enblend crahes without error message
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Enblend |
New
|
Undecided
|
Unassigned |
Bug Description
When blending a vertical panorama with hugin 2020.0, enblend crashes in Windows 10.
Here is the (German) log:
=======
Panorama zusammenfügen...
=======
Plattform: Windows 10 (build 19043), 64-bit edition
Version: 2020.0.
Aktuelles Verzeichnis: D:\Work\
Ausgabe-Präfix: 1040122-1040128
Überblendung: enblend 4.3-6b604e79e85b
ExifTool-Version: 12.08
Anzahl der aktiven Bilder: 7
Ausgabe-
Rahmengrösse: 4599x10224
ROI: (443, 254) - (4131, 9392)
FOV: 28x58
Projektion: Gradlinig (Rectilinear)(0)
Verwende GPU für Umberechnung: falsch
Panorama-Ausgabe:
* Mit Belichtungskorr
Erstes Quell-Bild
Nummer: 0
Dateiname: D:\Work\
Größe: 3712x2784
Projektion: Gradlinig (Rectilinear)
Kamerakurve: Benutzerdefiniert (EMoR)
HFOV: 23
Belichtungswert: 11,0
Umberechnung von LDR-Bildern…
Multiple images output
loading _1040122.jpg
remapping _1040122.jpg
saving 1040122-
loading _1040123.jpg
remapping _1040123.jpg
saving 1040122-
loading _1040124.jpg
remapping _1040124.jpg
saving 1040122-
loading _1040125.jpg
remapping _1040125.jpg
saving 1040122-
loading _1040126.jpg
remapping _1040126.jpg
saving 1040122-
loading _1040127.jpg
remapping _1040127.jpg
saving 1040122-
loading _1040128.jpg
remapping _1040128.jpg
saving 1040122-
Überblendung der Bilder…
---
In Windows Event Viewer I have these details (XML view):
<Event xmlns="http://
<System>
<Provider Name="Application Error" />
<EventID Qualifiers=
<Version>
<Level>2</Level>
<Task>100</Task>
<Opcode>
<Keywords>
<TimeCreated SystemTime=
<EventRecordI
<Correlation />
<Execution ProcessID="0" ThreadID="0" />
<Channel>
<Computer>
<Security />
</System>
<EventData>
<Data>
<Data>
<Data>
<Data>
<Data>
<Data>
<Data>
<Data>
<Data>2960</Data>
<Data>
<Data>C:\Program Files\Hugin\
<Data>C:\Program Files\Hugin\
<Data>
<Data />
<Data />
</EventData>
</Event>
Also, it seems Hugin does not detect the crash, because the temporary files are removed after that, and the resulting JPG has size zero.
The bug seems rather persistent for this vertical panorama; even when running with differnt controlpoints and alignments, enblend crashed.
So I made Hugin save the repapped files and ran enblend interactively with --verbose=2; it also crashed:
---
D:\Work\
enblend --verbose=2 -o test.jpj 1040122-
enblend: warning: unknown filetype of output file "test.jpj"
enblend: note: will fallback to type "pnm"
enblend: info: loading next image: 1040122-
enblend: info: loading next image: 1040122-
enblend: info: creating coarse/8 blend mask: 1/3 2/3 3/3
enblend: info: optimizing 1 distinct seam
enblend: info: Annealing Optimizer, s0: 1/4 2/4 3/4 4/4
enblend: info: Dijkstra Optimizer: s0
enblend: info: using 10 blending levels
enblend: info: generating Gaussian pyramid: g0 g1 g2 g3 g4 g5 g6 g7 g8 g9
enblend: info: generating Gaussian pyramid: g0 g1 g2 g3 g4 g5 g6 g7 g8 g9
enblend: info: generating Laplacian pyramid: l0 l1 l2 l3 l4 l5 l6 l7 l8 l9
enblend: info: generating Gaussian pyramid: g0 g1 g2 g3 g4 g5 g6 g7 g8 g9
enblend: info: generating Laplacian pyramid: l0 l1 l2 l3 l4 l5 l6 l7 l8 l9
enblend: info: blending layers: l0 l1 l2 l3 l4 l5 l6 l7 l8 l9
enblend: info: collapsing Laplacian pyramid: l9 l8 l7 l6 l5 l4 l3 l2 l1 l0
enblend: info: loading next image: 1040122-
