Jump to content


Photo

Unihedron SQM vs. Dark Sky Pro iPhone app

  • Please log in to reply
22 replies to this topic

#1 jchaller

jchaller

    Apollo

  • -----
  • Posts: 1350
  • Joined: 29 May 2008
  • Loc: Tenino, WA.

Posted 11 May 2013 - 03:23 AM

I have a Unihedron Sky Quality Meter (the SQM model) and decided to compare its readings to the Dark Sky Pro app (DSM) for the iPhone.

Probably not the best night to test - thin to thick cloud cover, which is going to amplify light pollution. Also, the
conditions weren't consistant as the clouds thined enough at times that I could see stars trying to peak through.


9:30
SQM - 15.51
DSM - 15.99

9:35
SQM - 16.40
DSM - 16.50

10:00
SQM - 19.00
DSM - 19.01

10:30
SQM - 19.6
DSM - 19.4

11:00
SQM - 19.7
DSM - 19.3

11:30
SQM - 19.6
DSM - 19.4

12:00
SQM - 19.7
DSM - 19.3

I took care to completly cover camera during calibration and pointed the camera at zenith each time I
took a reading that was used for comparison.

Typically my SQM readings range from 20.72 to 20.95 at my current location on a clear dark night.
I'll repeat the test when I get a good night.

#2 Tonk

Tonk

    Fly Me to the Moon

  • *****
  • Posts: 7082
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 11 May 2013 - 04:53 AM

Are you averaging a set number of readings for each measurement point? I suggest at least 5. Taking single readings at each point is subject to random variance issues

#3 richard7

richard7

    Not Quite

  • *****
  • Moderators
  • Posts: 5886
  • Joined: 02 Nov 2007
  • Loc: Sacramento

Posted 11 May 2013 - 09:17 AM

Before 10:00 the DSM shows darker readings and afterwards the SQM reads darker.
Is it some kind of bandwidth difference? About 10:00 is when drivers turn off their headlights and some outdoor lights are turned off maybe causing the different readings?

#4 jchaller

jchaller

    Apollo

  • -----
  • Posts: 1350
  • Joined: 29 May 2008
  • Loc: Tenino, WA.

Posted 11 May 2013 - 05:36 PM

I did not average the readings. I did take multiple readings with the DSM meter each time to ensure that I was getting good calibration at the start.

I live a couple of miles outside the nearest town, so there isn't much traffic, nor do I have many neighbors close to me.

I think the big variable here was the clouds. The following pic is what I believed the clouds looked like during the test. Took the pic this morning with my iPhone.

Attached Files



#5 jchaller

jchaller

    Apollo

  • -----
  • Posts: 1350
  • Joined: 29 May 2008
  • Loc: Tenino, WA.

Posted 12 May 2013 - 02:18 AM

Partly cloudy tonight, but I had nothing better to do, so I did some more comparisons.

Previously the DSM seemed to give higher readings than the SQM in lighter conditions, so I decided to start earlier to see if that would show itself more clearly. It seems the bottom of the scale on the DSM is shifted upward quite a bit.

I ended up with a large hole in the clouds at the 11:00pm readings, so I was able to get through the clouds.
Oh, I did take multiple readings and gave an average.

Attached Files



#6 Tonk

Tonk

    Fly Me to the Moon

  • *****
  • Posts: 7082
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 12 May 2013 - 07:05 AM

SQM are factory calibrated using standardised light sources. How is the DSM calibrated?

The difference between the SQM and DSM values is actually large in all cases as the values are powers

#7 Tony Flanders

Tony Flanders

    Voyager 1

  • *****
  • Posts: 11179
  • Joined: 18 May 2006
  • Loc: Cambridge, MA, USA

Posted 12 May 2013 - 07:57 AM

The difference between the SQM and DSM values is actually large in all cases as the values are powers


Well, yes and no. It seems clear that the DSM isn't as reliable as the SQM -- nor would one expect it to be, seeing as the cameras on cell phones are famously insensitive to low light levels because of their minuscule pixels.

On the other hand, being within 0.3 of the correct value is very useful data. It's far more accurate than people rating their skies by saying "I can see second-magnitude stars" -- which tells me nothing except that they're not looking very carefully.

Note also that magnitude per square arcsecond isn't really well-defined unless you specify what wavelengths are being measured. It's known that the SQM doesn't match the Johnson V band, and I doubt the DSM does, either. That's actually a pretty big deal, considering that light pollution is much redder than any of the major natural sources.

Consider that the SQM costs $120 and the app costs $5, I'd say it's a pretty good deal. However, it would be extremely helpful to know its limitations.

#8 REC

REC

    Fly Me to the Moon

  • *****
  • Posts: 5128
  • Joined: 20 Oct 2010
  • Loc: NC

Posted 12 May 2013 - 08:31 AM

Thanks for doing the comparison. I will keep this in mind when a I do some reading tonight.

Bob

#9 knightware

knightware

    Vendor (Knightware)

  • -----
  • Vendors
  • Posts: 153
  • Joined: 22 Mar 2006
  • Loc: North Carolina

Posted 12 May 2013 - 09:04 AM

On the other hand, being within 0.3 of the correct value is very useful data. It's far more accurate than people rating their skies by saying "I can see second-magnitude stars" -- which tells me nothing except that they're not looking very carefully.


