Update README.md

Author: plancherb1

README.md

@@ -17,26 +17,26 @@ This is a stripped down version of the [**Open Speech Recording**](https://githu
   python -m flask run
   ```
 
-3. Then open the link provided in the terminal in a web browser to run the application. Make sure to **run the application in a private or incognito window** which avoids any cacheing issues. Also we've found that the app works best when **using Chrome**. Once the app opens you'll need give access to your microphone, and then you can click ```Record```. Once you finish recording all of the words a popup will appear and ask if you'd like to download the data. Simply click ```OK``` and the files will be downloaded into the folder from which you are running the flask app (which should be the open-speech-recording folder). You can then close the browser window and kill the app with ```cntrl+c```.
+3. Then open the link provided in the terminal in a web browser to run the application. Make sure to **run the application in a private or incognito window** which avoids any cacheing issues. Also we've found that the app works best when **using Chrome**. Once the app opens you'll need give access to your microphone, and then you can click ```Record```. Once you finish recording all of the words a popup will appear and ask if you'd like to download the data. Simply click ```OK``` and the files will be downloaded into the folder from which you are running the flask app (which should be the open-speech-recording folder).
 
-Note: if you want to change the words or counts recorded make sure to kill and re-start the app and open the link in a new incognito window!
+Note: if you want to change the words or counts make sure to kill and re-start the app and open the link in a new incognito window to avoid any cacheing issues at the server or browser level!
 
 ### You can use the scripts to manipulate the data as follows:
 
-1. You now have a large collection of ```.ogg``` files in your directory. We now need to convert them to ```.wav``` files using ```ffmpeg```:
+1. You can convert the ```.ogg``` files to ```.wav``` files using ```ffmpeg```:
   ```
   sudo apt-get install ffmpeg
   mkdir wavs
   find *.ogg -print0 | xargs -0 basename -s .ogg | xargs -I {} ffmpeg -i {}.ogg -ar 16000 wavs/{}.wav
   ```
 
-2. Then we need to trim them using Pete's tool. Note: background noise does not need to be trimmed and does not need to be recorded by the tool — you simply need to convert them to ```.wav``` files. In fact, they are expected to be longer. See the [original dataset](https://storage.googleapis.com/download.tensorflow.org/data/speech_commands_v0.02.tar.gz) for examples. 
+2. You can then trim the ```.wav``` files with Pete's tool.
   ```
   mkdir trimmed_wavs
   make -C extract_loudest_section/
   /tmp/extract_loudest_section/gen/bin/extract_loudest_section 'wavs/*.wav' trimmed_wavs/
   ```
-3. Finally we need to create the directory structure as the training script expects that the data is organized into folders where there is one folder for each of the ```wanted_words``` filled with ```.wav``` files of the wanted words (as 1 second clips) and an additional folder called ```_background_noise_``` containing longer files of background noise. We will then compress this into a zip file so it can be easily uploaded to Colab. 
+3. Finally we can create the directory structure expected by the Tensorflow training script by running another of Pete's scripts adn then compress it into a zip file so it can be easily uploaded to Colab.
   ```
   python organize_wavs.py
   cd output_wavs