Latest revision as of 17:55, 25 November 2024

Press #1 For:

Description

From getting your BSN to installing internet in your new home to renting a car, rigid mazes of bureaucratic systems hold us all hostage. In the spirit of collective catharsis 'Press One For' invites you to share and listen to stories of how these systems frustrated you, or how you frustrated them.

T65 Rotary Phone bought by Sevgi n' Fred

Cookbook

Parts Required

USB interface to 3.5mm

https://www.adafruit.com/product/1475

3.5mm Cables

Old phone (T65)

https://www.marktplaats.nl/v/telecommunicatie/vaste-telefoons-niet-draadloos/m2158407457-ptt-t65-goed-werkende-nostalgische-telefoon

Raspberry Pi

TRS 3.5 Jack x2

Loads of wires

old headphones with a mic (apple earpod works best)

Things we done

- We got one phone from ScrapXL Rotterdam and one from Marktplaats. The first one was a T65

[11 (Neg), 12(Pos) speaker output]

We tried batteries for the carbon mic, didn't work

- We took the phone apart and got the speaker to work by connecting the blue and red cables from the phone out to the audio jack to the soundcard to the computer. played show me the body, was good.

- Then we tried the microphone but its a carbon mic and it didnt work. so we atttached a battery and hoped. didnt work.

- We took apart wired headphones that had mic in them to see if we could use the mic from it rather than the carbon mic in the telephone.-> didnt work, headphone cable is TRRS not TRS:

spent a day trying to wire up TRRS to an audio cable, turns out the audio jack was the wrong kind.

then wired it directly to a TRS jack, worked!!!!!!!!!!!!!

then cable managed, looks beautiful now. now we have working speakers and working mic.

This is how you could make it work too (Schema 1)

From the headphones you need to...

Connect th Blue (Ground) to the Sleeve of the TRS jack

Connect the Red (mic) to Pin (Ring) 1 or 2 of the TRS jack

MAKE SURE THE HEADPHONES CONNECTION IS CLOSED!

The wiring used

sketch of the wiring of the handset connection (Schema 1)

Next Steps:

We need to work out how to read the keypad input on the raspberrypi. Mark Powers has a snippet of code that supposedly works with this: https://marks.kitchen/blog/rotary_phone_running_linux/

To test our phone we used this GitHub project: https://github.com/ralphcrutzen/PTT-Tafeltjes-Telefoon?tab=readme-ov-file. > This project turns the phone into a math game where a child asks you a times table question and you are supposed to give the correct answer using the rotary keypad. When you dial the answer the py code reads it as 2 separate entires and gives you a Goed zo! if you gave the correct answer.

Fred is translating the code to eng. He also thinks I can't do times tables.

Set up the raspberrypi, named it telephone.(tut here:https://nematicslab.com/how-to-enable-ssh-without-using-a-monitor/)

Fred made me a user(Sevgi) as well and we tested the connections first, adjusted the code accordingly. Fred tested the connections like this (this is the final working version):

import RPi.GPIO as GPIO


GPIO.setmode(GPIO.BCM)
GPIO.setup(23, GPIO.IN, pull_up_down = GPIO.PUD_UP)
GPIO.setup(24, GPIO.IN, pull_up_down = GPIO.PUD_UP)
GPIO.setup(25, GPIO.IN, pull_up_down = GPIO.PUD_UP)

while True:
        
        print('-')
        if GPIO.input(23):
                print('23')
        if GPIO.input(25):
                print('25')
        if GPIO.input(24):
                print('24')

Here is how we set up the wires and the raspberry pi:

This wiring worked for us at first, but then we had to put the ground cables in different Rbpi GND's for it to work again.

Here is the eng version of the dutch code: https://pzwiki.wdka.nl/mediadesign/User:Wordfa/oldphonehack/PTT_Tafeltjes_Telefoon

We now reverted back to writing the code Ralph Crutzen's way, using pygame. We pulled this code from: https://github.com/pygame/pygame/blob/main/examples/audiocapture.py

""" pygame.examples.audiocapture

A pygame 2 experiment.

