Recording a Quick Memo
Simplified Guide to Automated Recording
Using a Bluetooth Headset
Exporting a Recording
Viewing/Playing Exported Recordings
Sharing a Recording
Deleting a Recording
Where Exported Recordings are Stored
The Simplified UI
Simplified UI versus Advanced UI
The Options Menu
The Spectral Display
What's a "dB"?
Setting the Triggers
Recording Phone Conversations
Monitoring Sleep Talk and Snoring
Capturing Bird Calls
Running Autocord Overnight
Select Record mode. Press the Start button. Speak. You may pause and continue recording at any point. When done press Stop.
Now select Play mode. Press Start, and you will hear your recording. You may pause and continue playback at any point.
That's all there is to making a quick voice memo. You may close the app, and when you return your recording will always be there for you until your record something else. Just select Play mode and Start to hear your recording again.
It is very simple to set up Autocord to trigger recording automatically...
Select Monitor Mode and tap Start.
You will see the spectrum of whatever sounds there are at the moment displayed, and the horizontal input level indicator bar changing. If you speak you will see these react. Just take that as an indication that everything is working.
Now tap the little gear icon and the Settings dialog will be displayed. On the upper right of this dialog you will see the icon to Auto Set triggering.
In a quiet moment, tap that icon. It will automatically adjust sensitivity to just above the current noise level. After a second or two, tap it again to stop it. That automatically sets the trigger.
If some unexpected noise occurred during the Auto Set operation, the horizontal line across the top of the spectrum representing sensitivity may jump way up, reducing the sensitivity unnecessarily. Just stop Auto Set and when it is quiet again, restart it.
The horizontal line will settle itself just above the peak noise level, meaning that if a sound occurs above the noise level, it will trigger recording on - as long as that sound falls within the frequency range that gets set as well.
Now dismiss the Settings dialog and start recording. When you speak, the word "Recording" will display in red, indicating sound is being captured.
When you stop speaking, the word "Recording" will no longer be red, and capturing of sound is stopped. That's it. You now have automated recording.
Please note, if you are in a very noisy environment like a party or gathering with a radio or TV playing in the background, or constant unexpected noise sources it will not be possible to use triggered recording.
If a Bluetooth headset is connected to the phone, when you start Autocord it will automatically route audio to and from the headset. If the headset was not already connected you must stop Autocord before pairing/connecting.
You can record/play via the in-car Bluetooth sound system. Start your car and pair it with your phone before you open the app. Now before you drive off, set triggering, then start the app recording. It will show up on the car's display as a "call in progress". Now forget about your phone and drive off. When you speak, recording will be automatically triggered.
If a call comes in, answer it via the car's phone interface. Recording will pause until you finish the call. Alternatively, if you want to record the call ask somebody with you to unpause recording, or pull over and do it yourself. Please do not attempt this while driving as the app's user interface is not suitable for use while driving.
If you want to preserve your recording, press the File icon while in RECORD or MONITOR mode. A little dialog will pop up and if you click "Ok" the recording will be saved in compressed format, with the name based on the date/time it was recorded in the format...
yyyymmddhhmmss (year, month, day, hour, minute, second)
For example, a file recorded on June 1st, 2021 at 11:33:21 am will have the name 20210601113321 plus the extension. You will find it in the folder Autocord.
To list and play exported audio files, press the File icon while in PLAY mode.
A dialog will appear with the file names translated into a friendly date/time format, for example 20210601113321.m4a will be listed as "Jun. 01/21, 11:33:21". The duration in the form hh:mm:ss will appear beside it.
Simply select the desired file from the list in the dialog and press the play button. You will be able to seek to any position in the audio file with the track bar, or pause and play again at any point. The exact time each sound was recorded will display above the track bar. If you are using the Advanced UI, the peak frequency and decibel level of that peak will also be displayed moment to moment.
To share exported recordings press the File icon while in PLAY mode. You will see your tracks listed and you can play them to determine which you want to share. With the track selected, click the share icon and you will be presented with a list of options to share to. Of course you can always just navigate in your files app to where Autocord's files are stored and share them from there.
