44 replies to this topic

[WheezyGod]

I can’t remember if this has been brought up before but has anyone compared two devices with lower vs. higher EBI as well as more vs. less narrow ha filters?

Everyone here has preferences and some of that differs based on how light polluted the site they tend to view from is. Non-astronomy NV use has suggested EBI can be important at a dark site which got me wondering if the preference here for how narrow a Ha filter is has nearly as much to do with EBI as it does with light pollution.

Most here in heavily light polluted skies prefer very narrow ha filters. Even in my more moderately light polluted, bortle 5/6 skies I had preferred the view from a 3.5nm ha filter over the 6.5nm ha filter I had. This was during the winter when my device’s EBI was between 0.0-0.1. Now during the summer I wonder if I would have the same preference. It’s possible that my device’s 0.7 EBI would mask some or most of the incremental benefit which could be why many here prefer something around 7nm instead of 3nm if observing from more moderately light polluted skies. As another example, Gavin prefers more narrow ha filters at darker sites, but one of his devices has an EBI of 0.1.

Anyways, has anyone with 2 or more devices done any comparisons for what I’m thinking about?

[sixela]

A device with a higher EBI will indeed make you want to either reduce the f/ratio or make the filter wider to keep the background above the "maximally grainy" one where EBI plays a majore role and becomes very distracting. The darker the site, the easier it is to make that background "too dark", and since the contrast ratio is going to be better at that site anyway, you might prefer a wider filter even though in theory it gives "less contrast".

 

If you're not using aggressive filtering, then the background is always bright, even at the darkest of the darkest site.

[Speedy1985]

Interestingly, I was just having a conversation with a new NV user who got a tube with great specs, with an EBI of 0.2 or 0.3. He lives in the next town over and we’re going to try and do a 3x side by side comparison with both units and different filters. My EBI is  0.6. 
 

In my experiences, and most recently with my big dob in a sky similar to yours at B5-6, I found the views of some objects to be more pleasing with a 6.5nm filter. The 3.5 just seemed to darken everything to the point of the object losing some definition as well as contrast. But there are so many variables, not just in each unit, but in each person’s eyes too. 

[sixela]

And in the effective f/ratio you're using at the photocathode and the surface brightness of the object itself...

Edited by sixela, 28 August 2024 - 02:50 PM.

[WheezyGod]

Interestingly, I was just having a conversation with a new NV user who got a tube with great specs, with an EBI of 0.2 or 0.3. He lives in the next town over and we’re going to try and do a 3x side by side comparison with both units and different filters. My EBI is 0.6.

In my experiences, and most recently with my big dob in a sky similar to yours at B5-6, I found the views of some objects to be more pleasing with a 6.5nm filter. The 3.5 just seemed to darken everything to the point of the object losing some definition as well as contrast. But there are so many variables, not just in each unit, but in each person’s eyes too.

I didn’t try it in my scope since it was a 1.25in filter I had borrowed for a few weeks, but tried it (3.5nm) handheld and saw more Halpha detail during the winter compared to the 6.5nm filter. It wasn’t a significant difference, but I could see the entire Barnard’s loop at 4x and f/1.8 with the 3.5nm, but couldn’t see part of it was the 6.5nm. Angelfish also stood out more.

It was enough of a difference that I’m considering getting a 3.5nm at some point, but it’s also hard for me to justify the expense for a small improvement when I don’t observe handheld as much as I do with my telescope.

In the summer though I might not notice a difference or as much of one because EBI could cancel out the benefit.

[Souldrop]

I got you fam. I just got in from a quick poke around. Apologies not a critical session, just a quick session with some notes. 