I would quibble with "very useful" and suggest "somewhat useful". I agree that either device provides much better information than the visual estimate technique as described, but I have found that different Sky Quality Meter models with lens produce readings within a few hundredths of a magnitude/square arcsecond. Thus, 3 tenths is a significantly large difference.

I think the important point to be taken from the discussion is that when one quotes one's sky quality readings, it is extremely important to state what type of device produced the reading. In the case of the iPhone app, it might need to include the phone model (hardware changes) and the app version (software changes). This is not as important with the SQM since these are calibrated at the factory. That said, professional research done with SQM devices includes complete version information taken from the meter, including calibration information.

Having readings and device information together can lead to meaningful analysis of light pollution which is what we really want in the end.

#10 core

core

    Apollo

  • *****
  • Posts: 1428
  • Joined: 23 Feb 2008
  • Loc: Mostly in Norman, OK

Posted 12 May 2013 - 09:23 AM

Before 10:00 the DSM shows darker readings and afterwards the SQM reads darker.


On the other hand, it looks more like DSM has a different way of interpreting readings above/below ~18? (ie, when it's darker, and there's issues with cameras and low-light)

Jim, any chance of plotting the readings you get? May help to spot some trends with regards to DSM.

One thing I wish is that the DSM coders might share a little of how they go about getting their numbers (eg, cropping the image to get a 'reading', how much crop, 'binning', averaging algorithm, bypassing AGC in the iphone's camera module?, temp issues, etc), it might help understand how the app gets to the value it reports. It's not just that readings are 0.xx value off from a SQM reading, needs to be consistent and repeatable.

Attached is a plot from Jim's number from his original post.

Attached Files



#11 jchaller

jchaller

    Apollo

  • -----
  • Posts: 1350
  • Joined: 29 May 2008
  • Loc: Tenino, WA.

Posted 12 May 2013 - 11:34 AM

SQM are factory calibrated using standardised light sources. How is the DSM calibrated?

The difference between the SQM and DSM values is actually large in all cases as the values are powers


Prior to each reading the user sets the dark point by covering the camera and press (1)DARK. When the app is done measuring the dark point, (2)SKY is highlighted. Point the camera at the sky and press (2)SKY.

The app does allow the user to shift/stretch the readings via app settings.

Attached Files



#12 GlennLeDrew  Happy Birthday!

GlennLeDrew

    Voyager 1

  • *****
  • Posts: 10643
  • Joined: 17 Jun 2008
  • Loc: Ottawa, Ontario, Canada

Posted 12 May 2013 - 12:50 PM

In cases of patchy brightness (as when clouds are present), larger differences between device readings could/should be expected.

If a uniform sky is not available, you could test under controlled conditions indoors. An LED flashlight, placed across a dark room and masked to the required degree will do. Make sure that any light reaching the device is well diffused, with no 'hot spots'. To this end, the light could be aimed at a sheet of paper, or the wall. The light is further diffused by a sheet of paper placed near to the device, the paper working in 'rear projection' mode (this obviates any self shadowing which might occur in 'front reflection' mode.)

Considering the fairly wide fields covered (~80 degrees for the non-lensed SQM), place the devices within a few inches of the paper. The distance can be allowed to vary, as surface brightness does not change with distance; as long as the sheet fully fills the device's field is all that matters.

#13 jchaller

jchaller

    Apollo

  • -----
  • Posts: 1350
  • Joined: 29 May 2008
  • Loc: Tenino, WA.

Posted 12 May 2013 - 12:55 PM

Jim, any chance of plotting the readings you get? May help to spot some trends with regards to DSM.


Here's a plot from my latest outing.

Attached Files



#14 H.A.Treur

H.A.Treur

    Vendor-Dark Sky Pro.

  • -----
  • Posts: 2
  • Joined: 13 May 2013

Posted 13 May 2013 - 06:44 PM

Good evening to all.

A very short introduction: Harro Treur from Netherlands and a newbie on this forum. I have assisted Mr.Norbert Schmidt with the realization of the DarkSky Meter-app.

We have tried the best we could to make the iphone "SQM-worthy". The app has been calibrated by using 5 different SQM-devices and more than 5000 measurements.

This was not easy, because the Apple-company refused to cooperate with this project. Apple did not allow us to have any control of the camera shutter speed. In spite of that, Mr.Schmidt found a way to make the iphonesensor detect enough light even with short exposures.

At the moment, the app is optimized for iphone 4S. Iphone 5 is a bit off from SMQ 21 and up.
Above SQM 21,5 the fluctuations on both iphones will increase.
Above SQM 22,5 the phone reaches its limits and will not detect enough light anymore.
The iphone 4 and ipad camera's are not sensitive enough at all above SQM 20, but...

This app is still a work in progress. We are currently busy with these issues. In coming weeks we will publish an updated version of the app.

