and move the X and Y positions to center the 3d object.
*******************************************************************
I have not tried this out using a lot of different camera’s or comp sizes… but, so far it has worked well for different camera movements and provides a close appoximation of center for most comp sizes.
Hopefully others can experiement with my solution and can find faults with my simplistic way of using AE camera’s and proanimators 3d objects.
FWIW
Joe Moya
-
Glad I could help…
… I don’t know what your original footage is… but, if it is coming from a m2t or m2v type codec… you might try and change the project/template settings (I think it called that)to where the project in Liquid is read as a YUV file.
What will happen is that no yellow bar (or slices in Liquid language) will appear and require a render to green. Instead, it will show as gray (i.e., indicating it is operating as a native format)… this means you don’t even need to render the video file. The only portions that will need to be rendered is any portions of the video you add an effect/filter. And, this rendering runs in the background all the time while you edit AND automatically saves any changes you make every 1/20th of a second… you can NEVER loose any work. Those two things can be a huge time saver for long editing projects.
Liquid is (for the price) one of the best editing software there is… to bad it is End of Line application… but, AVID plans on introducing an entirely new Liquid. If so, they have really big shoes to fill if they plan on trying to match what Liquid has done for about 5 plus years…many editing programs have only in the past 1 or 2 been able to match anything close to what Liquid has been doing for over half a decade (…and, none have been able to match the autosave Liquid uses)… and, doing it on a super stable platform and about 33% of the cost of AVID’s MC…it’s no wonder AVID bought it out. If you can’t beat your competition… buy them out… then, (as I predict) will re-release it as a new dumbed down version at a much higher cost.
-
I am not at my computer… so, I am working from memory… bear with me.
Where I am able to import the finished graphics when loaded onto the timeline the graphic only fills part of the screen as if the graphics is incorrectly set up in AE I have tryed a number of the options when setting up the composition settings and also again several options on the Export settings all with varing degrees of success but still with the overall result of a smaller graphic than required being seen on the Liquid timeline.
I noticed you are using 16:9 as your size setup… this might be the problem… Check the property settings for the video clip transferred into Liquid. Are the setting for the clip and/or entire project in Liquid 16.9 with 1.33.
A lot of times AE exported file gets convert to a 1.0 square pixel frm the 1.33 pixel format. All you have to do is change from the drop down menu that appears in Liquids Video Clip menu (not in AE)… you will find this drop down menu if you right click on the imported clip icon or clip dropped into a track.
Also, in Liquid… how is the video clip settings set for proportions and scaling. In liquid, you have a simple drop down menu that lets you pick your scaling options. This is a great little solution that AE doesn’t have. As a result, a lot of the short comings of AE can be easily overcome by Liquid if you just know where to look.
Next, what are your Liquid setting for you entire project? Sometimes those working in native file formats that don’t jive well with AE. In general, the best native format is the YUV file format if you are working in HD (not MPEG2 m2v). Liquid does a Fabulous job of rendering and smoothly using HD files and is a dream to use compared to After Effects or Premiere Pro. In fact, in the YUV format, there is NO rendering required in Liquid…it works natively… and takes 4:2:0 and converts it to 4:2:2 like Cineform (but is a freebe). Without knowing what is composited, this native yuv and 4:2:2 bonus may be advantages (or not).
If that is not the problem… then it could be how it is being exported from AE. In AE, make sure the setting of the frame size matches Liquids for best quality conversion. If you are using 1440×1080 in Liquid, then use the same in the AE compositon for best results. If not, then you will want to use Liquids resizing of the video clip scale in Liquid as your best option for making the video clip scaling size work between the two applications. However, doing this may result in loss of clarity and quality of the video.
The best generic export file type is AVI uncompressed windows media. For HD it is very slow to render in AE… but, it definately yields the best results and works natively in Liquid. The second best choice is Quicktime MOV uncompressed (not animation). These typically yield the best level of quality and have the least amount of trouble when it comes to complex AE composites. Oh… and, no OpenGL will also show best results and most stable render… but, without OpenGL the render time can be glacial. I would try with and without OpenGL…if you get the same results go with OpenGL to speed up render time.
The basic strategy to solving this problem is setting up the correct Liquids settings first… then, exporting from AE in a format that matches. Based on your post, I think you have a little bit of a problem on both sides of the equation – AE and Liquid. Which makes it pretty difficult to find the right combination of export and export.
There are tons of variation of options… but, I am not at my computer at this time and can’t look up some notes I keep with regard to AE to Liquid renders/transfers. If you could post a pic of how the import into Liquid looks, that would be VERY helpful.
-
Crap!!!… I have GOT to get a new pair of glasses…
I just realized that I consitently made the mistake in my posts of reference PSP instead of what I meant… I meant PSD … NOT PSP… ugh! Frustration is getting the best of me… for sure.
The animated character was created in Adobe’s Photoshop… The animation was done on a Photoshop file in AE using the Puppet Tool.
I have been lost for the past 5 days trying to figure this thing out… but through LOTS of trial and error… it appears to be a combination “OpenGL and RAM Memory Limits” when dealing with HD qualilty with AE. It was an issue that had a ton of various potential solutions once the problem was identified (which is no simple task by itself).
IMHO, this sort of problem is caused by AE’s inability to effectively use and handle Hard Drive cached memory in lieu of RAM memory. And, with large Hi. Def. video files and frame sizes this becomes a bigger issue once the project starts to grow in length.
The good news is that I fixed it and the cause was isolated to exactly what I have previously posted as a solution… or else, the problem would still exist.