 

Comparison setup:

Scope/BT: Vixen BT-70EDS (f/5.7)

Eyepieces TV 40mm Plossl

Filters: Antlia 4.5nm Edge; Astronomik

 

Sky characteristics:

Bortle 4/5

SQM ~20.7 

 

Observation Targets:

NGC6530 and M20

 

NV Specs:

Tube 1 (L3 Unfilmed):

SNR: 35.9

Res: 72

EBI: 0.9

Halo: 0.7

PCR: 2159

 

Tube 2 (L3 Unfilmed):

SNR: 34.6

Res: 72

EBI: 0.5

Halo: 0.7

PCR: 2266

 

Tube 3 (L3 Unfilmed):

SNR: 38.6

Res: 72

EBI: 3.9 (Not a typo)

Halo: 0.7

PCR: 2727

 

Tube 4 (L3 Unfilmed):

SNR: 36.7

Res: 72

EBI: 1.5

Halo: 0.7

PCR: 2717

 

NV Prep before observing:

 It was in the high 90's when I placed the NVD units out to warm up earlier this afternoon around 5PM local time before running some errands. It had cooled to mid 80's by the time I began observing, so the temperature effects on EBI should represent a pretty sub-optimal scenario.

 

Test setup:

 Each side of the BT had a different filter. One was an Astronomik 12nm and the other was an Antlia Edge 4.5nm. I would take about ~ minute turns at each eyepiece to get a feel for the differences. I then would cycle through the the different NV units. I did this multiple times in various orders. I would twiddle with the gain to reach what I felt was an image with the most relative contrast. Typically this was near full gain with the 4.5nm, but the 12nm I found the gain settling between 70-90% depending on the SNR of the tube being used. In addition Tubes 3 and 4 were compared at 1x with the filters in front of the objective.

 

Conclusion: At least for the relatively quick sessions with these targets the wider filter was preferred no matter the NV unit when used in the BT. More stars were visible. More structure of the targets was more easily seen.

 

Although it was mixed up a tiny bit when hand holding at 1x. I found I preferred the narrow filter for Tube 3 which has the worst EBI by far. Contrast for extended objects was noticeably better and they more easily popped out. For the lower EBI unit I preferred the the wider filter. The narrow filter felt like it was taking too much signal away from the image.

 

Apologetics:

I'm not what I consider a good astronomer, and coupled with the relatively quick observing session someone else could very well come to different conclusions. Additionally the results likely depend on the target in question I just stuck my eyeball towards the milky way and happened to land on the targets I did. 

[sixela]

The narrow filter felt like it was taking too much signal away from the image.

That is, of course, an incorrect deduction. A more narrow filter does not "take away signal", it takes away the background, which should normally count as 'noise' not 'signal'. But added background does help make the object's surface brightness larger (since you see the background plus the object), although it reduces contrast.

I'm not that surprised, BTW: with a 40mm eyepiece on an f/5.7 scope you're working at f/3.8, and many H-alpha objects will already be quite dark if you don't raise them with some background added to them. At that f/ratio (which for me us using the NVD in prime focus) I sometimes tend to prefer my 4nm Altair dualband filter *even* on objects that don't have a lot of OIII emissions (where it functions more or less as an 8nm H-alpha filter, the OIII band just adding some background).

 

Sometimes getting your eyes really well dark adapted (should be doable with SQM 20.7) will let you enjoy the darker image with better contrast of more narrow filters, BTW. Especially if you're an experienced glass eyepiece visual observer used to low surface brightness observation. You might need slightly averted vision even more, though.

Edited by sixela, 29 August 2024 - 02:45 AM.

[Souldrop]

That is, of course, an incorrect deduction. A more narrow filter does not "take away signal", it takes away the background, which should normally count as 'noise' not 'signal'. But added background does help make the object's surface brightness larger (since you see the background plus the object), although it reduces contrast.

I'm not that surprised, BTW: with a 40mm eyepiece on an f/5.7 scope you're working at f/3.8, and many H-alpha objects will already be quite dark if you don't raise them with some background added to them. At that f/ratio (which for me us using the NVD in prime focus) I sometimes tend to prefer my 4nm Altair dualband filter *even* on objects that don't have a lot of OIII emissions (where it functions more or less as an 8nm H-alpha filter, the OIII band just adding some background).