* record sound from a microphone
* play back the recorded sound
"""
import pygame as pg
import time
import wave

from pygame._sdl2 import (
    get_audio_device_names,
    AudioDevice,
    AUDIO_F32,
    AUDIO_ALLOW_FORMAT_CHANGE,
)
from pygame._sdl2.mixer import set_post_mix

# sound length (seconds, a float)
SOUNDLEN = 2.0
# sound frequency (in Herz: the number of vibrations per second)
SOUNDFREQ = 1000

# maximal amplitude
MAXAMP = 32767 / 2
# sampling frequency (in Herz: the number of samples per second)
SAMPLINGFREQ = 44100
# the number of channels (1=mono, 2=stereo)
NCHANNELS = 2

pg.mixer.pre_init(44100, 32, 2, 512)
pg.init()

# init_subsystem(INIT_AUDIO)
names = get_audio_device_names(True)
print(names)
sounds = []
sound_chunks = []


def callback(audiodevice, audiomemoryview):
    """This is called in the sound thread.

    Note, that the frequency and such you request may not be what you get.
    """
    # print(type(audiomemoryview), len(audiomemoryview))
    # print(audiodevice)
    sound_chunks.append(bytes(audiomemoryview))


def postmix_callback(postmix, audiomemoryview):
    """This is called in the sound thread.

    At the end of mixing we get this data.
    """
    print(type(audiomemoryview), len(audiomemoryview))
    print(postmix)


set_post_mix(postmix_callback)

audio = AudioDevice(
    devicename=names[0],
    iscapture=True,
    frequency=44100,
    audioformat=AUDIO_F32,
    numchannels=2,
    chunksize=512,
    allowed_changes=AUDIO_ALLOW_FORMAT_CHANGE,
    callback=callback,
)
# start recording.
audio.pause(0)
print(audio)

print(f"recording with '{names[0]}'")
time.sleep(2)


print("Turning data into a pg.mixer.Sound")

sound = pg.mixer.Sound(buffer=b"".join(sound_chunks))
snd = pg.mixer.Sound(sound)

# open new wave file
sfile = wave.open('pure_tone.wav', 'w')

# set the parameters
sfile.setnchannels(NCHANNELS)

# write raw PyGame sound buffer to wave file
sfile.writeframesraw(snd.get_raw())

# close file
sfile.close()

print("playing back recorded sound")
sound.play()
time.sleep(2)
pg.quit()

This didn't really work, just worked for us to record the sound and play it back on the handset.

We gave up on pygame and used pyaudio. Fred found this snippet of code which we will try to translate into our test.py:

import RPi.GPIO as gpio 
from recorder import Recorder 
gpio.setmode(gpio.BCM)  

class ButtonRecorder(object): 
    def __init__(self, filename): 
        self.filename = filename 
        gpio.setup(23, gpio.IN, pull_up_down=gpio.PUD_UP) 
        self.rec = Recorder(channels=2) 

    def start(self): 
        gpio.add_event_detect(23, gpio.FALLING, callback=self.falling, bouncetime=10) 

    def rising(self, channel): 
        gpio.remove_event_detect(23) 
        print 'Button up' 
        gpio.add_event_detect(23, gpio.FALLING, callback=self.falling, bouncetime=10) 
        self.recfile.stop_recording() 
        self.recfile.close() 

    def falling(self, channel): 
        gpio.remove_event_detect(23) 
        print 'Button down' 
        gpio.add_event_detect(23, gpio.RISING, callback=self.rising, bouncetime=10) 
        self.recfile = self.rec.open(self.filename, 'wb')    
        self.recfile.start_recording() 

rec = ButtonRecorder('nonblocking.wav')
rec.start() 

try: 
    raw_input() 

except KeyboardInterrupt: 
    pass 

gpio.cleanup()

We managed to get a recording with this code:

We made a first recording with this. (We were discussing if having a limited time to record is okay instead of hanging the phone to stop recording and its a second of that convo)

We combined these 2 snippets together but had errors, we had to install a bunch of audio related python scripts and Joseph told us to download sox (in his words: Swiss army knife Audio python scripts) :

sudo apt install sox

We had to do some troubleshooting for sox because it also didn't work perfectly of course.

