The biggest Christmas gift of to all of us comes this year from Russ - and it is a huge one. BXT Version 2 is here and it is not just an upgrade - it is ground breaking all over again! Watch this video to fully grasp what miracles BXT 2 can do - we reach today a new milestone in Astrophotography! Enjoy!!!!! https://youtu.be/IC0jSzLRmos
BlurXterminator 2 is here!!!!!!!
Yikes. Now I have to post process everything again 😬
Release is today at 16.00 UTC
Looks promising from the video!
Laurent - SP
Laurent - SP
Thank you Russ! BXT is a truly revolutionary deconvolution tool. Thanks for making it even better.
John
John
Well Written Respectful Supportive
Thank you Sascha,
I did some reprocessing yesterday and bought BXT just yesterday. I used it with default settings on material i shot some time ago with a telescope i sold already. It was the Vixen ED80sf with a TS Flattener.
I was so enthusiastic, that I did a comparison between this (still Version1 of BXT) and the "natural" Version and published it on my little blog (not updated yet):
https://astrocamp.eu/en/blurxterminator-2/
Now I read this! I did an Update in PixInsight and added my comparison. Wow! Thanks Russ!
This is only the upper left corner from the ed80sf, integrated, dc, dbe, spcc and then:
left: NXT; middle: BXT, NXT; right: BXT2, NXT.
I did some reprocessing yesterday and bought BXT just yesterday. I used it with default settings on material i shot some time ago with a telescope i sold already. It was the Vixen ED80sf with a TS Flattener.
I was so enthusiastic, that I did a comparison between this (still Version1 of BXT) and the "natural" Version and published it on my little blog (not updated yet):
https://astrocamp.eu/en/blurxterminator-2/
Now I read this! I did an Update in PixInsight and added my comparison. Wow! Thanks Russ!
This is only the upper left corner from the ed80sf, integrated, dc, dbe, spcc and then:
left: NXT; middle: BXT, NXT; right: BXT2, NXT.
Happy Happy Joy Joy
Man this cleans up some messy stars! No need for expensive scopes anymore (just kidding).
Oh well, more work to do - many stacks have to be reworked I guess! :-)
ok, no need to buy the new scope due coma 
This is a very nice Christmas Gift
Just compared: no modification - earlier version - new version...
Just compared: no modification - earlier version - new version...
Wow. Guess I can get rid of my EAFs?
As soon as I learned about it I reran an already stretched completed image through bxt2 like we aren't supposed to, but it still came out way im
proved, with all the misalignment coma gone! . Maybe I don't have to align optics anymore 😏
proved, with all the misalignment coma gone! . Maybe I don't have to align optics anymore 😏
I'm impressed by it. Could never get the 6th star by the brightest trapezium star to resolve fully until now.
Mind. Blown.
Just like last year, another amazing Xmas gift from Russ!
Just like last year, another amazing Xmas gift from Russ!
PI crashed trying to update BX this morning. I guess I didn't notice build 1593 released last week. So if you have not updated PI, you may not be able to update BX.
Ian McIntyre:
PI crashed trying to update BX this morning. I guess I didn't notice build 1593 released last week. So if you have not updated PI, you may not be able to update BX.
Have an "old" build of PI, updated BXT without problems.
This is a very nice Christmas Gift
Just compared: no modification - earlier version - new version...
Thanks for the comparison, that's extraordinary!
Ian McIntyre:
PI crashed trying to update BX this morning. I guess I didn't notice build 1593 released last week. So if you have not updated PI, you may not be able to update BX.
Have an "old" build of PI, updated BXT without problems.
Must just be me then
Getting the following error:
Error: ERROR: MLLoad() could not load tensorflow graph. Error 3: Converting GraphDef to Graph has failed. The binary trying to import the GraphDef was built when GraphDef version was 440. The GraphDef was produced by a binary built when GraphDef version was 1645. The difference between these versions is larger than TensorFlow's forward compatibility guarantee. The following error might be due to the binary trying to import the GraphDef being too old: NodeDef mentions attr 'explicit_paddings' not in Op output:T; attr=T:type,default=DT_FLOAT,allowed=[DT_HALF, DT_BFLOAT16, DT_FLOAT, DT_DOUBLE, DT_INT32, DT_INT64, DT_UINT8, DT_INT16, DT_INT8, DT_UINT16, DT_QINT8]; attr=ksize:list(int),min=4; attr=strides:list(int),min=4; attr=padding:string,allowed=["SAME", "VALID"]; attr=data_format:string,default="NHWC",allowed=["NHWC", "NCHW", "NCHW_VECT_C"]>; NodeDef: {{node BXT/max_pooling2d_14/MaxPool}}. (Check whether your GraphDef-interpreting binary is up to date with your GraphDef-generating binary.).
Anyone had this and knows of a solution?