Sometimes getting your eyes really well dark adapted (should be doable with SQM 20.7) will let you enjoy the darker image with better contrast of more narrow filters, BTW. Especially if you're an experienced glass eyepiece visual observer used to low surface brightness observation. You might need slightly averted vision even more,

I have to disagree with your first paragraph at least a tiny bit because it does also cut the amount of light of potential h alpha hitting the cathode. This very clearly caused some portions of extended objects to get lost in ebi washout. For an extended h-alpha object I wouldn’t expect a narrow line emission due to body motion (towards/away from us as well as motion within itself) so a more narrow filter could very well be cutting parts of the structure. This could be exacerbated by other mechanisms broadening the light profile. A narrow filter is creating a tighter bound on the relative velocities that would allow a “pixels” h-alpha emission through. Just checked and NGC6530 is gravitationally unbound so not outside the realm of reason that some parts would find themselves doppler shifted out with a narrow filter. This could also be exacerbated by other factors that may cause broadening of the emission line. ETA: not the mechanism here. Not moving fast enough to be close to the cutoff imposed by even a 4.5nm filter

The units that were testing at 1x had high EBI values and were operating in a warm environment. The incredibly high EBI unit showed less structure all around. Anything bright enough to survive the EBI washout likely didn’t really need as much background boost, so that preference could be a quark to that unit. I’ll do comparisons with a lower EBI unit the next evening I have some time.
The last paragraph is completely fair. Summer is not the time I enjoy getting out long enough to get really adapted not to mention I do not consider myself an experienced visual observer, so definitely take my above post with that consideration in mind.

Edited by Souldrop, 29 August 2024 - 08:19 AM.

[sixela]

I have to disagree with your first paragraph at least a tiny bit because it does also cut the amount of light of potential h alpha hitting the cathode.

Barely. A good filter has a transmission over 90%, sometimes as high as 97%. And a wider filter usually has exactly the same transmission on the emission line, so the net difference is exactly ZERO.

Edited by sixela, 29 August 2024 - 08:08 AM.

[Souldrop]

Nah wasn’t discussing transmission percentages. I did edit my above response as I was working from a wrong asdumption for ngc6530 at least, but in theory the above point stands for various objects in the sky. 

[sixela]

A wider filter doesn't "transmit more H-alpha" because the H-alpha emission is extremely narrow-band, much more narrow than the passband of even a 3nm filter. 

If you'd observe H-alpha objects in very distant and redshifted galaxies theoretically could indeed make a very narrow filter miss the H-alpha emission, but for the objects we usually observe that is not a real scenario.

 

I do have a faster scope than you (f/3.72) and couple it with a 3nm filter there is already a bit of extra H-alpha loss on the H-alpha regions of M51 (the redshift is 1nm for H-alpha, so it's still in-band for slower scopes but pushes transmission a bit lower on faster ones). And yes, I also have a 6nm filter (also with fairly large effective refractive index) just for corner cases like that. But that's what they are, corner cases, and it does not apply to your observations of Messier 8 and 20, which means the explanation for that is different.

 

But NGC 6530 or anything in this galaxy is unaffected --I think the shift is inferior to 0.1 nm--, and M33 (which has a lot of interesting H-alpha) has a slight blueshift that is actually beneficial on fast scopes ;-). The Magellanic Clouds (also with lots of great H-alpha) have redshift that are also in this neighbourhood.

Edited by sixela, 29 August 2024 - 09:23 AM.

[Souldrop]

TL;DR the universe is a huge, dynamic entity. There are several mechanisms which could cause the observed emission profiles to have some energy lost when using a more narrow filter, so I do stand by the argument that a narrow filter can cause some "signal" to be lost. Are you stating all observed h-alpha emission is a narrowline (<4nm FWHM)  centered at 656nm?

 

Whether it's the case here? I'm not willing to spend the time to confirm or deny. I could also be overthinking things and everything last night was more a lesson on perception in which case I'm more than happy to defer to you and others. WG mentioned comparing units with varying EBI with different bandwidth filters. Granted I don't have any incredibly low EBI units I felt the observation notes I made were worth pursuing and sharing. Last night was a shotgun approach. I'm more than happy to do more a targeted comparison if no one else decides to chime in.

Edited by Souldrop, 29 August 2024 - 09:05 AM.

[sixela]

I'm not doubting your observations -- if you read above I also state that at f/ratios like the one you use, I frequently also use less narrow filters. But that's to raise the surface brightness slightly by adding background, not by getting "more H-alpha", and that works well with a dualband 4nm *and* a 6nm H-alpha (only one of which could get some 'extra H-alpha').

 

In the case of the dual-band, of course, on an object like Messier 8 the bonus is that you make it brighter also by picking up extra signal instead, i.e. not only some more background but some OIII emission too (assuming that your NVD plus objective combination is at least slightly sensitive to it).

 

When I use much lower f/ratios, though, I tend to prefer the more narrow filter, even at a dark site. I do always compare them before I pick the 'final' one, since I have a filter slide (for nebulae it has a 6nm Astronomik H-alpha, a 4nm dual-band and a 3nm Chroma H-alpha).

 

FWIW: IIRC, generally the width of the H-alpha emission line for nebulae corresponds to a speed of around a few kilometers per second in terms of velocity for the emitter, which translates to a FWHM of about 0.1 to 0.2 nanometers. I think I did look it up in the past, just because I was wondering the exact same thing you were.

Edited by sixela, 29 August 2024 - 09:43 AM.

[Souldrop]

I'm not doubting your observations -- if you read above I also state that at f/ratios like the one you use, I frequently also use less narrow filters. But that's to raise the surface brightness slightly by adding background, not by getting "more H-alpha", and that works well with a dualband 4nm *and* a 6nm H-alpha (only one of which could get some 'extra H-alpha').

 

In the case of the dual-band, of course, on an object like Messier 8 the bonus is that you make it brighter also by picking up signal instead, i.e. not only some more background but some OIII emission too.

I think we are talking past each other a fair bit the past few posts. I will drop my end and leave it to other readers to suss they will of it. 

 

To get back on track I did take your initial response to heart. I hope to have the opportunity to do a more critical session with your post in mind here in the near future. Last night given it was so hot seemed like a good opportunity for really stressing the differences in EBI, so my focus was rushing to get comparisons in before it got too much cooler. 

[sixela]

Yeah, observational data is good, regardless of the theory behind what works and what does not. Sadly I can't really do your experiment, my devices have fairly comparable (and low) EBI, I can only get a gut feeling of when things start to go dark enough in the background for the EBI to really mess it up.

Edited by sixela, 29 August 2024 - 09:45 AM.

[ButterFly]

I can’t remember if this has been brought up before but has anyone compared two devices with lower vs. higher EBI as well as more vs. less narrow ha filters?

Everyone here has preferences and some of that differs based on how light polluted the site they tend to view from is. Non-astronomy NV use has suggested EBI can be important at a dark site which got me wondering if the preference here for how narrow a Ha filter is has nearly as much to do with EBI as it does with light pollution.

Most here in heavily light polluted skies prefer very narrow ha filters. Even in my more moderately light polluted, bortle 5/6 skies I had preferred the view from a 3.5nm ha filter over the 6.5nm ha filter I had. This was during the winter when my device’s EBI was between 0.0-0.1. Now during the summer I wonder if I would have the same preference. It’s possible that my device’s 0.7 EBI would mask some or most of the incremental benefit which could be why many here prefer something around 7nm instead of 3nm if observing from more moderately light polluted skies. As another example, Gavin prefers more narrow ha filters at darker sites, but one of his devices has an EBI of 0.1.

Anyways, has anyone with 2 or more devices done any comparisons for what I’m thinking about?

It's a matter of contrast to noise ratio - the Rose Model, again.  More goes into the device's contrast to noise ratio than just EBI, so if you want to focus only on EBI, everything else needs to be the same (such as SNR, sensitivity, the optics' MTF, ...).

 

When going to narrower filters, the target's light is entirely unaffected.  An object needs a velocity dispersion of about five hundred km/s to shift h-alpha by 1nm.  (500 / 299792) * 656.3nm.  That is insanely high for nebulae, so that doesn't really matter, even for the 1.5nms either way for a 3nm FWHM filter.

 

When going to narrower filters, the background light (which is actual signal reaching the device) gets lower.  The target light does not.  All that reduced signal reaching the device (again, including background as actual signal reaching the device) means more scintillation.  That's what affects the contrast to noise ratio.  The contrast of the signal (regions of background only compared to regions of object plus background) has to me made out against more noise from the device (scintillation).  For any given f/ratio, going to a narrower bandpass will always decrease the contrast to noise ratio, simply because devices scintillate more with less signal (again, signal here is the actual light that gets to the device, including the very real background).