Whenever you start a new recording, the previous is overwritten. If you left a very long recording you are no longer interested in, simply record very briefly over it to free up the space it took.
To delete exported recordings press the File icon while in PLAY mode. You will see your tracks listed and you can play them to determine which you want to delete. With the track selected, click the trash icon to mark it for delete. Tracks marked for delete will be deleted once you close this dialog. Upon closing you will again have an opportunity to change your mind or confirm delete. Once you click "Ok" the tracks are permanently deleted, along with their associated log files.
If deleting tracks manually, you need to use the Files app to find them and delete them. When you delete a recording, also be sure to delete its associated text file at the same time. For example, if you are deleting 20210601113321.m4a look for 20210601113321.txt and delete that too. Take care not to delete the wrong text file.
You can find your exported tracks with your Files app. For example, open Samsung's "My Files". From Categories select "Audio". You will see the folder "Autocord". Open the folder and you can play, share, or delete recordings. In File Manager you will see the files listed under "Autocord".
For editing or serious analyses save your recording with the lossless FLAC encoder and transfer to your computer. If you don't have a professional audio editor, the free and open source Audacity audio editor is recommended. It will load the flac files generated by Autocord. While you are at it, transfer the text file with the same name as the recording to your computer as well. It displays the time captured and sample offset for every sound event and you may refer to it to locate a particular part of the recording.
The operation of Autocord requires only a few UI elements. The image belows shows the simplified UI the user is presented with when first opening the app...
There are three modes of operation, Monitor Mode, Record Mode, and Play Mode, selected by a tap on one of the radio buttons.
As soon as you tap the Start icon, the name of the current mode you are in is displayed over top the radio buttons. When you Stop, you may select a different mode.
At any time in Monitor Mode or Record Mode you have access to the Settings dialog via the little gear icon, but in Play Mode this icon is replaced by the "Info" icon. When you tap this it will give you an abundance of information about the current recording.
Similarly, at any time when stopped in Monitor Mode or Record Mode you may save/export the current recording via the "Files" icon (Export/Play audio file icon), but in Play Mode this icon gives you access to already exported files so you can play them.
There are three values displayed above the row of radio buttons. During recording, the duration of the recording and its file size is displayed and constantly updated. In either MONITOR or RECORD mode, you will also see the Input Level displayed, both the actual value and visually on the horizontal level indicator bar.
During playback, instead of duration, time recorded is displayed, along with the file offset. This is constantly being updated in accordance with what you are hearing.
One more piece of information is also displayed, "Gain Applied". During playback Autocord "normalizes" each contiguous group of samples triggered to maximize the sound level. The value here will be the opposite of the peak Input Level for any given group of contiguous samples when recorded.
With the constant gain adjustments they all come out at the same maximized volume level, except a faint sound will have a poor signal to noise ratio compared to a sound that was originally loud.
You don't need to be tech-wiz to use this app. When you first open the app, you are presented with the Simplified UI, which shows fewer items on the screen so you will not be overwhelmed with complexity.
If you want to see what you are missing out on, go to the Options menu via the 3 little dots in the upper right corner and choose "Advanced UI". This adds the display of Peak frequency, Peak dB, and Note value on the main screen, things that you don't really need if you aren't much interested in technology. In the Options dialog and you will see you now have several more options available. Read on below to learn about them.
Besides the above mentioned option to choose the Simplified or Advanced UI, there are 8 more options available in the Advanced UI:
When you click on the Settings icon to set the trigger and click "Auto set", the frequency range of sounds that can trigger recording is set to one of the three presets selected here, which are Trigger on: Voice, Birds, or Any. Of course you can always use the sliders in the Settings panel to adjust the frequency as you like, but it will always begin with one of these preset frequency ranges. The Voice preset frequency ranges from 320 Hz to 4000 Hz, for Birds it's 3000 Hz to 5000 Hz, and for Any the range is set from 90 Hz to 8000 Hz.
