It’s not authenticating the material to be “true” for exactly the reasons you state. It’s authenticating that a stream has not been manipulated along the way.
A waste of time that provenance could have solved. Also, many users might spend time debugging their system or thinking that recording just sucks. Not everyone is as skilled as you are.
They don’t need to pay a penny. The future is streaming according to all the labels.
Yes. Will you make it happen?
To listeners, I believe it has value. To the labels, it might actually mean a loss of control if the artists start to get involved directly. Anyway, all that can be changed if someone has the desire to push it through the system. Bob is trying… Maybe he needs competition?
Anyone involved with moving content around at scale knows mistakes are made. We’ve seen enough of it.
In the early days of Roon+TIDAL, we were catching AAC files streaming on HiFi tier. We still see them once in a while and TIDAL reports them to labels and they get fixed. How do they get fixed? Unknown.
A download store in Germany (highresaudio.com) tests and invalidates files that have suspected upsampling. While their techniques may not be 100% accurate, they do throw out content.
I have other personal experiences but I’m under NDA for those examples.
If you don’t care about provenance, why do you care so much to be so passionate about being against it? What is your agenda?
If it’s like @evand and you don’t like the MQA baggage, that’s fine, but then stop polluting a topic about MQA + provenance. Take it to an anti-MQA baggage topic.
FLAC files have an internal checksum that can be verified at any time. So there goes that argument for places like Qobuz and getting content from the labels.
AAC files getting passed on the HiFi tier is not a provenance problem. That’s a bug…
A download store checking for tracks that have upsampling is not a provenance issue. That is a problem that happened at the label which could also happen with MQA.
You have an incorrect understanding of what that checksum does. It only validates file integrity, not integrity of the content against some original source.
So how do you know that FLAC that replaced it wasn’t just provided from decoding the AAC? I’m not saying it was, I’m asking how you may know. A digital signature fixes that. When I speak of “mistakes”, this is one such class of mistake.
Once again, I believe you have an incorrect understanding of the technology here. Upsampling of any digitally signed data breaks the signature as it constitutes a change in the stream. This is a property of the cryptography used, not of MQA. MQA just happens to be an implementor of this method of authentication. If you trust the signer, the only place for any funny business is in the chain of custody before the signer signed, and because you trust the signer, anything that happened before signing does not matter. Concerned about signatures from manipulator of the data? Read above where I wrote about about signatures given by a big label vs signature by Bernie Grundman.
Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
62
Rhetorical question? Pointing out that it would be a lot of work? Setting up the PKI infrastructure that would be required as well? Maybe we should just let Bob Stuart do it?
My background is more the IETF and GNU open-source communities. So I’d be inclined to (1) sign up for the flac-dev list and see what’s going on there (not a lot, apparently), (2) float the idea there to get some feedback and discussion, and (3) write up a draft.
There is multiple projects that try to get the signature into the blockchain. I wonder if any are making progress. Do you know?
MQA seems to have made the most in that they have the majors.
Yes, but I’d like to be pleasantly surprised.
Our end of the project is not so bad, it’s the wrangling at the labels and the artist that’s the biggest barrier to success.
I can’t think of a more patient and persistent man to get it done.
Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
64
Nope. Using a blockchain seems overkill to me for this.
Yes. But FLAC, after all, wasn’t designed by a “major”, was it? I don’t know who funds xiph.org. So, work out the design, register the metadata block, contribute patches to the flac library to incorporate it, add extra tools to support its incorporation in workflows, and see what happens. If it’s valuable (people are willing to pay extra for it) it will be used.
Heck, there are enough retired audio and software people hanging out on this forum to do a pretty good job at the design, I’d think.
I’m all for provenance, but putting that content through a DSP hamburger machine aka MQA fo get some provenance is not a solution. Knowing that johnny the copier boy at studio xyz cued the album for MQA conversion is of little reassurance.
Sorry, I think I misunderstood you before because I assumed you meant core decode when you said unfold.
If the DAC is rendering, only the source rate makes sense. Showing the unfolded rate is not accurate.
If not rendering (our case is pure decode), then it makes sense to show the core decoded rate + source rate in case rendering does occur.
…and this will fall out of style once a proper presentation of the data is available. See above about my comments related to big labels as the trusted identity vs someone like Bernie Grundman. When you move mountains, this is the type of thing that results in success.
Yah yah MQA baggage. Sorry, heard it multiple times above and it’s still as off topic as it was then.
Is there any available information about how frequently a man in the middle issue comes up for streaming audio? Meaning, how often someone alters the files we stream.
I’ve never seen an AES paper on this, music industry article on this, or really anyone talk about it other than MQA.
Sure I’ve run into a 22 kHz file from Tidal, but having the nuclear option of MQA (and all that comes with it) to solve this is kind of strange. For all I know this was the version sign-off on by the provider.
Also, does any other content providing industry have a provenance issue and if so how is it solved?
1 Like
Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
71
Well, we’re all about to have a provenance issue. Deep fakes are real and getting worse. So while the value of authentication is still kind of marginal, it will increase. Do you really want a FLAC track which when played backward [edit: how would you even do that?] says, “John is dead”? Will be easy to make in a few years. Think of what you could do with lyrics.
It would be quite strange for the streaming providers to ingest deep fakes from sketchy sources, but I’m sure YouTube will be full of them. I don’t believe YouTube is interested in provenance or stopping anything that could make it money however.
films had watermarks in the off-projector areas – i’m unsure how they do it digitally nowadays.
art has always had issues.
Sure, but how did you even know? Something told you it was that non-redbook file – otherwise, you may think it was just a shitty recording/encoding. How do you know this isn’t currently happening with the majority of the content that is a higher rate than redbook?
I’ve not seen deep fakes, but strange things happen.
In the long past, for a download store we did for HP, we saw download material that had file tags from a piracy group.
Imagine if a junior ops person accidentally deletes a FLAC file meant to go to the repository. To save their ass, they may grab torrent of MP3 and transcode to FLAC. I saw worse happen with interns on Wall St, and this is just music, not tons of actual money.
While I don’t think deep fakes are real, I do agree that we are about to enter a world where digital fakes are going to become far more prevalent that we can imagine. Those fakes will enter the ethos as the real deal.
Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
74
Sure we’ve all seen exceptions to the rule and as you guys frequently say, let’s focus on the big stuff and work on edge cases when there is time.
I just haven’t seen any compelling or for that matter any evidence at all suggesting a man in the middle issue is real. To me it’s about as real a voter fraud and MQA is equal to voter ID laws.
I’ll ask again to anyone who thinks provenance is a real non-edge case issue, can you help me find data and facts showing this? Honestly.
2 Likes
Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
78
I, for one, will welcome the unexpected appearance of previously unknown Benny Goodman Sextet albums!
