126 replies to this topic

[DAA]

Hi Wei-Hao and agavephoto,

 

I ran Wei-Hao's script successfully 3 times in a row without a crash, collecting 120 frames per trial.

 

Now, I'm really curious if my script crashes on Wei-Hao's setup.  Wei-Hao, If you have time to try it, please let me know.

 

My conclusion so far is that issuing camera change instructions (change of ISO, F-number, exposure time) too fast is the cause of the crash.  Without LVPSTART, the mirror mechanics limit frame rate and, maybe coincidentally, prevent crashes.  But with the mirror up and the mechanical limit lifted, sending instructions too fast might be a problem.

 

I still wonder how Xavier does it.  Please let me know what you think since this is the mystery at this point that stands in the way of using the really action-packed parts of Xavier's 'deluxe.txt' script.

 

All the best,

[whwang]

Hi David,

 

I will run your script tomorrow when I have access to the camera, and also try to change the exposure time or ISO in my script.  I will let you know the result.  

 

I am trying to recall the challenges I faced in the 2017 eclipse. And now I believe that changing parameters may cause crashes, as this requires additional time for the communication between the camera and the computer.  As a result, there was huge overhead between my 2017 exposures because I needed to make the exposures well separated in time to prevent crashes.  

 

In 2017, I used a D800.  It doesn't have the silent mode.  The need of MLU to suppress vibration also increased the overhead.  Because there were only limited number of exposures I could take, changing exposure parameters for each of the exposures was a must.  It was quite challenging.  This time, since D850 can deliver many exposures in a very short time, I can simply stack many short exposures to get the S/N that's equivalent to a longer exposure, while still having the unsaturated highlight in the short exposures.  This eliminated the need of many different exposure times.  During the totality, I can just stick to a small number of different exposure times and try to collect as many of them as possible.  I think this will bring huge improvements for the corona HDR.

 

You mentioned you tried an CF express card.  I would suggest to avoid that.  Although CF express can be much faster than XQD, on D850 it is actually slightly slower according to people's tests in the past.  I believe this is because D850 was made very early on so its hardware was not designed to take full advantage of CF express.  

 

Another thing we need to be careful is image compression.  So far I did the stress test with camera covered.  I think from tomorrow I should do it with a lens on it and open the lens.  A completely black exposure and an exposure with rich structure in the image may require slightly different compression time and definitely will lead to different compressed file size (and therefore card writing time).  It will be better to give some content to the image when doing stress test if we use image compression (lossless, of course).

 

Cheers,

Wei-Hao

[Gert]

Hello All,

I didn't have time to read the entire thread, forgive me if it's been asked before. I have watched the TAIC YouTube lecture on eclipse imaging giving examples of both the maestro and orchestrator tools. When they briefly showed the scripts on screen I had the feeling that the syntaxes are same or very close. Is it true?? Will that lead to an exchange capability of scripts between maestro and orchestrator??

Thanks & Clear Skies,
Gert

[DAA]

HI Wei-Hao,

 

Thank you for your answer.  I look forward to hearing whether SEM crashes on your set-up to help me understand whether my problem is pathological or simply a limit of the D850.  It is my experience that commanding cameras to change settings takes a frustratingly long time.

 

You are right about stacking images to improve SNR.  Your example of 20 images will add over 2 stops of dynamic range so you don't have to go to exposures that risk motion blur or saturation.  I have not stacked solar images before and will practice before the eclipse but does the small drift between images make a difference?  Do you align your image manually or use an HDR app?  (Even with a tracker, daytime polar alignment is not going to be precise, I think.)

 

Thank you for the info on the CFExpress card and D850.  I'll just send it back.

 

I took some solar images today and found your script worked flawlessly even with content.  I will try increasing the speed to 6 fps to increase the stress and remember to include content as well with lossless compression.  By the way, is there a time penalty for the compression operation?  In other words, what is the overall tradeoff of not compressing and having large files vs. taking time to compress and then sending smaller files?  I never considered this issue.

 

Looking forward to hearing back,

[DAA]

Hi Gert,

 

The YouTube video here, https://www.youtube....h?v=qLHfURDO1YY, suggests that the scripts were portable between the two eclipse apps you mention, or at least close.

 

David

[whwang]

By the way, is there a time penalty for the compression operation?  In other words, what is the overall tradeoff of not compressing and having large files vs. taking time to compress and then sending smaller files?  I never considered this issue.

 

Unfortunately I don't know.  I asked this on DPReview (can't remember which sub-forum it is) a while ago and no one really had an answer.  I guess for those who really need speed (like sport photography) and push to the limit, their cameras are all much faster than D850.  So even if they have an answer, chance is good that it doesn't apply to D850.  We can only test to find out.