To do list from 13/11/2024 on:

- discuss the intro in detail : maybe a connecting sound(?) at first, then hold music, then intro, then press 1 for: etc, beep sound before the recording, etc. , before the recording: disclaimer,

- decide whether we have different audios for each pickup (DONE)

- what is the detailed version of the 'annoying' bureaucratic protocol? make it a detailed question.

- discuss decor after the giveaway shop visit (DONE) File:Output-11.11.2024.mp3

- collect forms / print forms

- (eventually) make the pi go into the phone and fix the wiring (take everything out, solder the wires, ) (DONE)

13/11/2024

We went to the weggeefwinkel to grab some office related decor. Which was amazing, we got a desk lamp, plastic transparent folder, an old picture of some business people and a frame to go with it and a briefcase to put them all in.

Today everything went wrong... We discovered that the little metal tabs need to be warmed up by the solder machine and then get a bit of lead on them to make them work. Joseph told us we need to let the machine warm up the tabs until they are both ~325 °C, then let some solder into the tab. After the warm up, the solder should stick to the tab. Martina tried really hard and managed to connect 3 wires (without knowing any of this), Fred and Sevgi couldn't manage the other and gave up on that. (Martina said it was easy when She did it) We wired everything with different ways and took the case off of the pi to make it fit inside the telephone.

We took out the bells to be used in the jam session.

Fred cleaned the rotary part and the spiral came out which took him forever to put back. Eventually he figured out it is easier to attach the inner bit and then coil the rest of the spiral inside. Now the rotary is clean and it works smoother.

After putting everything back together, there seems to be a tone of connection issues. Audio seems to being played lower then expected but Maroon 5 still sounded good. Fred stayed behind and managed to fix it by uninstalling and reinstalling pulseaudio and playing around with Sox.

We still have rest of the to-do list to go through

Everyone hated today

14/11/2024

Martina typing:

Fred added a function to clean up the recordings and make them louder. He also tidied up the code, added some beep tones when numbers are dialed in. In the code, he also made a function to loop through all the recordings randomly.

Press 1 for: Flowchart

Code for Today -> User:Wordfa/oldphonehack/code14nov

Fred and I discussed the next steps to be done for the weekend:

I will work on the script and the forms that will be placed on the text. The inspiration is to make them look like survey forms, maybe participants can also interact with these forms.

Fred also made this flowchart of how the call can follow different outcomes.

At the end of the day, Fred and I tried to close up the telephone with the pi and cables inside, but unfortunately it does not fit.... we re-wired some of the cables to make the pi fit on the outside of the telephone. We will do the rest on Monday as we didn't have enough soldering wire and were very tired </3

Figma flowchart with draft text: https://www.figma.com/board/aKK1GVkBnmIyScXWcyUmGP/telephone-flowchart?node-id=0-1&node-type=canvas&t=T2I1UidcYxM7jHld-0

These are the tasks that we will work on Monday:

Re-wire the rest of the cables to make the pi fit on the outside of the telephone. (DONE)
Read through the draft script together and arrange it. (DONE)
Record the script audios and upload everything into the code. (DONE)
Make sure that everything sounds good on the telephone. (DONE)
Check in with Worm to see if they have an office table? (will be done wednesday)
Print forms (with HR printers or publication studio?) (DONE)
Pack up all the decor, equipment, etc.. (DONE)
Check ways to reduce noise (DONE)
Record a few complaints for option 2 / maybe ask the class on Zulip to record some complaints

18/11/2024

Martina typing:

During this session, we continued the rest of the soldering to make the pi rest outside. All went well but we did have some sound issues, but all got fixed when Fred adjusted the volume settings.

Sevgi, Fred and I also went through the script together (spoken script for telephone). Later on, I recorded the audio so Fred would be able to put it into the code. We also added a 'secret' feature where you can dial in 9 to listen to the last recording. This is mainly for us to know that the telephone is working correctly.

19/11/2024:

Martina typing:

One day left!!!!

We started straight with working on the final details. Sevgi and Fred dressed Sassie back up and she looks gorgeous with her shell on! I worked on printing the survey forms on the paper we found from scrap. Feline kindly helped me navigate the HR printers and we have beautiful forms, they will be placed in the white folder on the office desk installation.

I packed up all our office decor to be ready for tomorrow while Fred cleans up the code. He also made a secret new function: if you press 0, you hear a lovely tune.