The Song Remains The Same.

[Tub]

Often hearing the same songs played on many different radio and television stations, I've wondered how any song always sounds the same, even though it is delivered on the many different frequencies that the different broadcasters use. Why don't those different wavelengths affect the original sound of the song? ( Thanks to Led Zeppelin for the thread title ).

[Sensei]

Right now almost the all radio stations here are transmitting digital signal, not analog anymore.

https://en.wikipedia.org/wiki/Digital_radio

 

Also, your audio experience is pretty subjective.

To really compare signal, you would have to digitalize it and record as raw file on disk, and then compare couple files together. I am pretty sure data would be slightly different in many places (transmission errors, slightly different decompression algorithms). Digital single sample has 65536 ^ 2 = 4294967296 possible values (left and right channels 16 bit both). Multiply 4 bytes per sample by 44100 Hz (CD quality) = 176400 bytes per second of raw data.

 

Edited August 4, 2017 by Sensei

[Tub]

Thank you, Sensei. Sadly, i'm behind the times and all my radios aren't DAB, lol.

[imatfaal]

1 Rather than answer the question directly - which would involve describing how the information is encoded - I will ask an analogous question.  Do you think that the meaning, import, and beauty of a Shakespeare sonnet is changed if it is printed in 12 point, or 72 point? 

2. The frequency of a radio band is where the radio "looks" for the signal - any signal can be encoded in the specific modulation chosen.  The signal, I believe, is encoded in a small change in the main ideal frequency

3. Note the frequency of the sound waves for middle C is about 261 hertz.  FM is broadcast at around 500 times higher frequency than that.  There is no necessary connection between the cycles per second of the soundwave produced by the speaker and the cycles per second of the carrier wave.

[swansont]

Short answer is that the carrier frequency has no effect, it's the modulation, which is the same for the same signal.  

[Tub]

Thank you, imatfaal; thank you swansont. Good, clear answers. I was under the impression that the carrier signal could alter the pitch of the song.

Shakespeare is Shakespeare however it is written, yes, but if he'd heard me trying to sing one of his songs he may have said: " If Music be the food of Love, that's what salmonella sounds like !"

[StringJunky]

Although this is more biology, if you know a song well your brain will make adjustments to the incoming signal to match your expectations and will even synthesise sounds/frequencies that may be missing, derived from memory. .

[Tub]

13 hours ago, StringJunky said:

Although this is more biology, if you know a song well your brain will make adjustments to the incoming signal to match your expectations and will even synthesise sounds/frequencies that may be missing, derived from memory. .

Seems the old saying is true: " I couldn't believe my ears! ". I imagine we are susceptible to subliminal frequencies too. Thanks, SJ.

[swansont]

18 hours ago, Tub said:

Thank you, imatfaal; thank you swansont. Good, clear answers. I was under the impression that the carrier signal could alter the pitch of the song.

It occurs to me that if filtering out the carrier is imperfect it could add a bias to the resulting signal

[imatfaal]

42 minutes ago, swansont said:

It occurs to me that if filtering out the carrier is imperfect it could add a bias to the resulting signal

Any directly generated additional noise would have to be at a fraction of the carrier frequency (not sure if that is the correct terminology) the threashhold of human hearing is just under 30kHz for young adults (who I think have the best range) - whereas FM transmission is about 3000 times that frequency.  There is about 100kHz variation in the carrier frequency of  of about 100MHz..  You are much more likely to get annoyed at mains hum if you are an audio-phile audio-snob

For a decoding bias I just cannot envisage what happens to a 100MHz carrier modulated with a ~10-10000KHz signal when the carrier signal which is removed is actually 99MHz.  Just read that commonest form of demodulator is two rf transformers - #1 on a closed loop with fixed frequency of the carrier and #2 on the received signal; if the frequency is the same on signal and closed then no voltage difference between to transformers and no resulting output.  However if frequency is not the same then the difference will be expressed as a varying out put voltage which is proportional to the difference in frequency.  And that system was developed over 80 years ago and is still the basis of FM radios

[Tub]

5 hours ago, swansont said:

It occurs to me that if filtering out the carrier is imperfect it could add a bias to the resulting signal

I'm sure that's a possibility; all technology is fallible. Conversely, my old mono FM radio is perfect at filtering-out important parts of stereo broadcasts so that some older stereo-recordings, ( particularly Beatles and Led Zeppelin songs), emerge minus vocals or bass or drums or even the whole rhythm-section!  Quite interesting, really.

I'm not sure how to digitalize signals etc, as Sensei suggested above, so the only other way i can think of to compare different signals carrying the same song would be to listen to that same song being played simultaneously on two or more different radio-stations,which is probably impossible to do. Then again, there could be other variables due to the actual radios in use.