[mcordier]

Greetings all,

 

I was fortunate to be in Nebraska for the totality for the 2017 eclipse.  Since that was my first time being in totality, I just wanted to view the 2 minutes and 36 seconds without worrying about my cameras and technology.

 

So, this April I do plan to image the eclipse using Solar Eclipse Maestro. I purchased an older Mac on ebay for $75.00 and loaded the software. 

 

I have found a few scripts online from some generous people who have posted their scripts.  Attached is the one that I tested and am happy that the camera shutter responded to the simulation.

 

I would very much appreciate if other people had success, or have improved upon their scripts that were used in previous eclipses and share their work.

 

Clear Skies!

 

Mark Smith

 

 

 

Final R5 Script.txt

how did you get your R5 to connect?  I'm running Mojave and the  "snow leopard +" release of SEM and when I try to connect to my camera I'm seeing an error that says "camera #1 is not supported yet by this intel release of solar eclipse Maestro.  please upgrade to a newer release of solar eclipse maestro that supports this camera model", which is weird because the R5 is listed in supported cameras.  i checked to make sure that no other applications are trying to use the camera per the information on the website.  I'm running V1.9.0 of solar eclipse maestro, the most current release.

[agavephoto]

how did you get your R5 to connect?  I'm running Mojave and the  "snow leopard +" release of SEM and when I try to connect to my camera I'm seeing an error that says "camera #1 is not supported yet by this intel release of solar eclipse Maestro.  please upgrade to a newer release of solar eclipse maestro that supports this camera model", which is weird because the R5 is listed in supported cameras.  i checked to make sure that no other applications are trying to use the camera per the information on the website.  I'm running V1.9.0 of solar eclipse maestro, the most current release.

Until a new version is officially released, you'll need to contact the developer to see if he'll supply a link to a beta version that has some support for the R5. Not all is fully functional for me, though I will be able to get the corona shot sequence I want just fine.

Edited by agavephoto, 12 March 2024 - 10:05 PM.

[whwang]

Hi David,

 

I tried your script a few times.  It only crashed my SEM once.  However, for the rests where it didn't crash, sometimes it produced 4 files, sometimes as few as just two.  And I also noticed that when it successfully triggered 4 exposures, sometimes the time label of the exposures shown by the camera indicated a time period (like 4 seconds) much longer than what should be according to the script (about 1.8 seconds).  So there is definitely some large overheads between the exposures, which cased crashing, or delayed triggering, or even no triggering.

 

As for my script, I changed the exposure time and ISO in each of the series of burst, without increasing the gaps between the bursts.  I tried it twice, no crash, and no dropped exposures.

 

If I were to draw a conclusion from this, it would be that the burst mode is the way to go.  Somehow, the overhead in single picture mode is too much and can cause all kinds of troubles.

 

Cheers,

Wei-Hao

[Smitty14]

how did you get your R5 to connect?  I'm running Mojave and the  "snow leopard +" release of SEM and when I try to connect to my camera I'm seeing an error that says "camera #1 is not supported yet by this intel release of solar eclipse Maestro.  please upgrade to a newer release of solar eclipse maestro that supports this camera model", which is weird because the R5 is listed in supported cameras.  i checked to make sure that no other applications are trying to use the camera per the information on the website.  I'm running V1.9.0 of solar eclipse maestro, the most current release.

I contacted Xavier and he sent me the beta version of SEM.  I had the same message attempting to use the Snow Leopard release of SEM.  The beta version recognizes my R5 and 90D cameras.  I hope this helps.

 

Mark

[DAA]

Hi Wei-Hao,

 

Thank you for testing my script.  Yes, when the interval between new instructions (e.g., changing exposure) is 0.6s, crashes are occasional.  As the interval decreases to 0.5s and shorter, crashes are more frequent.  I agree with your conclusions for the D850.  It is simply too slow in updating frame parameters to run at the high frame rate shown in Xavier's deluxe.txt example.

 

With that better understood (so helpful to have two of the same hardware configurations under independent control!), I will use your burst mode method.  I will also play with bracketing.  The minor disadvantage of bracketing is the fixed bracket size, like 0.3 or 1.0 EV rather than the flexibility of specifying an arbitrary exposure.  But maybe (not likely) bracketing gets around the crashing problem in which case it is an alternative for burst mode in the case of very rapidly changing imagery, like during chromosphere or Baily's bead phases.

 

This is very helpful.  I will let you know the results of my next set of tests.  When I get a script that works reliably, I will sent that as well.