Lesson learned… be very careful when building long projects that use HD size frames when using the Puppet Tool. If anything, these series of posts may be useful to some other poor soul who unwittingly delves into the dark world of HD 2D animation using Adobe’s Puppet Tool in CS4.
-
Well… if I am using temporal footage… then, After Effects created it…
This is a 100% AE created footage. The source is a PSP file… that’s it. ALL animation is created with the Puppet Tool and Camera movements… there is NO other source of animation in this project (temporal or othewise).
The slow render or failure to render is caused by these effect elements:
1) top on the list… and major cause… Puppet Tool.
2) followed by: Shadows, Color Correction, Lights, and Camera.The slow render time and poor quality of results is caused by OpenGL in combination with the Puppet Tool.
This loss of quality could be caused by the video source… but, if it is it… then it was created by Adobe since there is no external animation used except what is done by AE. Somehow I doubt this, since I have a perfectly nice HD quality output when I eliminated the OpenGL in combination with Puppet Tool…OpenGL works nicely until it hit the frame where the Puppet Tool effect is applied. To me, this rules out temporal compressed footage (…much less the fact there is NO externally created footage used).
I am not looking for instantanious anything… but, I certainly am not expecting total render failures or loss of quality in video either… which is what happens when I try to do the project in a simple and typical AE work flow. A LOT of juggling around of this project has been done just to get it to render to a final format in HD quality.
The final result seems to be this… Puppet Tool dumb downs the video quality when rendering approaches max. memory useage during render (..in fact, the actual error message during render indicates the fault to be blur… but no blurs are used… nor DOF).
The way to over come this render failure and/or poor quality render results is to fix the compression settings as “Full” (or “Current Settings” if Full is chosen in the project)… “Best” settings does not work because… (I’m guessing) is that it defaults to an “Adaptive” resolution/compression to save memory. The “Adaptive/Dumb Downed” resolution reduces the videos quality (and, not a by product of the zooming of camera).
Of course, this “Full” resolution typically results in render failures because of low memory. The solution this problem is to go to the Secret Pref. and purge the video frames. The net results is a nice looking video that takes a very long time to render. I can handle long renders, but failure to render is a bit hard to deal with…
More RAM would be great, except for one big limitation. The max. amount of RAM that can be used with a 32bit XP is 3G’s… I currently have 6G of low latency RAM in Dual Quad 7i 3.2Ghz chip in an ASUS mobo using Adaptec Raid Controllers with Dual RAID 0 setups with 15k rpm Fujitsu e-sata HD’s. Just short of a full fledge Render Farm… you can’t get a faster CPU, Raid Controller and HD combination than what I use. 64 bit will take care of the RAM issue in the future…but, until then… everyone is stuck with 32bit with 3G limit of RAM. AE biggest limitation is that it very RAM dependent. And, with HD size frames this poses big problems.
As for solving the long render time problem with the RAM-use intensive Puppet Tool is not truely solveable unless you want crappy looking results that is created when using OpenGL. I think Adobe has a lot of issues to work out with regard to HD stuff… and, quite honestly other compositing software can do a marginally better job dealing with HD assuming you have the hardware to work with HD source media. Hopefully, CS5 may have a better handle on HD size file (or, not… lot depends upon how the competitors deal with the same 64 bit change).
-
Nothing personal… but, I think the phrase is this…”don’t pee on me and tell me it’s raining”….
…the problem isn’t that everyone knows about Adobe’s decision to no longer support 32 bit… the problem eventually comes down to simple economics. From Adobe’s standpoint the cost of conversion is being placed squarely on the backs of the customer and the third party plug-in developers.
It is not a question of whether or not the third party plug-in developers know about the is 64bit conversion ONLY road Adobe has choosen to go down… but, rather the question should be asked, “Do you know of any specific developers that are planning on creating a 64-bit compatible version of their plug-ins for free?” I doubt it (nor do I expect it)… and, so the fundamental issue is still the same. The cost of this lack of 32 bit compatability is being transferred from Adobe to it’s customers and developers… end of story… no need to sugar coat this fact. I don’t see how it is difficult to communicate this simple fact (even in a short post on a AE forum).
All this being said, I am happy to see Adobe communicate this key fact beforehand. It gives me and it’s customer’s the opportunity to evaluate my next upgrade in system applications. Which in code means this… since I am going to fork over a significant additional cost per station (because of lack of 32 bit compatibility). Then, the question becomes, “Why not just scrap Adobe applications all together and go with another compositing software?”
In the end, this decision will come when we look at stability of the applications and it’s relative costs. In that respect, 64bit only application may save Adobe important development costs but may at the same time reduce revenue.
When I see these sort of decisions come into play… I can almost always find the accountants (or, the accountant mindset) running the business instead of the sales and marketing departments. Such decisions tend to be simple and shortsighted. It is always easier to cut costs than to sale more. It also tends to have negative implications to it’s image and ability to market the products that keep the revenue side with a positive growth rate. At that point, enter the spin-doctors of marketing…. and, the resulting “kick me” sign pinned on the back of the Adobe customer.
As for the band-aid… exactly who’s skin is being ripped off? Adobe customers already know who is doing the ripping and who will be doing the screaming. I really wish Adobe would at least kiss the boo boo with a significant cost reduction compensation or rebate system for upgrading plug-ins to a 64bit application.
So,… Adobe… why not implement a “proof of purchase” rebate system for those who have to upgrade to 64bit plug-ins and prove you do have the customer’s thoughts in mind (…but, shhh… just don’t let the accountants find out)?