 

5 hours ago, imatfaal said:

Any directly generated additional noise would have to be at a fraction of the carrier frequency (not sure if that is the correct terminology) the threashhold of human hearing is just under 30kHz for young adults (who I think have the best range) - whereas FM transmission is about 3000 times that frequency.  There is about 100kHz variation in the carrier frequency of  of about 100MHz..  You are much more likely to get annoyed at mains hum if you are an audio-phile audio-snob

For a decoding bias I just cannot envisage what happens to a 100MHz carrier modulated with a ~10-10000KHz signal when the carrier signal which is removed is actually 99MHz.  Just read that commonest form of demodulator is two rf transformers - #1 on a closed loop with fixed frequency of the carrier and #2 on the received signal; if the frequency is the same on signal and closed then no voltage difference between to transformers and no resulting output.  However if frequency is not the same then the difference will be expressed as a varying out put voltage which is proportional to the difference in frequency.  And that system was developed over 80 years ago and is still the basis of FM radios

http://www.electronics-tutorials.ws/category/systems

There is a good explanation of  Loop-systems on the above site.It seems they are still used now as part of  " Error Correction Coding " which is designed to filter-out errors in digital-broadcasting too, but, again, they aren't infallible: if the ECC is itself faulty or weak it won't correct signal-errors and a so-called " bubbling mud " interference or a complete loss of signal, ( a " digital-cliff " ), can occur. I also read that varying bit rates used by different digital broadcasters can also have an effect on sound quality and authenticity,

 

For a very simple experiment, i tuned my new DAB radio and my 50 year-old FM radio to the same station. A Mary Winehouse song was playing and, apart from the DAB delay, i couldn't really discern any difference and both songs were in tune with each other, in the same key of D minor. As i said, I suppose a truer test would be to hear the same song played simultaneously on two or more different radio stations but, again, that's not really feasible. Anyway, what i did hear was good enough for my ears, but maybe not for someone prepared to pay tens of thousands of pounds for a sound-system, or someone with perfect-pitch or just a very good ear for music who could , perhaps,hear a difference between the same music played on different stations,or in different formats, compared with what they hear on their sound-systems at home.

 

 

[imatfaal]

Is Mary Winehouse a bastard mix of Amy Winehouse and Mary Whitehouse?  A great soulful voice who sings about censorship and the decline of family values

if you sit in quiet room and listen to your favourite tune on FM radio, DAB radio, tape, cd, bad encrytion, good encryption, etc.  You will notice a difference but you pay your money and take your choice of where you draw the balance between money and fidelity.

 

[Tub]

2 hours ago, imatfaal said:

Is Mary Winehouse a bastard mix of Amy Winehouse and Mary Whitehouse?  A great soulful voice who sings about censorship and the decline of family values

 

 

Ha, Ha. Yes. Sorry. And sorry, Amy ( R.I.P. ).  Ironically, it must have been the glass of red wine i was drinking to help me sleep. ( Just woke up on the sofa with a stiff neck!)

2 hours ago, imatfaal said:

 

if you sit in quiet room and listen to your favourite tune on FM radio, DAB radio, tape, cd, bad encryption, good encryption, etc.  You will notice a difference but you pay your money and take your choice of where you draw the balance between money and fidelity.

 

What's underlined  above ( my underlining ) is exactly what i was going to write in my last post: " You pays your money and you takes your choice ", but i thought the phrase might get  a bit lost in translation - especially to people who wouldn't know anything about Mary Whitehouse, ( again, R.I.P. ), who would have been completely horrified at being described as  anything like a " bastard mix ", lol.  Language,Timothy!

Just by the way, my favourite old FM radio is my 1967 Bang and Olufsen Beomaster 900 and it still sounds good to my ears even though it's only worth about £10 now.

 

[Sensei]

On 8/5/2017 at 11:34 PM, Tub said:

I'm not sure how to digitalize signals etc, as Sensei suggested above,

It's done by A/D converter

https://en.wikipedia.org/wiki/Analog-to-digital_converter

It can be done by any modern smartphone and any computer with microphone, when you press record in appropriate application.

Store it in lossless audio file format like raw or wav, not mp3.

https://en.wikipedia.org/wiki/Audio_file_format#Lossless_compressed_audio_format

Edited August 8, 2017 by Sensei

[Tub]

Thanks again, Sensei. Raw looks like the best bet.

[Sensei]

5 hours ago, Tub said:

Thanks again, Sensei. Raw looks like the best bet.

.WAV file format is typically just raw (uncompressed) sample data (PCM) plus very small header with information about frequency, bits/bytes per sample, number of channels, etc. etc.

http://soundfile.sapp.org/doc/WaveFormat/

PCM (Pulse-code modulation) is digitalized analog signal

https://en.wikipedia.org/wiki/Pulse-code_modulation

 

Later, you would have to load two such .wav files (with the same song recorder twice from various sources), in some audio/sample editor software on desktop computer, and overlap them, slightly adjust offset of one, and compare their waveforms, subtract one from another (this would reveal where are differences).

 

Edited August 9, 2017 by Sensei

[Tub]

Thanks for all your help, Sensei. I have downloaded the free Audacity digital audio workstation and can use it to analyze the waveform and spectrogram of two ( or more ) wav files playing simultaneously. I have plenty of wav files on my PC so all i need to do now is find out when my radio is going to play something i already have on file: i can do this by monitoring my weekly radio-programme listings magazine which, for a specific radio station, ( BBC 2 ), lists forthcoming  choices of  10 " Tracks of My Years " that are played at the same time every weekday, and which are usually very common popular songs by well-known performers. The same station also has an " Album of the Week " feature so i know on two counts what songs/singers are going to be played so i can record whatever is appropriate. ( It won't be piracy as i would have already purchased those same songs that i'd be comparing ).