 

Cheers,

[DAA]

Hi again, Wei-Hao,

 

Moving on from crashing, I experimented with frame rate to find that 4 fps and 5 fps drop no frames but it seems like 6 fps begins to drop frames using burst mode.  For now, I will limit bursts to 5 fps or slower.

 

Also, in RAMPUP mode, which changes camera settings "as quickly as possible" (according to the Script File Format document for SEM), the D850 acquired 20 frames at 1/3 EV intervals in 19 seconds or about 1s per frame.  This is slow but if it is "as quickly as possible," it might be avoiding crashes.  This might be a convenient method for acquiring data for building up a corona HDR image.  (See also, RAMPDN and RAMPSUD.)

 

Next, I will look at bracketing using burst mode.  Perhaps with the mirror locked up, it will acquire images quicker than other methods that change image parameters.  Or not.  If not, then your method of burst at a give set of [ISO, F-number, exposure time] and then moving on to the next set of parameters appears to be the best way to get a lot of bracketed data without crashing SEM. 

 

For very quick changes and a camera that limits parameter change time, the compromise will be to limit the number of changes and use bursts as a substitute for extended exposures.

 

Best regards,
David

[whwang]

Hi David,

 

Thank you for letting me know about your finding of the RAMPUP mode.  The burst mode is really only suitable for the corona during totality, where relative few different exposure times are needed especially with stacking.  For the Baily's bead, chromosphere, prominence stage, I think constantly changing exposure time is needed, even at the cost of lower frame rate.

 

Also, by paying close attention to the countdown display in SEM for the script and the little LED light on the back of D850, I believe there is a 0.5 sec delay between the start of a burst script and the start of its actual exposures.  This is consistent with the about 0.5 sec of overhead in the single exposure mode that you found.  It's just that all the subsequent exposures in the series of burst does not require passing camera parameters so they can be done extremely fast.

 

Yesterday I settled my corona plan.  Today I will move to testing the various phases around C2 and C3.  I will try the RAMPUP mode.

 

BTW, did you try putting more than 20 exposures in a burst script?  It never worked for me.  20 seems to be my maximum number of exposures in a burst.

 

Cheers,

Wei-Hao

[DAA]

Hi Wei-Hao and agavephoto for the last post of the night,

 

I tried bracketing + low burst speed with the mirror first engaged and then mirror locked up.  All successful and fast without crashing SEM.

 

Here's the script, run with my camera aimed at a clock with millisecond display.

 

BKTDATA,9,0.3EV,u0o
EXPMODM,C2,-,00:32.0,D850, , , , , , , ,set camera to manual (M)
CHGCSSL,C2,-,00:31.0,D850, , ,3, , , , ,burst low speed 3 fps
LVPSTART,C2,-,00:30.0,D850, , , , , , , ,mirror up
SETBKT,C2,-,00:28.0,D850, , ,1, , , , ,activate bracketing for the burst
TAKEBST,C2,-,00:27.0,D850,1/100,8,1000,9,RAW,None,N,burst of 9 bracketed pictures
SETBKT,C2,-,00:05.0,D850, , ,0, , , , ,deactivate bracketing
LVPSTOP,C2,-,00:03.0,D850, , , , , , , ,mirror down

 

Both with mirror engaged (with no delay) and with mirror locked up, the images rolled off with an average interval of 0.33s, determined by the burst speed.  This is at least as fast as Xavier's 'deluxe.txt' sample script with the small disadvantage of loss of flexibility in choosing exposures arbitrarily.  I see that is a minor inconvenience.

 

Still to test is (1) the case when the exposure time is longer than the burst interval.  Will the camera slow gracefully to accommodate the exposure or will something crash?  And (2) how fast a burst speed with this method tolerate if the case of exposure > burst interval is tolerated?  I already saw that 6 fps causes frame dropping for the D850 but what about 4 and 5 fps?

 

In any case, I think that this method is the key to getting HDR-ready images for the very dynamic phases of the eclipse, namely Baily's beads and the appearance of the chromosphere where a bit of HDR will be a big help but only if there is reasonably small change between exposures.  Worst case -- HDR doesn't work and a single shot within a span of about 1 second is a good exposure.

 

So, with Wei-Hao's burst method during the coronal phase and mirror-up bracketing bursts during the dynamic phases and just TAKEPIC for the pre-C2 and post-C3 phases or some simple looping, SEM should provide really good results.

 

Once again, thank you Cloudy Nights for bringing making this thread possible and to all the contributors.

[FXM]

Here are mine tested and so far no issues. These are by no means perfect but work for me as I am trying to burst mode for Diamond and Baily's Beads. Trying to fill the time taking images. 