Error: ERROR: MLLoad() could not load tensorflow graph. Error 3: Converting GraphDef to Graph has failed. The binary trying to import the GraphDef was built when GraphDef version was 440. The GraphDef was produced by a binary built when GraphDef version was 1645. The difference between these versions is larger than TensorFlow's forward compatibility guarantee. The following error might be due to the binary trying to import the GraphDef being too old: NodeDef mentions attr 'explicit_paddings' not in Op output:T; attr=T:type,default=DT_FLOAT,allowed=[DT_HALF, DT_BFLOAT16, DT_FLOAT, DT_DOUBLE, DT_INT32, DT_INT64, DT_UINT8, DT_INT16, DT_INT8, DT_UINT16, DT_QINT8]; attr=ksize:list(int),min=4; attr=strides:list(int),min=4; attr=padding:string,allowed=["SAME", "VALID"]; attr=data_format:string,default="NHWC",allowed=["NHWC", "NCHW", "NCHW_VECT_C"]>; NodeDef: {{node BXT/max_pooling2d_14/MaxPool}}. (Check whether your GraphDef-interpreting binary is up to date with your GraphDef-generating binary.).
Anyone had this and knows of a solution?
Anthony Glas:
Getting the following error:
Error: ERROR: MLLoad() could not load tensorflow graph. Error 3: Converting GraphDef to Graph has failed. The binary trying to import the GraphDef was built when GraphDef version was 440. The GraphDef was produced by a binary built when GraphDef version was 1645. The difference between these versions is larger than TensorFlow's forward compatibility guarantee. The following error might be due to the binary trying to import the GraphDef being too old: NodeDef mentions attr 'explicit_paddings' not in Op output:T; attr=T:type,default=DT_FLOAT,allowed=[DT_HALF, DT_BFLOAT16, DT_FLOAT, DT_DOUBLE, DT_INT32, DT_INT64, DT_UINT8, DT_INT16, DT_INT8, DT_UINT16, DT_QINT8]; attr=ksize:list(int),min=4; attr=strides:list(int),min=4; attr=padding:string,allowed=["SAME", "VALID"]; attr=data_format:string,default="NHWC",allowed=["NHWC", "NCHW", "NCHW_VECT_C"]>; NodeDef: {{node BXT/max_pooling2d_14/MaxPool}}. (Check whether your GraphDef-interpreting binary is up to date with your GraphDef-generating binary.).
Anyone had this and knows of a solution?
Russ has a FAQ on this: https://www.rc-astro.com/faq/blurxterminator-2-0-ai4-error-in-pixinsight-mlload-could-not-load-tensorflow-graph/
Francesco Meschia:Anthony Glas:
Getting the following error:
Error: ERROR: MLLoad() could not load tensorflow graph. Error 3: Converting GraphDef to Graph has failed. The binary trying to import the GraphDef was built when GraphDef version was 440. The GraphDef was produced by a binary built when GraphDef version was 1645. The difference between these versions is larger than TensorFlow's forward compatibility guarantee. The following error might be due to the binary trying to import the GraphDef being too old: NodeDef mentions attr 'explicit_paddings' not in Op output:T; attr=T:type,default=DT_FLOAT,allowed=[DT_HALF, DT_BFLOAT16, DT_FLOAT, DT_DOUBLE, DT_INT32, DT_INT64, DT_UINT8, DT_INT16, DT_INT8, DT_UINT16, DT_QINT8]; attr=ksize:list(int),min=4; attr=strides:list(int),min=4; attr=padding:string,allowed=["SAME", "VALID"]; attr=data_format:string,default="NHWC",allowed=["NHWC", "NCHW", "NCHW_VECT_C"]>; NodeDef: {{node BXT/max_pooling2d_14/MaxPool}}. (Check whether your GraphDef-interpreting binary is up to date with your GraphDef-generating binary.).
Anyone had this and knows of a solution?
Russ has a FAQ on this: https://www.rc-astro.com/faq/blurxterminator-2-0-ai4-error-in-pixinsight-mlload-could-not-load-tensorflow-graph/
Ahh thank you!
Works well with astigmatic lines but not with high order coma nor with signficant tilt (but better than what it used to do).
Sean Mc:
Man this cleans up some messy stars! No need for expensive scopes anymore (just kidding).
I know you are joking, but really...
Version 2 is incredible! As someone who images extensively with very fast lenses, stars across the frame are nearly indomitable even after extensive adjustment of backfocus. The previous version of BXT could not correct them. Now, the stars look so good I could probably just rely on BXT entirely.
Here are a few tests on an image I took with my f/1.4 lens over the summer. I paid almost no attention to spacing, and indeed I don't even think I was perfectly focused.
Original:
Previous version of BXT:
BXT v2:
Here are a few tests on an image I took with my f/1.4 lens over the summer. I paid almost no attention to spacing, and indeed I don't even think I was perfectly focused.
Original:
Previous version of BXT:
BXT v2:
Well Written Helpful Insightful Concise Engaging