A "newspaper in the photo" but for audio? Audible timestamp hash/signature/password/watermark? What (live?) audio channel works as an audio timestamp?



Short Version ("TL:DR"):


  1. My parent er landlord screamed and or banged pots and pans by my room from 3h to 7h (3AM to 7AM) and did not stop the entire time (but my door was closed and not open so I could not see just hear), how do I prove that happened that very long-term, with just my smartphone to timestamp?

  2. At exactly 03h07, a heavy machinery WM.com dumpster truck backs in ten (10) meters away, starts revving their engine like a racecar and running their forklift at extremely high decibels, reverberations physically shaking me and my home (and my neighbors, does that matter?). How do I prove, with just the sound/audio signature recorded from my smartphone, that I was woken by that at the brink of dawn 3:07AM?

Long Version ("Extended/Detailed"):

What best practices to prove as evidence that that sound was at that timestamp? Evidentiarily best/wisest is a live stream to timestamp audio? What live audio stream (and is radio/tv news, what about reruns?) should a user use to Timestamp Audio? What is suggested, if I need to be able to prove that a sound happened and (is clockable as proof of) at a certain exact time.

Any self-recorded audio would be the same example, can you tell me exactly what second a word was said in Universal Standard Time for each word?

My personal angle, to use me as just one example, was from asking by search engine speaking "best live audio stream (is radio/TV news?) to timestamp audio?" I meant 'What is the suggested method for timestamping/dating the

audio recording I just made' (note the timestamp, as the trusted time the record began, at the beginning of the YT title and .wav record file name, which you have to trust the time written by me/you the User) and https://steemit.com/csychology/@scribe/qd9npv to prove my experience from seconds ago.

Well because of technological reasons, partly done to follow professional workflow standards, posting process from www.AudioShip.io to www.YouTube.com took/delayed hours for, again, mainly technology reasons, seemingly another Digital Divide beyond what would be the value a "newspaper in the photo" is, for there is not an equivalent for audio. Comparatively, I can send a timestampable sexting message with greater proofing than a single word I say with my audible mouth [scroll feature to track linear sound is emotionally tolling (literally an emotional experience to rehear, a page of text does not bite like an hour of audio equivalently, and the evidence process is generally manual), to make matters difficult] as sound waves, which is very technically provable.

Is radio/TV news verifiable to the second, a secure reference point, what about reruns/repeats? If I change to a radio channel and the commercial is the same as before, that's not proof of life, er proof of timestamp, correct?

I am worried that try to timestamp I just made is not enough proof for the legal system, if I did not activate a radio or TV with some verifiable knowledge 'I know that sound happened at that time because I heard the radio anchor say that the first time then, that was that issue's first ever word, very first words'.

I am not asking Recognition of timing information by sound device about if my sound recorder knows what min/sec I am at in the tracking bar. The issue is proving as a matter of fact that what the "sound device" thinks is the time is not the (painfully sadly for record-keeping standards if) only proof we have to ask for. Why should I (scientifically, evidentiarily) trust my/your tape recorder said that cursed/claimed/alleged/accused/"documented" hexspeak of a timestamp?

I am asking how to mechanically or socially make sure for myself and others that that sound happened at that time. Is that a security feature, do any sound recorders record timestamps of pending audio streams before you're able to save fully to Google Drive, in my test case? Is there an audio channel that works as a timestamp?

Do we need a radio frequency AM/FM/HAM that has, a voice counting a clock every second?, a newscast that never has commercials or repeating sections?, or a constantly changing rhythym and sound or whatever, to 100% know (like a password for sound itself (audible timestamp hash/signature/password/watermark), maybe a blockchain for sound itself), can we verify that sound definitely happened at that time the sound happened, that sound means that timedatestamp?

prosody-Gab Vereable Context

Posted 2020-07-10T21:37:37.267

Reputation: 107

This sounds very much like an XY Problem Why not tell us what you are trying to achieve rather than ask us about how you think you might achieve it.

– Tetsujin – 2020-07-11T10:40:49.643

1This might be a forensics question, the answer to which is is the ENF signal (Electrical Network Frequency) which is used in forensics to locate certain associated acoustic and electrical signals in time for forensic purposes. – Mark – 2020-07-11T14:55:21.660

@Mark By "audio channel that works as a timestamp", you mean there is an ENF signal which works as a timestamp? That would be an answer to my question. – prosody-Gab Vereable Context – 2020-07-11T15:12:56.777

@Tetsujin I need to be able to prove that a sound happened at a certain time. Does that require a "newspaper in the photo" but for audio? – prosody-Gab Vereable Context – 2020-07-11T15:29:33.577

