DBE vs Graxpert - A quick look.

[@michaellorentzen1672]

I am so happy you’re back! I still LOVE the previous dbe method you posted, still use it all the time.

[@alexanderreinders4009]

So do I. It even works on difficult dust objects like the Plejades. I always get much better results with the visible Dark dbe Methode then with Graxpert.

[@davebillybish]

Thanks for the straightforward video It’s just the job for us budding astrophotographers. I have just used DBE to remove a gradient on the Cigar galaxy M81. Before the colour in the centre was visible but DBE blew it out. I haven’t got Graxpert but used the new Pixinsight Gradient Correction tool and the results were much better with the colour retained. I will be getting GraxpertAI to give it a go also. Thanks again!

[@riversresearch]

Thanks, Shawn! I am having good success with GraXpertAI too and glad you took the time to make this video. You’ve shared a ton of tips over the years and I appreciate all the help you provide to astrophotography novices like myself. 👍🏻. Clear skies!

[@briangriffiths937]

I use GraXpert all the time now. It's far easier than DBE, and the results are very impressive.

[@somethingclever1234]

same

[@MikeTettenborn]

Hi Shawn. I got a first view of your video. Wo Hoo. And it s a great comparison video too. Thanks!

[@dmcmahon999]

Thank you for the big yellow cursor. It really helps those of us who are unfamiliar with the programs.

[@VisibledarkAstro]

Glad it was helpful!

[@AstroForumSpace]

Thanks Shawn! I agree, Graxpert looks a bit smoother. Are you getting similar results when using devision and/or normalize in DBE? Graxpert sure looks like a time saver, especially when processing multiple, mono stacks.

[@AstroCloudGenerator]

Hi Shawn good to see you back. Wouldn’t it be easier to visualise the difference if both the background models had the same STF?

[@VisibledarkAstro]

Perhaps yes. I could have done that I suppose. But the difference isn't to much really. I think the advantage here is graxpert makes it easy.

[@AstroCloudGenerator]

You're right Graxpert does a great job on most occasions and for sure it is a lot easier than DBE. That said, I'm really pleased they added the possibility to see the background image, I have noticed that on more complex images with little to no true background the background starts to look a lot like the Nebula. Like anything in this hobby, I think it needs to be used judiciously. Thanks for doing the comparisons they are always useful. Perhaps you could do a quick video if you find a Nebula that Graxpert doesn't like. @@VisibledarkAstro

[@terrizittritsch745]

Nice video. When you said DBE is more aggressive, it would be interesting to measure and compare the aggressiveness of each quantitatively. When you showed the backgrounds it looked like the GX background was lighter, wouldn’t this mean the amount of background extracted it higher? I think it may be difficult to compare with out actual numbers or consistent display stretches.

[@juangonzalezsanz3840]

Thanks Shaun. In yout video for DBE you recommended applying division and then substraction. Here you only use substraction... why?

[@VisibledarkAstro]

I was just doing a comparison. There wasnt need to use division on this image that's all. Thanks for watching!

[@KurtFB-r6x]

Okay, now you’re a mind reader! Geez, I was wondering about this (I got GraXpert a few weeks ago) when I was having difficulty with DBE & M45 and GraXpert worked perfectly. I’m still keeping your method of DBE as two of my saved icons (division and subtraction). As always, good stuff! BTW, why do you think that the less aggressive gradient removal would be better for later in your processing?

[@AntonioPena1]

Shawn, thanks for taking your time, I have some cases where Graxpert is great option but others that fails, for example M33 galaxy was able to control it using DBE, Graxpert in fully automatic I believed in some cases fails. Thanks

[@VisibledarkAstro]

Thanks for sharing!

[@gregmckay666]

Nice, fair comparison for this subject. Another test would be on a target that has a fairly large amount of nebulosity across a large portion of the image. I think that is when DBE is going to shine because GraXpert destroys much of that nebulosity. Nothing beats manual placement of gradient points.

[@zborkz]

Have you had any issues with GraXpert with narrowband? I have experienced some unexpected results in narrowband channels only.

[@johnadastra1754]

Yes. NB with GraXpert gives more mixed results.

[@VisibledarkAstro]

I haven't myself. What issue are you having?

[@zborkz]

@@VisibledarkAstro It's often over aggressive near a nebula, creating a darkened area adjacent to it compared to the rest of the starfield. I have not tried removing stars first to see if the issue remains. Only NB seems affected, same target in RGB for stars is not affected. Also thanks for the videos.

[@VisibledarkAstro]

@@zborkz If for NB you are using it on the combined Ha, OIII, SII false colour image it won't work well for that. You have to apply graxpert to the individual mono channels first. The AI was trained on OSC images I believe. It still works well in mono images. Just not combined NB.

[@Steve_The_Ignorant_Astronomer]

Shawn quick question about darks, I am finally getting back into things here and i am taking darks. If I want to make a master flat should I use wbpp and image calibration > I dont have any lights just want to get a just on making darks. And once made can I delete the subs ??

[@VisibledarkAstro]

Hi. Not sure I understand your question exactly. Can you email me instead? Easier via email for this topic. Shawn (at) visibledark.ca

[@bezain7663]

I dif a lot of tests with graXpert and DBE. I must Say GraXpert does 90% of the job , with 1% of thé effort required, but if you want that extra accuracy, DBE with accurate sample points and adequate parameters gives you that extra 10%. Actually , everything lays in the accuracy of the sample points and parameters .

[@messier7849]

Thanks, Shawn. I like GraXpert. And I might be .. no, I am.... a bit lazy... but I frequently use ABE. Most of the time I get good results. 😬

[@N4GW]

Great video

[@andysmith7903]

Would have been helpful to see the effect of smoothing at 1 in Graxpert. Cuiv the lazy geek has looked into it and suggests the smoothing at 0 is more aggressive and default of 1 maybe the way to go first. Thanks for the review though.

[@VisibledarkAstro]

Cuiv. Lol. The graxpert devs I think know better. Zero is their own guidelines based on having created graxpert and the AI functionality.

[@terrydoran8689]

This comparison on gray scale images is all well and good but it doesn't tell us anything about how they both handle color gradients. I have come across more than a few circumstances when DBE just will not deal with a color gradient where GraXpert will. I also find if I open the GraxPert program and manually place sample points and use the RBF model I get better results than using AI. i think their AI model has a ways to go to give accurate results on complex nebula filled images.

[@VisibledarkAstro]

Let's do a video about it then! Good idea.

[@Jonasastrophotos]

I have made the "switch" to Graxpert. Its so much easier

[@carvrodrigo]

Leaving smooth at 0 in Graxpert removes details from most of nebulas there are articles showing this

[@VisibledarkAstro]

Haven't read this or experienced it. The leaving it at zero was a graxpert dev suggestion. I will look into this though. Interesting. Thanks!