Telescope Live Bad Data? Other · Quinn Groessl · ... · 10 · 624 · 1

I can't speak to the quality of narrow-band data, but I have processed quite a few of their RGB images.  I posted on their data previously.  Here is my post:  https://www.astrobin.com/forum/c/astrophotography/other/telescope-live/?page=1#post-150353

The single point that is important to realize about their data is that one of their "one-click observations" is generally not enough.  I usually download at least three and sometimes one of their bundles that may have hours of exposure.  A single bundle will often be noisy, and sometimes, there are not enough subs to eliminate satellite tracks and other anomalies.  

But overall, I like the service: reasonable prices and a good variety of objects.   The latest version of Blurxterminator (AI4) can do wonders to correct star problems in images from their FSQ106 scopes.  Game changer for me. 

George

I had a data set that had hexagonal stars (one of the Australian data sets I think).  I texted support and pretty much got a "not much I can do about it"  Didn't love the fact that there was no discount on the next set or anything.  
I think when I pay for a subscription like that I should get premium data that someone reviews and is calibrated well.  
Why pay for garbage data when I can get that from my own backyard?
That being said, it is a rare occurrence thankfully.  
Did you blink the subs that you downloaded?  I have chucked a couple in the bundles not one clicks.

As with all paid data; "Caveat Emptor".  I am a member of the TL site and have been a member of Roboscopes and used Chilescope.

In my experience, the consistently best data was from Chilescope. Refunds were alos readily given for poor quality data, including those taken in poor seeing.  But in this case I was doing the Quality Control (QC) myself and the data was uniquely mine.  

In the One-Click observation model [TL] and consortium [Roboscopes], QC is either absent or very cursory before offering the data to the paying public. 

In many ways, this is understandable since these models are *very* inexpensive [compared to the cost of bespoke observations].  No doubt this is only possible by minimising salary costs associated with the models.  QC is very expensive in this regard.  Either in assessing the conditions at the time of observation, the calibration process or the pipeline output.  I note that both Roboscopes and TL have much more expensive bespoke observation models, where I presume the customer can demand a refund if the quality of the data falls below a certain threshold and where more QC may exist as the data is taken. [Refunds are also an expensive business in terms of staff time]. 


Despite the apparent lack of QC,  I have found that about 75% of the One-Click TL data is good. The depends on how stringent one wishes to be, and which telescope one uses.  I always use the One-Click observation bundles as single One Click sessions simply have too few frames to make a decent image.  Making a decision on TL based solely of single One Click observations probably won't give you the best idea of whats really on offer. 

I always use Blink and SubframeSelector to weed out bad frames. A lot of the data is taken with a bright moon too, reducing the efficiency, but not necessarily providing bad data.]. I principally use TL data to supplement my backyard observations [L band data to improve seeing, H band data to correct ghosting]. Almost exclusively I use data from the Chilean CDK24.  I have found the Australian and Spanish data of lower quality on average.

However, that is not to downplay the data from TL.  In the 75% of cases where the data is good - it is *really* good.  It works out about $1/hour of data.  On a CDK24 this is an amazing value-for-money bargain.   Even allowing for the poor data [and the fields that are advertised DSO Panel 1, DSO Panel 2… which don't mosaic together.  Grrr!], it still works out at a bargain $1.50/hour.     

If I were to buy the kit myself it would take me at least 150000 hours to break even on the capital cost with the .  Around 50 years observing in a location with no loss due to weather.  I won't live that long, and such a place doesn't exist. So I am very happy to put up with some bad data from the One-Click  observation bundle model.

Messierman3000….Could these vertical lines be bad columns in CCD detectors?  Most of the images from Telescope.live come from CMOS cameras (e.g., QHY600M), but some of the older data can come from CCD cameras with bad columns.  I've found that these lines can be removed using CosmeticCorrection in Pixinsight.  

George

Most of the time the data is good, and on rare occasions of poor focus or bad seeing conditions they always refunded me. They have this live chat window where you can text the support. They reply very quickly, almost instantly.