ENF is a technique commonly used in audio forensics. It requires analysts to have recorded ENF signals over a long period of time. I am aware of this technique, but not a practitioner. I can only recommend you conduct further research into this technique and reach out to local audio forensics professionals. – Mark – 2020-07-12T05:10:54.140

@Mark I think your comment is actually an answer, though I think you may have gone deeper than me if you think my question not theoretical. I would appreciate your comment as a solid answer to my theoretical question. Thank you for the answer, and you have my vote once posted officially. (My question is for future sound recording techniques, and not to analyse that particular recording, that's just me citing an example from before of not having a good timestamp and wanting a better solution for next time.) – prosody-Gab Vereable Context – 2020-07-12T05:13:48.027

1thx mate - sorry it's a rudimentary answer - I'm knee deep in code right now so will update at a future point in time. – Mark – 2020-07-12T05:21:25.050

1I'll also mention that a "newspaper in the photo" doesn't prove the photo was taken on a specific date - it proves the photo was taken no earlier than date on the newspaper, but could have been taken at any later date. Any audio clip will be the same, as it could simply be recorded and played back at a later time. Hearing MLK's "I have a dream" speech in the background, for example, only proves the audio clip is from sometime after 1963, not that it was recorded in 1963. – Nuclear Hoagie – 2020-07-21T17:54:38.337

@NuclearWang So optimally you need a "live audio" feed (affirmative?)? That is then part of the answer to the question, so if you know how to elaborate an answer I would appreciate the help defining that for audio/verbal/oral evidence. I figure that line of thinking is the "best a user can do", in reality. To say that another way, I mean your best bet is have live CNN coverage blasting by radio. Most optimally, during an important crisis with many news organizations live recording the same conference... that's (for just audio) even a greater technological challenge than a newspaper citation? – prosody-Gab Vereable Context – 2020-07-21T18:08:15.570

@prosody-GabVereableContext But there's nothing that would prevent you from recording the "live audio" feed and playing it back later, which would be indistinguishable from the true live feed. I can't think of any "self destructing" media that would work by limiting the latest date it could be played, and if it existed, you'd never be able to verify it past the expiration date anyway! – Nuclear Hoagie – 2020-07-21T18:24:46.727

@NuclearWang We might need to clarify/differentiate the requirement for a live audio feed, from the separate requirement to save/upload to the cloud live. I mean there is some degree of trust if I am timestamping a live audio stream, uploaded to a server at said time which the ISP confirms? Same ideal if invented blockchain mark/stamp for Proof of Audio Time, only the method by which time is recorded and then synced to the chain is the proof? In combination, a Live Audio Feed (TV, Radio) and Live Audio Upload (Cloud, FTP), are the best proof? Or is there an algorithm like ENF signal for Users? – prosody-Gab Vereable Context – 2020-07-21T18:34:34.873



This might be a forensics question, the answer to which is is the ENF signal (Electrical Network Frequency) which is used in forensics to locate certain associated acoustic and electrical signals in time for forensic purposes

ENF is a technique commonly used in audio forensics. It requires analysts to have recorded ENF signals over a long period of time.


Posted 2020-07-10T21:37:37.267

Reputation: 7 535

I would comment that your idea is (just currently?) cost prohibitive but that does not change the fact it is a good idea. If that's the only way, I find that astounding. I wonder if there will be efforts to make the "time domain"/timestamp analysis ability common so normal people can establish extremely credible audio recordings. (So normal people who need that kind of credible proof/evidence/assessment are stuck technologically til the gov't open sources the ability of law enforcement to use ENF time domain/timestamp analysis, but for giving personal protection and security if not integrity?) – prosody-Gab Vereable Context – 2020-07-12T05:37:07.303


How about embedding a smart contract into the audio using digital watermarking?

Essentially, a trusted encrypted ledger entry could embed the date (or any particular information) such that later on you could retrieve the value and compare it against what is in the ledger.

Don't know if that's a thing, but it sounds like it could work.

Drew Mills

Posted 2020-07-10T21:37:37.267

Reputation: 31

A quick Google found some audio watermarking tools at https://audiowatermarking.info/awt1_main.php.

– Drew Mills – 2020-07-22T00:41:04.180

I was thinking the process maybe as a https://sound.stackexchange.com/a/48548/19675 "ENF blockchain"(?), an Electrical Network Frequency backed by the blockchain, where the sound (of a single word at a certain time, for example) is recorded with an analyzable audio timestamp then hashed on chain. (I did not think to call the idea ~"smart contract audio watermarking". That nearly implies the same sort of idea I was trying to express by saying the word "blockchain" and "audio" together, but seems for Digital Rights Management mainly, interestingly not sold as specific Audio Timestamp Products.)

– prosody-Gab Vereable Context – 2020-07-22T02:12:13.697