 

The change in the amount of scintillation as signal decreases doesn't only depend on EBI.  EBI is merely an effective noise floor for NO signal.  All that a lower EBI can do is place a floor on the effective background illumination.  That says NOTHING about scintillation of actual signal (yet again, sky background is actual signal reaching the device), or how it changes with lowering signal.

 

This is a matter of taste.  Even though there will always be lower contrast to noise ratio with decreasing bandpass, for a given device in a given setup, whether that matters for you depends solely on you.  Some don't mid the fizzles as much as other do.  Some can "see through" the fizzle better than others can.  If you don't like the results of a given bandpass with your device, when used in some particular setup, don't use it.  If you find yourself never using it, get rid of it.

 

If you insist on making a comparison between EBI and bandpass, keep everything else constant.  That is nearly impossible to do with real tubes.  Increased scintillation with decreasing bandpass is what's affecting whether you can "discern a target against the background", and EBI has nothing to do with how much scintillation changes with decreasing light.

[ButterFly]

A little side note on velocity dispersion's affect on bandpass.

 

Doppler shift is affected by both radial (line of sight) and transverse velocities, when special relativity is taken into account.  The generic z ~ v / c above only applies when v << c, and only for the radial (line of sight) direction.  For v << c, z ~ 1/2 (v/c)^2 for transverse motions.  Both of these come from approximating the gamma factor of special relativity, which is affected by both radial and transverse velocities.

 

In practice, imagine a toy model nebula as a uniformly expanding spherical shell.  If one side of the shell is coming toward you at 250 km/s, and the other is going away from you at 250 km/s, then the velocity dispersion in the radial direction is 500 km/s.  Nonetheless, each line of sight side is still only moving at 250 km/s, so each line of sight side only shifts by about 0.5nms.

 

In reality, we are seeing all parts of the nebula whose light gets through other parts of the nebula.  For that same toy model nebula, if we can see the plane that cuts through the middle of the nebula, there is NO line of sight motion at all.  It's all entirely transverse motion in that plane cutting through the middle.  The redshift there is about 0.5 * (250 / 299,792) ^2 * 656.3nm at h-alpha ~ about two milliAngstroms.

 

Apply this to motions of solar flares at the limb of Sun.  For things that go straight outward, the light is still well within the sub-Angstrom band.  What about the tops of curving arcs of solar flares though?  Look up some of the radial velocities for those.

[Souldrop]

Radial “body” velocity does matter. Thats going to shift your observed “centerline” emission.

Yes, as you pointed out the magnitudes of macro velocities we are looking has a marginal impact, but there is also some doppler broadening due to molecular speed distributions. Earlier in the thread I had a hard time believing there weren’t situation where some h alpha emissions wouldn’t be filtered out. It was an off the cuff thought/theory without digging into relative velocities of the more observable h-alpha objects. Until that moment I had not really ever thought about what order of magnitude for velocities to expect. Even though this wouldn’t really come into play until observing objects with ~1000km/s it was fun to think about and gave me an excuse to look into the properties of a couple different objects.

[Souldrop]

While thinking on it lets also not forget potential Nii emission which would be cut off by a narrower filter.

[ButterFly]

Radial “body” velocity does matter. Thats going to shift your observed “centerline” emission.

Yes, as you pointed out the magnitudes of macro velocities we are looking has a marginal impact, but there is also some doppler broadening due to molecular speed distributions. Earlier in the thread I had a hard time believing there weren’t situation where some h alpha emissions wouldn’t be filtered out. It was an off the cuff thought/theory without digging into relative velocities of the more observable h-alpha objects. Until that moment I had not really ever thought about what order of magnitude for velocities to expect. Even though this wouldn’t really come into play until observing objects with ~1000km/s it was fun to think about and gave me an excuse to look into the properties of a couple different objects.

Again, 500 km/s of radial motion gives a 1nm bandshift.  Even though the Sun is orbiting the center of the Milky Way at about 230 km/s, the vast majority of nebulae we're going to see are in our neighboring arms.  The relative radial motion between us and those nebulae is much, much smaller.

 

 

... but there is also some doppler broadening due to molecular speed distributions.