Input Filter is selected by default. Input filtering removes subsonics and very low frequencies below the range of the male voice that only serve to muddy up the sound. These sounds are produced by sources such as machinery, electrical equipment, and car engines. The cutoff is set to 90 Hz. The only time you wouldn't want this is if you are recording music.
Autocord is optimized to capture distant sounds. Input boost helps to amplify very low level sounds and improves signal to noise ratio and the effectiveness of the input filter. It provides a boost of about 8 dB at the input stage, before further processing and amplification. It is recommended to disable Input Boost if your phone is very sensitive, or if you typically speak directly into the phone when recording. If your phone is very sensitive, you will notice that the input sensitivity bar often deflects all the way to the right when you speak. In these cases there is no need for the extra gain and it may even cause a bit of distortion on the peaks.
There are 3 options for encoding an audio file for export. These are: AAC at 64,000 bits per second (the default), AAC at 128,000 bits per second (higher quality), and FLAC, which is a lossless encoding scheme. Audio files would quickly become very big without some form of compression. AAC at 64K BPS reduces the file size to 12.5% yet maintains good quality sound, while AAC at 128K BPS only reduces file size to 25%, in exchange for higher quality sound.
FLAC compression gives the highest possible quality in exchange for very poor compression, typically only reducing a file to around 62% of its original size. However, when a FLAC file is decoded, it is identical with the original. This is the preferred format for scientific analyses of the captured audio, but you wouldn't want too many of these large files to accumulate on your phone as they can take up a lot of space.
An hour long recording will occupy ~28 mb when encoded to AAC at 64K BPS, ~56 mb when encoded as AAC at 128K BPS, and 130 to 190 mb when encoded as FLAC. It will take around 2 minutes to encode an hour of audio (as measured on a Samsung Galaxy Note 8).
Select Monitor mode, and press Start. If you speak or whistle you will see the spectrum of the sound you make. Remain quiet for a moment, and you will see the peak level of the ambient noise.
The frequencies of the spectrum range from 16 Hz to 16,000 Hz, divided into 10 octaves, running from octave 0 on the left to octave 9 on the right. These are the same octaves you will find on a piano, except a piano doesn't have such a large range, extending only from octave 1 to octave 7.
Dividing a spectrum into octaves gives the viewer a frame of reference. Zoom in on the legend below the display in one of the screen shots here. If you look closely a little to the left of centre, you will see the label "Middle C", which is octave 4. This octave extends from the note C4 to B4, or in terms of frequencies, from about 261.63 Hz to 493.88 Hz. You will also see the frequency that all the other octaves begin at.
Each octave is conceptually divided into 48 bands. When a sound is processed, the highest peak in each of these bands forms one vertical line of the graph. Its height shows the magnitude of that peak - if indeed a peak was found in that band. You will note that the lower the octave, the more gaps will be seen in the display as no peak was found for several bands.
The fundamental frequency of the male voice is typically around 100 Hz to 120 Hz, while the fundamental of the female voice ranges from 165 Hz to 255 Hz. It is not quite that simple, however. Most sounds in nature generate a range of frequencies at once, beginning at some fundamental value and running up the spectrum. Then when we see the spectrum of a male speaking compared with the spectrum of a female speaking, we will see a considerable overlap.
Also, most of the energy may be well above the fundamental. For example, a male saying "hello" may produce the highest peak at around 488 Hz. It all depends on the syllables being spoken.
Meanwhile, many bird songs have frequency ranges between 1,000 Hz and 8,000 Hz. You can play with different sounds to see their spectrum - always superimposed upon the ambient noise level.
With the Advanced UI (selected from the Options Menu), you will see the peak frequency and its dB level. If that happened to coincide with a musical note, you will also see its musical note value.
"dB" is short for "decibel", which is a measure of sound level. We use the decibel scale instead of displaying the amplitude value because it conforms to how humans interpret loudness. Our ears are more sensitive to quiet sounds than to loud sounds. Our hearing sensitivity is logarithmic and the decibel scale is also logarithmic to match.
