fix speedup bounds

app.py

@@ -288,7 +288,7 @@ with st.form("my_form"):
     url = st.text_input(
         "youtube url", value="https://www.youtube.com/watch?v=_3MBQm7GFIM"
     )
-    speedup_factor = st.slider("speedup", min_value=1.0, max_value=10.0, value=1.5)
+    speedup_factor = st.slider("speedup", min_value=0.5, max_value=5.0, value=1.5)
     min_speedup = st.slider("minimum speedup", min_value=0.5, max_value=5.0, value=1.0)
     max_speedup = st.slider("maximum speedup", min_value=0.5, max_value=5.0, value=2.0)
     max_num_segments = st.slider(
@@ -323,6 +323,6 @@ st.markdown(
 5. fit a piecewise-constant function to the information rate vs. time data using a decision tree regression model from [scikit-learn](https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeRegressor.html).  this lets us control the number of segments that will be stitched together in step 8, which can run slowly if the number of segments is too large.
 6. compute speedup for each segment: 1 / information rate (induces constant bit-rate over time)
 7. clip speedups with user's min and max, and use binary search to find linear scaling factor that matches the user's desired overall speedup
-8. apply scaled and clipped speedups to each segment, and stitch the segments together using ffmpeg
+8. apply scaled and clipped speedups to each segment, and stitch the segments together using [ffmpeg](https://ffmpeg.org/)
 """
 )