This is a joke.  The broadening is about sqrt(T) for a Maxwell distribution.  For the 10,000K of the Orion nebula, that is a broadening of about a factor of 100.  So, what's the FWHM of h-alpha emission, so that it's broadened by a factor of 100.  Recall that its an atomic transition, and not complex molecular transitions.  Broaden it by a factor of a million, and it still won't matter.

[Souldrop]

Again, 500 km/s of radial motion gives a 1nm bandshift.  Even though the Sun is orbiting the center of the Milky Way at about 230 km/s, the vast majority of nebulae we're going to see are in our neighboring arms.  The relative radial motion between us and those nebulae is much, much smaller.

 

 

This is a joke.  The broadening is about sqrt(T) for a Maxwell distribution.  For the 10,000K of the Orion nebula, that is a broadening of about a factor of 100.  So, what's the FWHM of h-alpha emission, so that it's broadened by a factor of 100.  Recall that its an atomic transition, and not complex molecular transitions.  Broaden it by a factor of a million, and it still won't matter.

You do you friend. No wonder so few people contribute to this forum. I will try my best to refrain from publicly sharing any more “jokes” in the future. 

 

I looked at the spectra for some objects of interest. I felt validated that a narrow band is indeed cutting signal of interest and not just background signal 

 

I will take my leave of this thread to no longer “sully” the science. 

Edited by Souldrop, 04 September 2024 - 07:26 AM.

[ButterFly]

Telling people to avoid narrowband filters because of nebula motion is indeed a contribution.

 

 

I looked at the spectra for some objects of interest. I felt validated that a narrow band is indeed cutting signal of interest and not just background signal 

Which one?

[Souldrop]

Telling people to avoid narrowband filters because of nebula motion is indeed a contribution.

 

Which one?

Stay your blade keyboard warrior. I never told anyone to avoid anything. I provided my subjective thoughts and initial musings when comparing multiple units with 2 different filters if that general dialogue is an unwanted contribution I'm more than happy not post further in this sub forum.

 

Everything you're arguing was already hashed with Alexis. I conceded then that my initial musings were incorrect. I valued his "constructive" input.

 

I guess you couldn't pass an opportunity to post your own critique. Some individuals do seem to feel those compulsions. FWIW I feel I even acknowledged where my initial, base assumptions were incorrect directly in response to you. Quote by me to you below:

Earlier in the thread I had a hard time believing there weren’t situation where some h alpha emissions wouldn’t be filtered out. It was an off the cuff thought/theory without digging into relative velocities of the more observable h-alpha objects. Until that moment I had not really ever thought about what order of magnitude for velocities to expect. Even though this wouldn’t really come into play until observing objects with ~1000km/s it was fun to think about and gave me an excuse to look into the properties of a couple different objects.

 

 

Last post for realsies since I should have provided examples of spectrum I felt validated my argument that a narrow filter is cutting some "target signal" (not background). Some examples where their spectrums indicate a narrow h-alpha filter would decrease signal of interest: M57, NGC 1952, IC 418, NGC 6888, and M42. There's probably more. 

Edited by Souldrop, 04 September 2024 - 01:23 PM.

[Thierry Legault]

Last post for realsies since I should have provided examples of spectrum I felt validated my argument that a narrow filter is cutting some "target signal" (not background). Some examples where their spectrums indicate a narrow h-alpha filter would decrease signal of interest: M57, NGC 1952, IC 418, NGC 6888, and M42. There's probably more. 

 

I really doubt about that, such finding has never been highlighted during decades of CCD/CMOS narrowband imaging, despite the fact that digital imaging provides extremely precise measurements (not affected by guessings from spectrums or by feelings )

[ButterFly]

Much more pressing for us is that we don't stack anything.  Our real time sky backgrounds, even from Atacama, severely limit how much of those tails we can even reach.  Even if those tails get really, really broad, that's not any target light that will ever have any chance of getting close enough to our backgrounds to matter.

 

But even when stacking, the sky background doesn't get reduced to zero.  At best, stacking gets rid of uncorrelated noise.  Some of that background is correlated, even when stacked over long integration times.

 

As a general rule of thumb, we can make out things that are about two mpas higher (as in dimmer) than backgrounds.  That applies regardless of whether we're looking at an images on a screen or through an eyepiece.  Stacking just artificially reduces that background as much as it can.