A decibel is the smallest change in volume we can perceive in a laboratory setting, though it typically requires a change of 3 dB before we can all agree there has been a change in the volume of some sound source in the real world. If the amplitude of some tone is reduced by half, our ears interpret that as a drop of only 6 dB.
Autocord deals in terms of relative dB, since it requires specialized equipment to determine absolute sound level. For Autocord, a sound at full amplitude is arbitrarily labeled "0 db", and anything less than that necessarily must be negative. For Autocord, -96 dB is perfect silence, so the level of any sound captured by Autocord ranges from -96 dB up to 0 dB.
If you see the noise peaking up around -50 dB, some specific sound embedded in that noise at -70 dB will be barely discernible, depending on its frequency in relation to the noise of course. Then typically we don't want to set the trigger lower than 20 dB below the peak noise level.
This app is all about triggered recording, and the reason it is named "Autocord" - for "Automated Recording". Audio capture can be triggered not only by the magnitude of the signal, but also only if it falls within a desired frequency band. Thus it can even pick out a sound of interest from below the ambient noise level. Click on the Settings button (the little gear icon) and you will see the settings screen as shown here...
In Monitor Mode, tap Start, then tap the little gear icon to bring up the Settings dialog, and here tap the Auto Set button during a quiet moment. You will see a horizontal line being drawn across the spectral display just above the highest peaks, demarking the sensitivity threshold, and vertical lines marking the minimum and maximum frequency on the spectral display. If audio is Stopped, the Auto Set feature will not be available, because of course it requires live audio input to operate. However, you can still set or adjust all the parameters manually if you want.
Only sounds that fall within the frequency band set can trigger recording, and only if that sound exceeds the sensitivity threshold marked by the horizontal line, which is automatically set to be 3 dB above the peak of the noise.
When you click "Auto set", the frequency range of sounds that can trigger recording is set to one of the three presets selected in the Options menu, which are Trigger on: Voice, Birds, or Any. Of course you can always use the sliders to adjust the frequency as you like, but it will always begin with one of these preset frequency ranges. The Voice preset frequency ranges from 320 Hz to 4000 Hz, for Birds it's 3000 Hz to 5000 Hz, and for Any the range is set from 90 Hz to 8000 Hz, a range to capture just about any sound that could be of interest.
During the Auto Set operation you may notice the horizontal line representing the sensitivity jump a bit as higher peaks come and bump it up. If some unexpected loud sound occurs it may kick the sensitivity line way up. Just stop the Auto Set operation and tap Auto Set again to restart it. After a second or two, stop Auto Set to freeze the setting.
If you are dictating to your phone in your hand or on the desk beside you, use the preset for Voice and slide the sensitivity line up a bit higher so it won't trigger on every little extraneous sound.
Of course it all depends on the noise level where you are at. Observe this before opening the Settings dialog, then ensure you don't set the sensitivity to more than 20 dB below the peak ambient noise level of the full spectrum. There is no point in capturing sounds that are buried in noise.
If the noise level is so high that the input kevel bar is very active, often deflecting nearly all the way across, triggered recording will be nearly impossible unless the spectrum of the noise is outside the range of the sounds you want to capture.
To dismiss the Settings dialog tap the phone's Back button, or just tap the screen outside of the Settings dialog.
Now set the app into Record Mode and tap Start when ready. The word "Recording" will be displayed. When a sound occurs that triggers audio capture, that word will turn red for the duration of the triggering events. When the sound level drops below the trigger threshold, the word "Recording" will no longer be red.
While recording you will see the portion of the spectrum that you selected to trigger on framed between two faint vertical lines. During playback you will see the spectrum as it was captured. Experiment with this triggering a bit and you will understand it quickly enough. It becomes intuitive with the visual display of the triggers to guide you.
Autocord can capture phone calls as a byproduct of its fundamental capabilities but was not created specifically for this purpose. There are many other apps available specifically for capturing phone calls that would manage the task more conveniently.