enblend: info: creating coarse/8 blend mask: 1/3 2/3 3/3
enblend: info: optimizing 1 distinct seam
enblend: info: Annealing Optimizer, s0: 1/4 2/4 3/4 4/4
enblend: info: Dijkstra Optimizer: s0
enblend: info: using 10 blending levels
enblend: info: generating Gaussian pyramid: g0 g1 g2 g3 g4 g5 g6 g7 g8 g9
enblend: info: generating Gaussian pyramid: g0 g1 g2 g3 g4 g5 g6 g7 g8 g9
enblend: info: generating Laplacian pyramid: l0 l1 l2 l3 l4 l5 l6 l7 l8 l9
enblend: info: generating Gaussian pyramid: g0 g1 g2 g3 g4 g5 g6 g7 g8 g9
enblend: info: generating Laplacian pyramid: l0 l1 l2 l3 l4 l5 l6 l7 l8 l9
enblend: info: blending layers: l0 l1 l2 l3 l4 l5 l6 l7 l8 l9
enblend: info: collapsing Laplacian pyramid: l9 l8 l7 l6 l5 l4 l3 l2 l1 l0
enblend: info: loading next image: 1040122-
enblend: info: creating coarse/8 blend mask: 1/3 2/3 3/3
enblend: info: optimizing 1 distinct seam
enblend: info: Annealing Optimizer, s0: 1/4 2/4 3/4 4/4
enblend: info: Dijkstra Optimizer: s0
enblend: info: using 10 blending levels
enblend: info: generating Gaussian pyramid: g0 g1 g2 g3 g4 g5 g6 g7 g8 g9
enblend: info: generating Gaussian pyramid: g0 g1 g2 g3 g4 g5 g6 g7 g8 g9
enblend: info: generating Laplacian pyramid: l0 l1 l2 l3 l4 l5 l6 l7 l8 l9
enblend: info: generating Gaussian pyramid: g0 g1 g2 g3 g4 g5 g6 g7 g8 g9
enblend: info: generating Laplacian pyramid: l0 l1 l2 l3 l4 l5 l6 l7 l8 l9
enblend: info: blending layers: l0 l1 l2 l3 l4 l5 l6 l7 l8 l9
enblend: info: collapsing Laplacian pyramid: l9 l8 l7 l6 l5 l4 l3 l2 l1 l0
enblend: info: loading next image: 1040122-
enblend: info: creating coarse/8 blend mask: 1/3 2/3 3/3
enblend: info: optimizing 1 distinct seam
enblend: info: Annealing Optimizer, s0: 1/4 2/4 3/4 4/4
D:\Work\
---
Running with --verbose=4, the last lines were:
enblend: info: loading next image: 1040122-
enblend: info: assembled images bounding box: [(0, 4228) to (3688, 7062) = (3688x2834)]
enblend: info: image union bounding box: [(0, 0) to (3688, 7062) = (3688x7062)]
enblend: info: image intersection bounding box: [(0, 4228) to (3688, 6218) = (3688x1990)]
enblend: info: creating coarse/8 blend mask: 1/3 2/3 3/3
enblend: info: optimizing 1 distinct seam
enblend: info: difference statistics: overlap size = 1849532 pixels
enblend: info: difference statistics: mismatch average = 0.0333519 [delta-e]
enblend: info: difference statistics: standard deviation = 0.0519767 [delta-e]
enblend: info: Annealing Optimizer, s0: 1/4 2/4 3/4 4/4
I had the idea to try the same input files on openSUSE Leap 15.3 (enblend- enfuse- 4.2-bp153. 2.1.x86_ 64):
There the program did not crash; instead it seems to output a warning:
...
enblend: info: Annealing Optimizer, s0: 1/4 2/4 3/4 4/4
enblend: info: Dijkstra Optimizer: s0enblend: warning: unable to run Dijkstra optimizer
enblend: note: seam-line end point outside of cost-image
enblend: note: contour #1 of 1, segment #1 of 1, vertex #4 of 183
enblend: info: mask transition line bounding box: [(0, 5109) to (3688, 6235) = (3688x1126)]
enblend: info: region-of-interest bounding box: [(0, 0) to (3688, 7062) = (3688x7062)]
enblend: info: using 10 blending levels
enblend: info: generating Gaussian pyramid: g0 g1 g2 g3 g4 g5 g6 g7 g8 g9
...
The resulting image was 3688x9107 pixels, and I could not see any artefacts; in fact the result looks very good, considering that the contributing shots were made without a tripod.