One important note: the Field of view of the DSM-app is only 10 degrees (comparable with the SQM-Lens). Therefore, measurements taken with an oldschool SQM-device (80 degrees FOV) will show differences, especially when the sky is not uniformly lit (i.e. clouds, milky way or lightdomes).

Please keep up these measurements, but preferably with a SQM-Lens.

Best regards
H.A. Treur

#15 jchaller

jchaller

    Apollo

  • -----
  • Posts: 1350
  • Joined: 29 May 2008
  • Loc: Tenino, WA.

Posted 13 May 2013 - 09:34 PM

I wish I had the SQM-L, but I don't.

I was planning on doing one more comparison when I get a clear dark night, but from the looks of it that could be awhile.

I don't think I can test above 21, as the highest reading I have seen at my present location was 20.95, but that might not be repeated. Although, I should be able to get above 21 by traveling a few miles further east though, which I might do at some point.

I am using an iPhone 4s.

#16 Nop

Nop

    Vendor (iPushto)

  • -----
  • Vendors
  • Posts: 115
  • Joined: 12 Feb 2006
  • Loc: Netherlands, Valkenburg

Posted 14 May 2013 - 03:24 AM

Yes, we use binning, cropping and stacking, substracting, even gamma correction.
The iPhone is different from a Unihedron SQM. It has no temp meter and a different kind of sensor. And we have the limitation of exposure control. That's why we can't support iPhone 3GS or iPad 2 because of the limited light gathering abilities for short exposures. However, with millions of iPhone 4S and 5's around and a 'lite' version aimed at a general audience we can put a SQM in the hands of thousands of users worldwide. That generates a lot of useful data for the IDA and scientists, but also for amateur astronomers looking for suitable skies. A map containing all the (valid) measurements will be up soon. This is our development map:

Attached Files



#17 Tony Flanders

Tony Flanders

    Voyager 1

  • *****
  • Posts: 11179
  • Joined: 18 May 2006
  • Loc: Cambridge, MA, USA

Posted 14 May 2013 - 04:16 AM

However, with millions of iPhone 4S and 5's around and a 'lite' version aimed at a general audience we can put a SQM in the hands of thousands of users worldwide. That generates a lot of useful data for the IDA and scientists ...


Exactly. That's why I said being within a few tenths generates very useful data. What it lacks in accuracy it potentially makes up in quantity.

For the first time, one can imagine actually blanketing a metropolitan area with skyglow readings. That would be a tremendously valuable check for theoretical models.

#18 Dig

Dig

    Lift Off

  • -----
  • Posts: 7
  • Joined: 28 Oct 2010
  • Loc: Cary, NC USA

Posted 15 May 2013 - 11:01 AM

Norbert,
For the iPhone4s and 5 it would be great if we could off load the data instead of having to manually transcribe our collections by hand. Also, is there anyway we could edit the data as well? If I for get to add the weather report I'd like to add it later.

#19 Dig

Dig

    Lift Off

  • -----
  • Posts: 7
  • Joined: 28 Oct 2010
  • Loc: Cary, NC USA

Posted 15 May 2013 - 11:03 AM

... I should add that it would nice for the app to state what kind of device is making the report.

#20 Nop

Nop

    Vendor (iPushto)

  • -----
  • Vendors
  • Posts: 115
  • Joined: 12 Feb 2006
  • Loc: Netherlands, Valkenburg

Posted 15 May 2013 - 01:09 PM

The maps and underlying database with measurements will be available for everyone within a few months. This is a snapshot of the raw, unfiltered data submitted:
Posted Image

#21 iam1ru12

iam1ru12

    Mariner 2

  • *****
  • Posts: 225
  • Joined: 25 Feb 2006
  • Loc: Raleigh/Durham, NC

Posted 16 May 2013 - 05:57 AM

Norbert/H.A.,
Keep up the good work. I have the Pro version and looking forward to updates as they come available.

At $130 plus (in the US) for the SQM-L, that's more than I'm willing to pay to measure light-pollution. However your Pro version at $3.99 is an phenomenal option! Thanks for coming out with this app.

BTW, a fellow club member shared news of your app with the club a few weeks ago. Several of us have purchased the app and we've had some great conversations and comparisons.

Thanks!
-Mike

#22 Jon Isaacs

Jon Isaacs

    ISS

  • *****
  • Posts: 43855
  • Joined: 16 Jun 2004
  • Loc: San Diego and Boulevard, CA

Posted 17 May 2013 - 09:32 AM

Consider that the SQM costs $120 and the app costs $5, I'd say it's a pretty good deal. However, it would be extremely helpful to know its limitations.



It would seem that for light polluted skies it works quite well. If the skies are quite dark, I would think that would be where the problem lies.

Jon

#23 Tonk

Tonk

    Fly Me to the Moon

  • *****
  • Posts: 7082
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 17 May 2013 - 09:41 AM

Depends what you are using the data for. I'm doing a long term survey of the LP profile of my city along a 15 mile section from city center to countryside - its been going 5 years now. Hence I need good calibrated values to make year on year comparisons meaningful.

If you are just interested in estimating how long to expose your camera for DSO imaging etc then a roughly OK value is all that you need. I have created a table of MPSAS values vs camera exposure time vs ISO for my imaging setup






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics