What exactly is encoded in the telemetry strip of a NOAA APT image?

KE7IZL

Member
Joined
Aug 2, 2011
Messages
187
Location
Seattle, WA
I was reading this specification here https://www.ncei.noaa.gov/pub/data/satellite/publications/podguides/N-15 thru N-19/pdf/2.1 Section 4.0 Real Time Data Systems for Local Users .pdf where it lists the 16 fields of the telemetry strip of the image.
but it neglects to say how to actually INTERPRET the values there. Of particular interest to me is the Channel ID field.

I know that just as any other part of the image, each field contains a value between 0 and 255 because it is composed of pixels. However I also know that due to the analog nature of the image transmission, the range of values that exists in this field would need to be quantized into larger steps (like maybe all values between 0 and 4 should be interpeted as 0, values between 5 and 9 as 1, values between 10 and 14 as 2, etc). This way each value in the field could be represented as a range of pixel values so it would be more resistant to error (like just being one pixel value above or below the intended value would not result in a complete change in the value of the field's data).

However, I don't know how to interpret these fields. For one, I don't know what quantization steps are used to encode values in any of the fields, other than the brightness steps (which have 9 brightness values, starting at the darkest gray above black, going up to white, and then back down to black). The calculations of these are easy. They are step_val/8*255. The first 8 steps have step_val going from 1 to 8, and the ninth step has step_val at 0.

I don't know how the other telemetry fields are calculated though. This is not mentioned at all in the PDF file I linked to above. And that appears to be the only official APT specification that NOAA has made public. Has anybody managed to reverse engineer how to interpret the various telemetry fields in a NOAA APT transmission? In particular I'm interested in the Channel ID field.
 

merlin

Active Member
Joined
Jul 3, 2003
Messages
3,094
Location
DN32su
You would need software that will parse the data into its segments and decode that. Off hand, I know of no such animal.
 

KE7IZL

Member
Joined
Aug 2, 2011
Messages
187
Location
Seattle, WA
You would need software that will parse the data into its segments and decode that. Off hand, I know of no such animal.
I was interested in the specs, so that I could write such decoder software myself. Right now, there's a discontinued software called WXtoIMG that can do exactly that, and generate a composite image, based on exactly which cameras on the satellite were being used at the time of the transmission. For example, if the thermal imaging camera was in use, the software can incorporate temperature data into the composite image, while if the thermal imager wasn't being transmitted then temperature data can't be put into the composite image. So of course this software knows which cameras were in use at the time of the transmission because of the Channel ID that's in the telemetry strip.

I'm just trying to figure out the specs. And the specs published by NOAA are definitely INCOMPLETE.
 

radar_hunter

Member
Joined
May 29, 2023
Messages
40
I think the SatDump team had some plans to process the telemetry strip, but I'm not sure about this, and not sure if such a feature is already implemented.

But the channel ID field is not very hard to use manually. The brightness/intensity of the channel ID field should match one of the first six reference fields.

So in order to use the channel ID field, locate the 9 reference fields that show 8 increasing intensity levels and the 9th field is zero intensity. Then find the channel ID field and check which of the reference fields has the same intensity level as this one. Use the specification to find out which sensor channel was transmitted.

The other fields contain sensor calibration values (mostly temperatures I think) and the specification should (?) tell how to use this data.

If you are interested in the telemetry data, making actual measurements or anything other than just getting cool images, it's better to use a decoding program that follows the specification precisely. Not all decoding software do. Image post-processing may ruin the actual measurements, especially if it's not done according to the telemetry and specification. I have even seen some decoding software changing the brightness values of the telemetry strip when post-processing was used which instantly makes it useless for anything.

As far as I know the SatDump team aims to follow the official specifications, but you can discuss with them if you are interested in the details. Use a github ticket or join their matrix channel.

edit: From SigID Wiki:
 
Last edited:
Top