P Cyg last night!

[Gasman]

Hi guys

Not been on for a while!.  After a very promising start I`ve been trying out Ekos/Kstars for controlling the scope/ccd setup and big respect to the Kstar/Ekos guys but on my setup I found it just a little too buggy for now so drifted back to Windows!. Anyway back to the important stuff

Had a clear evening the other night and had a try with  P Cyg - very active with well defined lines so did a quick process tonight and compared John`s BASS with ISIS!. Note - strip not aligned with profile!

and the comparison ISIS plot

cheers

Steve

[Jack Martin]

Nice spectra, good work.

Regards,

Jack

[Gasman]

Thanks Jack. P Cyg always seems to display well when sky conditions are right!

Steve

[Merlin66]

Steve,

I'm assuming same pre-treatment, for any slant/ tilt, similar binning and sky removal...

Obviously I can't confirm calibration from the images but I assume they agree.

The major features recorded seem to compare... but the shape of the continuum.. hmmm something different there....

Any ideas???

[Ernst Pollmann]

Hi Steve,
maybe you know the AAVSO-ASPA-BAV amateur long-term campaign of the star. We are monitoring the star since 2008 in terms of study the Halpha intrinsic flux in collaboration with the professional astronomy. See:

http://astrospectroscopy.de/PCyg_report_2013.pdf

http://astrospectroscopy.de/PCYG_AAVSO.pdf

Before that background I am very interesting in your spectrum. Would you be willing to send me it?

Ernst Pollmann

--------------------------------------------

International Working Group ASPA

Active Spectroscopy in Astronomy

http://www.astrospectroscopy.de

http://www.astronomie.de/astronomische-fachgebiete

[Gasman]

Hi Ken & Ernst

No worries Ken. those profiles are exactly what ISIS and BASS churned out with no continuum altering at all. Bear in mind its a whole month since I did any spectro so the Gasman grey matter might not quite be back up to speed (if it ever was ).

Ernst

Would you like the 10 fits images used to stack/create the spectrum or just the final ISIS processed spectrum ?

regards

Steve

[Ernst Pollmann]

Steve,

the processed spectrum would be fine.

Ernst

[iksobarg]

What in the out of this world?

I have some familiarity with this topic but am quite uninitiated.

Would you or someone give a nice report/detail of this?

[Gasman]

Hi iksobarg

Welcome to the little wiggly lines section of SGL  . The above images are the end result after processing some fits  taken by using a spectroscope attached to (in my case) a 120mm refractor. The top image has been processed with the excellent free spectro app called BASS and you can see I`ve been able to id some elements from within the software. The bottom is from an equally excellent free app (but a slightly steeper learning curve) ISIS from Christian Buils site. Thats basically it but we can go into more detail if you want!

cheers

Steve

[Ernst Pollmann]

Hi Ken & Ernst

No worries Ken. those profiles are exactly what ISIS and BASS churned out with no continuum altering at all. Bear in mind its a whole month since I did any spectro so the Gasman grey matter might not quite be back up to speed (if it ever was ).

Ernst

Would you like the 10 fits images used to stack/create the spectrum or just the final ISIS processed spectrum ?

regards

Steve

Steve,

here comes the evaluation of your spectrum.

The lower Fig. shows the evaluation of the sum spectrum after subtraction of the sky background: EW = 79 A (+/- 3%)

The upper Fig. shows the evaluation of your scanned profile: EW = 72 A. I assume, that the sky background has been not subtracted already. That would explain the lower EW.

The EW of the lower Fig. has been inserted in the AAVSO-ASPA-BAV long-term monitoring and does fit very well.

If there is a Vmag value in the AAVSO data base, then we could calculate the intrinsic Halpha flux to supplement the corresponding AAVSO-ASPA-BAV flux monitoring. That is the sence of our long-term program since 2008.

Ernst Pollmann

--------------------------------------------

International Working Group ASPA

Active Spectroscopy in Astronomy

http://www.astrospectroscopy.de

http://www.astronomi...che-fachgebiete

[Ernst Pollmann]

Steve,

here comes the evaluation of your spectrum.

The lower Fig. shows the evaluation of the sum spectrum after subtraction of the sky background: EW = 79 A (+/- 3%)

The upper Fig. shows the evaluation of your scanned profile: EW = 72 A. I assume, that the sky background has been not subtracted already. That would explain the lower EW.

The EW of the lower Fig. has been inserted in the AAVSO-ASPA-BAV long-term monitoring and does fit very well.

If there is a Vmag value in the AAVSO data base, then we could calculate the intrinsic Halpha flux to supplement the corresponding AAVSO-ASPA-BAV flux monitoring. That is the sence of our long-term program since 2008.

evaluation.jpg

pcyg_Halpha-EW.jpg

Ernst Pollmann

--------------------------------------------

International Working Group ASPA

Active Spectroscopy in Astronomy

http://www.astrospectroscopy.de

http://www.astronomi...che-fachgebiete

Steve,

there is a contemporaneous Vmag at the AAVSO database. This enables to calculate the intrinsic Halpha flux (= EW corrected for the visual brightness). See the last (your) measurement point in the flux monitoring.

Ernst

[Gasman]

Interesting Ernst! at least my obs seems to fit in with the others. It would also be interesting to see the process of calculating the EW.

regards

Steve

[Ernst Pollmann]

Interesting Ernst! at least my obs seems to fit in with the others. It would also be interesting to see the process of calculating the EW.

regards

Steve

... the EW is the Intergration of the area of the Halpha peak of the on 1 normalized spectrum, with an integration range from 6520 to 6610 Angstr.

The program VSpec performs this calculation in a convenient manner.

Ernst

[Merlin66]

Steve,

BASS also calculates EW the same way....

Look under "Tools"/"Measurements and Elements", tick "Equivalent width"

[robin_astro]

... the EW is the Intergration of the area of the Halpha peak of the on 1 normalized spectrum, with an integration range from 6520 to 6610 Angstr.

The program VSpec performs this calculation in a convenient manner.

Ernst

Hi Ernst, everyone,

The calculation of EW, while simple on paper is one of the most difficult to make accurately  in practise with many potential pitfalls.  One major one concernsthe fact that the measurement of the area is made relative to the continuum but defining where the continuum actually is and getting the program to recognise this is not trivial. There are as many techniques to define this as there are programs calculating it and in some cases the differences can be large so it is important to know what a particular program is doing and what the consequences are to the measurement if you want consistent results.

For example you might be interested that I found when using VSpec for the epsilon Aurigae work that the standard calculation in Visual Spec is not particularly good in some circumstances and can give significantly different results depending on the exact values at the calculation limits. (Moving just them by just 1 pixel  can at times have a dramatic effect on the result which was clearly a problem.) 

I therefore  worked with Valerie Desnoux to explore this and developed some alternative calculations which are also included in Vspec under "spectrometry"  "EW lin fit", "EW cnst", "EW cnst 1" .  If you try each of these options you can see exactly where VSpec assumes the continuum is in each case. (an orange line plot of the assumed continuum is overlaid on the spectrum)

"EW lin fit" is the same as the standard calculation on the right mouse click.  It assumes the edges of the selection lie exactly on the continuum and that the continuum is a straight line slope between them.  This assumption is reasonably ok as long as the continuum is reasonably flat and noise free but if it is noisy then a small change to the exact points selected (eg at a noise peak or trough) can make a big difference to the result, particularly when measuring weak lines. It is because of this that I developed with Valerie some alternative calculation techniques.

"EW cnst fit  is as above but it assumes that the continuum is level ie not sloping. This still suffers from the effect of noise in the continuum though. 

"EW cnst 1"  is  the one I prefer as it assumes the continuum has already been rectified (ie normalised so the continuum = exactly 1 everywhere) and that any variations from this are just due to noise (or the line being measured of course)  It is up to the user how this rectification is done and the best technique to use is in itself a big subject but, since it has to be done at some stage in the measurement of EW, I prefer to separate it out rather than it be combined with the EW calculation as is sometimes the case, for example in SpcAudace.

Once this has been done, the calculation is simple as all VSpec has to do is assume that the value of the continuum  between the EW calculation limits is exactly 1 everywhere (even though in practise it will not be exactly 1 everywhere due to noise)

Note that these are relatively simple techniques to deal with the continuum and there are  more sophisticated methods around in the professional sphere. In general though different programs measuring EW will give different results, (sometimes the differences are small , sometimes significantly large) and I found from experience that at the end of the day,  the variability between observers could be reduced significantly by  having one person make all the measurements on the spectra using the same technique, re-rectifying them using a common procedure if necessary.  (Note that this assumes the spectra have already been accurately reduced to remove camera, background and instrumental effects which is an important prerequisite for accurate measurement of EW)

Cheers

Robin

[Ernst Pollmann]

Hi Ernst, everyone,

The calculation of EW, while simple on paper is one of the most difficult to make accurately  in practise with many potential pitfalls.  One major one concernsthe fact that the measurement of the area is made relative to the continuum but defining where the continuum actually is and getting the program to recognise this is not trivial. There are as many techniques to define this as there are programs calculating it and in some cases the differences can be large so it is important to know what a particular program is doing and what the consequences are to the measurement if you want consistent results.

For example you might be interested that I found when using VSpec for the epsilon Aurigae work that the standard calculation in Visual Spec is not particularly good in some circumstances and can give significantly different results depending on the exact values at the calculation limits. (Moving just them by just 1 pixel  can at times have a dramatic effect on the result which was clearly a problem.) 

I therefore  worked with Valerie Desnoux to explore this and developed some alternative calculations which are also included in Vspec under "spectrometry"  "EW lin fit", "EW cnst", "EW cnst 1" .  If you try each of these options you can see exactly where VSpec assumes the continuum is in each case. (an orange line plot of the assumed continuum is overlaid on the spectrum)

"EW lin fit" is the same as the standard calculation on the right mouse click.  It assumes the edges of the selection lie exactly on the continuum and that the continuum is a straight line slope between them.  This assumption is reasonably ok as long as the continuum is reasonably flat and noise free but if it is noisy then a small change to the exact points selected (eg at a noise peak or trough) can make a big difference to the result, particularly when measuring weak lines. It is because of this that I developed with Valerie some alternative calculation techniques.