 

The D7000 will burst up to 11 frames but then the buffer will fill and slow down but I am ok with that.

 

One is for a Nikon D7000 with 2.2 seconds between shots, Diamond Ring, NO Baily’s Beads, Chromosphere, Prominences, Corona fill.

 

The other script is for D7500 0.8 seconds between shots, NO Diamond Ring, Baily’s Beads, Chromosphere, Prominences, Corona fill.

 

Both emergency scripts by Jerry Lodriguss.

 

The audio is custom and I can post a link if anyone wants.

Attached Files

•  MSTR DLUX 7500 No Dia _8 sec 10 sets.txt   12.57KB   41 downloads
•  emergency script D7500.txt   160bytes   28 downloads
•  MSTR DLUX 7000 2_2sec 5 sets DIA NO BAILEY.txt   6.74KB   26 downloads
•  emergency script D7000.txt   160bytes   19 downloads

Edited by FXM, 13 March 2024 - 09:16 PM.

[whwang]

Talking about audio file, does anyone know an online AI that can created voice audio based on your input text and output files that can be accepted by SEM?

[Nn1982i]

Does anyone have a script with bracketing and auto-rank that one can share? Thanks. Having a series of clear skies the next few days that I would love to use the script to test out.

[whwang]

Have you tried SEM's auto configuration wizard?  It can generate a script that does most of the things, and you can start from there.

[Nn1982i]

I have. But it does seem to create a set of bracketed exposures (or maybe I just could not find it).

[astronerd1]

Hello All,

I didn't have time to read the entire thread, forgive me if it's been asked before. I have watched the TAIC YouTube lecture on eclipse imaging giving examples of both the maestro and orchestrator tools. When they briefly showed the scripts on screen I had the feeling that the syntaxes are same or very close. Is it true?? Will that lead to an exchange capability of scripts between maestro and orchestrator??

Thanks & Clear Skies,
Gert

I noticed that there are differences in the syntax between EO and SEM.

For one, EO can have commas together when a field id empty (i.e. PLAY,C2,-,00:10:00.0,Sounds/10minutes.wav,,,,,,,,"10 minutes" voice prompt) where as SEM must have spaces for an empty field (i.e. PLAY,C2,-,00:10:00.0,10minutes.wav, , , , , , , , "10 minutes" voice prompt).

[astronerd1]

I noticed that there are differences in the syntax between EO and SEM.

For one, EO can have commas together when a field id empty (i.e. PLAY,C2,-,00:10:00.0,Sounds/10minutes.wav,,,,,,,,"10 minutes" voice prompt) where as SEM must have spaces for an empty field (i.e. PLAY,C2,-,00:10:00.0,10minutes.wav, , , , , , , , "10 minutes" voice prompt).

My problem is in the use of the reference 'LAST'. The code in EO works well when one is spacing shots on the fly, as in:

##
TAKEPIC,C2,+,00:00:05.0,D800,1,8.0,100,0.000,RAW,,Y,
TAKEPIC,LAST,+,00:00:02.0,D800,1,8.0,100,0.000,RAW,,Y,
#
TAKEPIC,LAST,+,00:00:02.0,D800,0.793701,8.0,100,0.000,RAW,,Y,
TAKEPIC,LAST,+,00:00:02.0,D800,0.793701,8.0,100,0.000,RAW,,Y,
#
TAKEPIC,LAST,+,00:00:02.0,D800,0.629961,8.0,100,0.000,RAW,,Y,
TAKEPIC,LAST,+,00:00:02.0,D800,0.629961,8.0,100,0.000,RAW,,Y,
#

This allows 2 sec between shots and does not depend on specific timing for each shot (i.e. Take a pic 2 sec after last shot).

'LAST' is listed in the manual of SEM but I have yet figured out how to use it.

Anyone have any ideas?

Edited by astronerd1, 15 March 2024 - 12:49 PM.

[DAA]

Hi Wei-Hao,

 

I keep forgetting to test maximum number of images per burst with the D850 but now have tested

-- at 3 and 5 fps and

-- without and with mirror lock-up and

-- without and with the deprecated command 'BURSTNB'.  

 

I confirm your finding -- 20 images per burst seems to be the limit with the D850, despite the Script File Format suggesting, "maximum of 200 for exposures under 2 seconds and 960 for exposures over 2 seconds.'

 

No explanation except some D850 hardware limitation.

 

David

 

PS:  Is anyone using a different camera finding a limit on burst number?

[DAA]

... and another finding for Nikon D850, this time about bracketing.

