Flatshot issues in AutoCAD 2015

[alvespereira]

To my surprise the FLATSHOT command is no longer working in ACAD 2015. If I try to project a view i«which is in PERSPECTIVE=1 all I get are rubbish results. This used to work flawlessly in 2014 and it became a important part in my workflow. Combining linework with renderings and post-producing in Photoshop. There are a couple of other thread addressing this issue but none was given an acceptable  answer/solution. (VIEWBASE is not an alternative).

I find it deeply disturbing to find a good tool (altough lacking some features) being trashed like this. Any help on the subject would be greatly appreciated.

Thank you

[Alfred.NESWADBA]

Hi,

>> This used to work flawlessly in 2014 

Do you have a drawing that worked with 2014 for us which shows what 2014 and 2015 created as results, just to compare results you get with the ones we get?

Because when I look to the help about the command FLATSHOT I read:

     The flatshot process works only in model space. Start by setting up the

     view you want, including orthographic or parallel views.

Source of the help: >>>for 2014<<< and for >>>2015<<< at least the same content.

That means (for me) ... not perspective, am I wrong?

- alfred -

[alvespereira]

Hi Alfred,

Thank you for your interest. I attach two files: one with a working example of flatshot used with perspective=1, made in a previous version of AutoCAD. The other contains the result obtained in 2015. The filenames are self explanatory.

About the help content, "including orthographic and parallel" doesn't explicitly exclude perspective. Again thank you.

Paulo

[dbroad]

alvespereira,

Thanks for posting the files.  I am surprised to see the difference.  I haven't seen the same problems here but don't usually use 2015.  I can verify that the difference in behaviors exists here.

You should report this behavior as a bug to Autodesk.

[nestly2]

Unfortunately, FLATSHOT does appear not to work as well in 2015 as it did previously  (and not just for PERSPECTIVE View either)

I don't think there's anything that can be done to improve it on our end, but If you're on Subscription submit a support request, otherwise file a Bug Report. and hopefully it will fixed in a future release or update.

[SEANT61]

It seems to work better with units set as meters.

[nestly2]

That's a very interesting observation....  especially since the drawing was based on an Imperial template, (-DWGUNITS=1, MEASUREMENT= 0, &  acad.lin is the Linetype file)

[SEANT61]

Those anomolies notwithstanding, the problem relates to the 15 significant digits, and AutoCAD truncating critical information out there at the fringe.  That's my theory, anyway.

[Alfred.NESWADBA]

Hi,

sorry for my late response, and thank you for sharing the drawings. Sorry to say: yes, I can reproduce it too.

At least I would follow the suggestion of @dbroad and report that as defect (in your subcription center).

I don't really believe on units (for this issue here) as the problem would be the same in 2014 then (with the same drawing/same object/same position/same size).

But to draw something in meter instead of mm and so get 3 decimal places more (or to scale the object down) and closer to 0,0,0 to get more digits reserved for geometrical calculation is a good suggestion.

- alfred -

[SEANT61]

In the interest of clarity, I, too, believe that an algorithm change between 2014 and 2015 is responsible for the different levels of performance. Some process of that new algorithm, a math calculation, some tolerance setting, suffers a double precision floating point operation mishap. I say that specifically because Flatshot does works better at more sensible unit settings.

With regard to CAD operations, and AutoCAD in particular, the logical ‘bug’ is the notion that millimeters should be used at architectural scale. Sure, it saves the users the hassle of a inputting radix points, but it cause numerous problems as math calculations waste significant digits needlessly.

[alvespereira · Accepted solution]

Dear participants,

Strangely enough I was able to obtain a perfect FLATSHOT in perspective from the model attached in my previous post.

Following SEANT's suggestion of a rounding error somewhere along the calculation process, I fooled around with different sizes and placements, and voilá, out of the blue, a perfect FLATSHOT. I believed at the time that the problem was the unit (m) and hence dimensions with a small integer part, and scaling it by 1000 (mm) would solve the issue everytime; Sadly no. I tried it in another computer, with a model in mm and got the same very crooked results. 

I am now in the process of remembering all the steps I took before getting FLATSHOT to work again. Nevertheless it's not working how it should and I will report it to Autodesk as suggested by Nestly2 and DBroad.

Thank you all for your help,

P

(sorry for the broken english)

[negativenine]

I have found the FLATSHOT works fine in perspective if UNITS is set to the lowest PRECISION settings,Length Precision = 0'0" / Angle Precision =0.

I must be a bad rounding issue, not sure.  I have not played with it to see exactly where the precision would begin to cause errors though.

This of course I figured out after reading through this thread.

[negativenine]

Okay, doesn't seem to work every time. Got one good FLATSHOT out and can not seem to repeat this.

[binayashahi]

I have the same problem with autocad 2015 and 2017. I haven't tried out in 2016. It works fine in autocad 2014. Autocad 2015 and 2017 displays error message- "Flatshot operation did not produce any geometry. It happened in the 3d models with non uniform scaling. I tried flatshot on simple cube. It worked in autocad 2017. So, the problem is with the non uniformly scaled 3d blocks. However, flatshot works with non-uniformly scaled 3d blocks in autocad 2014 but produce flatshot of original size( not the scaled one).

[alvespereira]

Unfortunately the issue still persists. All versions since 2015 are affected by this bug. Reports have been sent to Autodesk to no avail.

[SEANT61]

I can understand, somewhat, that FlatShot has problems with non-uniformly scaled blocks.  That elevates the level of complexity enough to assume more time is needed to hammer out all the details.  Likewise, unusually large/small scale factors tend to play havoc with numerous AutoCAD operations.  Those factors play havoc with math computations in all digital endeavors.

My assumption, though, was that 3D solid based drawings can always produce good/clean Flatshots given sensible scale and placement.  Is that not true?  Do you have an example of 3D solid drawings that refuces to cooperate with FlatShot?    

[zhtao]

This case (Flatshot_notworking.dwg) is fixed in AutoCAD 2017.

[wynand]

Hi, been following this hoping for a solution. I have flathots that was generated, but now seem to not want to update:

"FLATSHOT
Select block to replace:Flatshot operation did not generate any geometry."

The fact that it's solved in other versions is irrelevant, I don't have those versions. How can it work one moment and the next not?

[alvespereira]

Sorry Tao Zheng but it is not. The issue persists. When creating a flatshot from a viewpoint with perspective=1 the results are unexpected (and wrong). Several reports have been made to Autodesk. If you want i can screencast. Thank you

[zhtao]

There are several flatshot bugs are introduced in AutoCAD 2015. The bug drawing Flatshot_notworking.dwg is attached in this post months ago and this one is due to arc segments in the drawing and already fixed in AutoCAD 2017.

And one other bug is due to blocks contain 3D solod, the fix is now reproducible and the fix is on going.