"EW cnst fit  is as above but it assumes that the continuum is level ie not sloping. This still suffers from the effect of noise in the continuum though. 

"EW cnst 1"  is  the one I prefer as it assumes the continuum has already been rectified (ie normalised so the continuum = exactly 1 everywhere) and that any variations from this are just due to noise (or the line being measured of course)  It is up to the user how this rectification is done and the best technique to use is in itself a big subject but, since it has to be done at some stage in the measurement of EW, I prefer to separate it out rather than it be combined with the EW calculation as is sometimes the case, for example in SpcAudace.

Once this has been done, the calculation is simple as all VSpec has to do is assume that the value of the continuum  between the EW calculation limits is exactly 1 everywhere (even though in practise it will not be exactly 1 everywhere due to noise)

Note that these are relatively simple techniques to deal with the continuum and there are  more sophisticated methods around in the professional sphere. In general though different programs measuring EW will give different results, (sometimes the differences are small , sometimes significantly large) and I found from experience that at the end of the day,  the variability between observers could be reduced significantly by  having one person make all the measurements on the spectra using the same technique, re-rectifying them using a common procedure if necessary.  (Note that this assumes the spectra have already been accurately reduced to remove camera, background and instrumental effects which is an important prerequisite for accurate measurement of EW)

Cheers

Robin

Hi Robin,

great, that you took time, to supplement with this explanations the complex topic in that way.

Are you participant of the OHP workshop this year?

Ernst

[Gasman]

I think a spectro workshop here in the UK would be most useful and I think very welcome for many astronomers especially those new to spectroscopy to deal with a lot of concepts such as the above. I'll be honest and confess I didn't realise the importance of EW and also a few other concepts are still grey area's to me! Maybe a subject for another post ;-)

Cheers

Steve

[Ernst Pollmann]

I think a spectro workshop here in the UK would be most useful and I think very welcome for many astronomers especially those new to spectroscopy to deal with a lot of concepts such as the above. I'll be honest and confess I didn't realise the importance of EW and also a few other concepts are still grey area's to me! Maybe a subject for another post ;-)

Cheers

Steve

... in general it would be a wonderful idea, but the question is, whether there are enough interested amateurs who are willing to take part. This depends on you to find out that and to start an attempt respectively. But you have to know, there are some important frame conditions: location, attractive talks, date, accomodation costs and so on....

I would be interested in it.

Ernst

[robin_astro]

Hi Robin,

Are you participant of the OHP workshop this year?

Hi Ernst,

Not this year I am afraid. Say hello from me when you are there.  

I was in Cologne a couple of months back for the VdS conference and caught up with a few old friends there including Olivier Thizy. He was saying It is very popular this year with many participants off site in local hotels. I need to get more organised and get back there one year. I can definitely recommend it for anyone wanting to get up to speed. It helped me get to grips with my LHIRES back in 2007

Cheers

Robin

[chiltonstar]

Coming back to the original post, wouldn't we expect to see the absorbtion dip  on H Alpha at this resolution?

Chris

[Ernst Pollmann]

Coming back to the original post, wouldn't we expect to see the absorbtion dip  on H Alpha at this resolution?

Chris

... ... in order to see the blue shifted absorption component you need a corresponding dispersion and a spectral resolution respectively. The dispersion of the original spectrum of Steve had 3.4 Angstr./pix. For comparison I attached one of my last spectrum (2014-07-03) with a dispersion of 0.029 Angstr./pix.

Ernst

[robin_astro]

Coming back to the original post, wouldn't we expect to see the absorbtion dip  on H Alpha at this resolution?

Chris

Hi Chris,

Not at the ALPY resolution (~10A). The separation of the emission and absorption is ~5A. Additionally because the emission is much stronger than the absorption, the shoulder of the emission tends to bleed into the absorption  and fill it in even at higher resolutions.  Not sure at what resolution the absorption starts to be detected though.

Cheers

Robin 

[robin_astro]

Not sure at what resolution the absorption starts to be detected though.

Found this spectrum using the LISA (~5A resolution)  

http://www.spectro-aras.com/forum/viewtopic.php?f=17&t=165

The H alpha emission absorption still bleeds into the absorption at this resolution so still not  resolved  but the the absorption in the nearby redward He line is just visible, due to  the much lower He emission 

Robin

[robin_astro]

It is resolved though at ~2.5A resolution (eg L200 or LHIRES III with 600 l/mm grating)

http://www.spectro-aras.com/forum/viewtopic.php?f=17&t=466

Robin

[robin_astro]

It is resolved though at ~2.5A resolution (eg L200 or LHIRES III with 600 l/mm grating)

http://www.spectro-aras.com/forum/viewtopic.php?f=17&t=466

An interesting comparison and discussion of different EW measurements there too

Robin