diff --git a/src/Glossary.tex b/src/Glossary.tex
index 32b6721..230e4d9 100644
--- a/src/Glossary.tex
+++ b/src/Glossary.tex
@@ -660,3 +660,8 @@
 % network
 % noise
 % workstation
+% mean
+% standard deviation
+% max
+% timestamp
+% frame
diff --git a/src/SNOUG.tex b/src/SNOUG.tex
index 8ebd43c..47c6564 100644
--- a/src/SNOUG.tex
+++ b/src/SNOUG.tex
@@ -361,7 +361,7 @@ leftmargin=1cm,rightmargin=1cm
 \chapterconf{Solve}{Plate Solvers}{Place from Pixels}
 \chapterconf{Detect}{Detect}{Find Satellites}
 \chapterconf{Identify}{Identify}{More Than Just Old Rocket Bodies}
-\chapterconf{Upload}{Upload}{Send Data to the SatNOGS Network}
+\chapterconf{Upload}{Upload \& Share}{Send Data to the SatNOGS Network}
 \chapterconf{Support}{Support}{Help!}
 \chapterconf{Contact}{Contact}{Email, Chat, Forum}
 %% END MAINMATTER CHAPTERS %%%
diff --git a/src/Upload.tex b/src/Upload.tex
index d9e450b..1f7503a 100644
--- a/src/Upload.tex
+++ b/src/Upload.tex
@@ -9,10 +9,70 @@
 % International Public License (CC BY-SA 4.0) by Jeff Moe.
 %
 
-\section{Upload Data}
+\section{Push Data Upstream}
 \label{sec:upload}
 \index{upload}
-
-Push.
 The \gls{SatNOGS} network isn't ready for optical data yet.
+So there is nothing that can be uploaded. This is under
+development.
+
+\section{File Extensions}
+\label{sec:upload-extensions}
+\index{formats}
+The following file extensions are in use:
+
+\begin{mdframed}[backgroundcolor=blue!10,linecolor=blue!30]
+\begin{description}
+        \item [.cal] --- Cal % XXX
+        \item [.cat] --- Cat % XXX
+        \item [.csv] --- \gls{CSV}
+        \item [.dat] --- Data.
+        \item [.fits] --- \gls{FITS}.
+        \item [.png] --- \gls{PNG}.
+\end{description}
+\end{mdframed}
+\index{FITS}\index{PNG}\index{CSV}
+
+
+\section{Data Files}
+\label{sec:upload-data}
+\index{data}\index{stvid}\index{process}\index{acquire}
+After the \texttt{stvid} acquisition and processing steps have been
+run, files such as below will be generated:
+
+\begin{minted}{sh}
+2022-08-23T04:16:26.633.fits
+2022-08-23T04:16:26.633.fits.cal
+2022-08-23T04:16:26.633.fits.cat
+2022-08-23T04:16:26.633.fits.png
+2022-08-23T04:16:26.633_48473_catalog.dat
+2022-08-23T04:16:26.633_48473_catalog.png
+2022-08-23T04:16:26.633_52718_catalog.dat
+2022-08-23T04:16:26.633_52718_catalog.png
+2022-08-23T04:16:26.633_90000_unid.dat
+2022-08-23T04:16:26.633_90000_unid.png
+2022-08-23T04:16:26.633_hough.csv
+2022-08-23T04:16:26.633_predict.csv
+2022-08-23T04:16:26.633_threshold.csv
+\end{minted}
+
+
+\subsection{\texttt{.fits} FITS Files}
+\gls{FITS} files are the main files created during acquisition and
+are the main files used in processing. It has the data of
+the actual image taken with the camera. FITS files will be significantly
+larger than the other files, at around 35 megabytes each, depending
+on sample rate and frame size.
+\index{FITS}\index{acquire}\index{process}
+
+The \texttt{stvid} program creates it's customized \gls{FITS} format
+file, which is a stack up of multiple images.
+I'll leave explanation of the \texttt{stvid} \gls{FITS} file structure
+to the program's author, Cees Bassa:
+
+\begin{quote}
+With stvid, I typically store 10 seconds worth of video frames, 100 frames at 10 \gls{Hz} frame rate or 250 frames at 25 \gls{Hz} frame rate into a so-called ``four frame''. Here, for each pixel of the initial 100 or 250 frames, the mean, standard deviation, maximum and argmax are stored, each as its own frame. Objects that move will stand out because they will have a high (max - mean) / standard deviation, which allows me to detect trails. Once a trail is detected, the argmax frame tells me the frame number and hence the motion of the trail, and using the timestamps of the individual 100 or 250 frames, I can reconstruct the pixel position as a function of time during the 10 second exposure, using only 4 frames instead of the 100 or 250 individual video frames.
+\end{quote}
+\index{Bassa, Cees}
+