 

1.  Bracketing works at 6 fps and accommodates longer exposure times that the nominal 1/6 s if they are part of the bracket.  For example, if you set bracket interval to 1.0EV with 7 images in the bracket centered around 1/100s, you will get 1/800, 1/400, 1/200, the center image at 1/100, then 1/50, 1/25 and 1/13.  But if you center at 1/25, you will get 1/200, 1/100, 1/50, the center image at 1/25, then 1/13, 1/6 and 1/3.  But even though the frame rate is set to 6 fps, the last two frames will be 1/6 an 1/3 seconds.  This is useful for cramming in short exposures with spacing of 167ms but accommodating longer frames as long as you don't jam this bracketing step against the next step with too small an interval.  

 

2.  The maximum number of image in a bracket for the D850 appears to be 9.  I cannot figure how to get past this limit.  (Like the limit in number of frames in a burst, this might be written in the camera hardware.)  Still, 9 images with a good EV interval and well chosen center exposure seems very useful since it proceeds so fast!  Certainly good for the corona phase when time isn't as much in demand but maybe even for bracketing beads with 3 or 5 images.

 

Next -- look for D850 limits on frames in RAMPUP, RAMPDN and RAMPSUD sequences.

[DAA]

... and one last observation, this time about RAMPUP, which seems like a useful way of bypassing the bracketing limit of 9 frames at a slight loss of speed.

 

RAMPUP (and presumably RAMPDN and RAMPSUD although I have not yet tested these other commands) takes "as quickly as possible an upward ramp of exposures auto-adjusting the settings."  Unlike bracketing, which is written into the D850 hardware, RAMPUP must send commands to the camera between frames, I am guessing.  For this reason, "as quickly as possible" is not the 6 fps for short exposure bracketing.  Rather it turns out to be about 3 fps, which is slower but not as slow as TAKEPIC.  

 

Also, RAMPUP naturally accommodates for long exposures by slowing down appropriately if the exposure exceeds 1/3 second.

 

If RAMPUP has a downside of being slower for short exposures than bracketing, its advantage is that there seems to be no limit in number of frames in a ramp, at least up to the 25 that I tried.  (I mean how many exposures do you want in a ramp?)

 

So, to summarize results for the Nikon D850,

 

TAKEPIC seems to not like intervals shorter than 0.7s, especially when the mirror is locked up.  At 0.6s intervals with mirror up, SEM crashes occasionally and with still shorter intervals, SEM crashes more often.

TAKEBST seems to work up to 5 fps and up to 20 frames in a burst.  At 6 fps, frames are occasionally dropped.  There doesn't seem to be a way of passing the 20 frame limit with the D850.

TAKEBKT runs up to 6 fps, accommodates frames longer than 1/6s by slowing down gracefully but only allows up to 9 frames in a burst.

RAMPUP (and maybe its cousins RAMPDN & RAMPSUD) run at about 3 fps, probably sending commands to the camera for each frame, accommodate for any exposure time gracefully and seem to be unlimited in number of frames in a ramp.

 

How certain am I of these conclusions?  Not all that certain because the experiments were short and simple so edge cases were not tested.  If you find exceptions, please post.

 

But now it is time to optimize an image acquisition script through totality from beads to diamonds to chromosphere to corona and back again.  Xavier's 'deluxe.txt' is a very useful starting point for timing and events to capture.

[whwang]

Now I basically have finished my preparation of the D850 scrip, so I move on to prepare my GFX-100S script.

 

My 2017 experience told me that SEM suggests an exposure time that's way too long for the no-filter version of Baily's beads.  The solar photosphere was terribly over-exposed and the beads were not sharp at all.  So for D850, I pushed to the lowest ISO and shortest exposure time.  For the scope that goes with D850, this will be -3.5 EV from SEM's recommendation.  This is likely to still saturate the photosphere, but should be much better.  Then I found that for GFX-100S, I can only push to -1.6 EV from SEM's recommendation.  So chance is good that I will terribly over-expose the photosphere again on GFX-100S.

 

The reason I cannot push down GFX-100S's exposure is that although SEM shows that there are 1/16000, 1/13000, and 1/10000 secs available in the camera property list, TAKEPIC command for these exposure times all ended up with just 1/8000 sec according to the record of the pictures.  So 1/8000 sec is practically the shortest exposure time.  (BTW, in case you are wondering, I am using electronic shutter to avoid any shutter vibration.  Without SEM, the camera's shutter button can trigger 1/16000 sec exposures without problems.)

 

If it were half year ago, I would probably write to Xavier to ask about this.  I think now he is occupied with all kinds of requests/questions, and his own preparation for the eclipse.  I will not bother him and I will just live with this.