If the app is recording and the phone rings, it will automatically pause. Answer the call, and navigate back to Autocord and unpause it to capture the conversation.
If you are not already in record mode, ask for a moment while you navigate to the app and set it to record.
Only caveat here is if you are using the speakerphone, you will not capture your end of the conversation very well because of the echo cancellation done by the phone. If you are using headphones or earbuds, no problem! You will capture both sides of the call.
By the way, Android will briefly pause and unpause itself at any time during Play or Record when notifications sound on the phone.
Did you ever wonder if you talk in your sleep or snore? Use the preset "Trigger on: Voice" in the Options menu and tap the Auto Set button in the Settings dialog to set the trigger. Leave your phone recording by the bed or even on the bureau across the room, but away from the window or any vents. It will only trigger on record when you speak or snore, and ignore sounds like cars passing in the distance. You may be quite surprised by what you hear, or even the conversations that go on between couples in their sleep!
For bird watchers, the "Dawn Chorus" is one of the most inspiring moments of the day. To capture bird calls use the preset "Trigger on: Birds" in the Options menu and tap the Auto Set button in the Settings dialog to set the trigger.
Of course the frequency range you will set all depends on what kind of birds you are monitoring. While listening to birds, set Autocord in monitor mode and observe what the frequencies peak at when a bird calls. While 3000 Hz to 5000 Hz is perfect for starlings, crows for example make their calls in the range of 700 Hz to 1,700 Hz. Set the trigger and leave your phone on the window sill or veranda overnight on record.
It's becoming a hobby for some who like running Autocord all night, or to leave it capturing for hours in some location just to see what sounds they catch. For some of these sounds it becomes a bit of a mystery to determine what created them.
The thing about triggered recording is, that if all goes well, at the end of several hours you may only have several minutes of interesting sounds, so you are not going to run out of disk space even running it all night if you set the triggers correctly.
To capture a broad range of potentially interesting sounds use the preset "Trigger on: Any" in the Options menu and tap the Auto Set button in the Settings dialog to set the trigger.
If you are in a very noisy environment try raising the minimum frequency threshold. Typically, the higher you raise this, the lower the noise level you will observe within the frequency range set. You will still trigger on interesting events and even voices if you raise this to 320 Hz or more because remember, most sounds are broadband. Raising the minimum frequency threshold will allow you to bypass a lot of sounds that arise from "the roar of the city".
A final note about triggered recording... You can set it just right and leave it running all night long if you wish but watch out for some unexpected source of continuous noise that can ruin it all for you.
For example, an unanticipated motor, fan, or air conditioner switching on and staying on some time after you set the trigger. Or for a phone exposed to the outdoors, the wind picking up and making noise as it blows past the phone, or the sound of rainfall. If any of these things should occur, the noise can permanently trigger recording for the duration, which may be all night if that's what you are attempting. As you gain experience with setting the trigger you will be better able to anticipate these kinds of situations.
As far as battery usage goes, you may expect it to drop up to 10% with Autocord running all night. Of course you shouldn't leave the screen on. Autocord will continue to run with the screen off. You may leave it charging while you run Autocord if you wish.
During recording a log of triggering events is maintained. The time, sample offset, peak frequency, and magnitude of each event is individually recorded. When you export the audio file, you will find the event log in the same folder as the audio, with the same name except the extension will be ".txt".
The event log will begin with a header showing date/time capture began, what the trigger frequencies and trigger sensitivity was set to, the total number of samples recorded, and the duration of the file in hours, minutes, and seconds.
Then later if you open your audio file in an audio editor you will be able to locate every event. Of course remember that events will be interspersed with the blocks of samples recorded before and after the events that give continuity to each sound captured.
Typically a group of contiguous samples will begin with the half-second of audio from before the events (16 K samples), followed by some number of events, which will in turn be followed by a half-second of audio from after the events (another 16 K samples).
The events themselves may be interspersed with "non-events" (up to 3 blocks of 4 K samples each), a "non-event" being audio that briefly fell below the triggering threshold or outside of the frequency range.