PG Music Home Forums General Discussion Recording, Mixing, Performance and Production "Gremlins" in my audio track

I have an audio track downloaded from YouTube that contain what I call "gremlins"; unwanted audio artifacts that don't exist in the original YouTube video. I have seen this before and have just lived with it. Now I'm smart enough to at least ask the question if there is a way to remove them without seriously disrupting the audio that I want to keep. In other words, a "cure" that is worse than the "disease" is not a cure.

I'm pretty sure these gremlins are a result of the downloading of the video and may involve the process of converting the video file to an MP3 file.

Here is an image of a section (roughly between 4:14 and 4:17) where a prominent gremlin lives. (I call him grandaddy troll #17 ) There is no visual anomaly that I can see at this timescale in the waveform and the gremlin is not described as white noise but rather it sounds more like a ringtone on a phone. From what I've seen over the years each gremlin has a somewhat unique sound profile. And is somewhat faint, so a good pair of ears are needed to hear it; but it's bothersome enough to me.

For any mixing pros out there, have you seen this and have you been able to solve it? There is a ton of EQ capability in Studio One so I'm hoping I already have the required tools.

I don't think I can post the MP3 file because I don't have the copyrights. But I can send it privately.

Thanks for any ideas.

Why would you convert a video file to mp3 and how? Assuming that you know this is not possible, what are you really doing and how are you converting the audio?

Yes, we would need to know the steps you are taking and what software and settings you use. I’m assuming you are recording the audio while watching a video.

I’ll pm you my email and you can send the mp3 file.

Why? So I can operate on the audio in my DAW.
I don't know this is not possible. I know this is possible.

How? RealPlayer

Is this on a Mac?

No.
Win 11 PC

Hello Steve,
So there are numberous apps to transfer the Audio Only from a Youtube video. I know this because I have used most of them. And most of them work well. But since this is a sketchy operation it is not unusual for all of these to fail for one reason or another at one time or another. As a result, after years of hit or miss, I settled in on what I find to be the most reliable for me.

I use Audacity, a free PC Recording application. What it does is simply record to audio what ever you are sending to your PC speakers. So, first I configure Audacity to record my Speaker output, then I arm the Audacity track. Hit record on the Audacity Track and immediatley hit play in YouTube. After all is done. I then save the recorded Audacity track as mp3 or .wav or whatever audio format I care for. Drag and drop that into Reaper and all is good. If there are any audio glitches, I simply repeat the process.

The only challange with this is the Audacity configuration can be tricky at first. So I have written down exactly what I need to do. Fact is if I only use Audacity for this single purpose, I just set it and forget it. That is until I have an update and all settings go back to default.

I tend to do those kinds of things from Linux and I haven't done it for quite some time, but I believe there Windows-runnable versions of both youtube-dl and yt-dlp. I know not if they have any artefact issues. Maybe worth a look if Dan's method doesn't work for you.

Dan, has anyone ever called you a walking library?

As it turns out I have attempted this with Audacity in the past and I think it messed up or disconnected my audio drivers. And I found that RealPlayer [usually] does this well with a small footprint on my other applications. These "RealPlayer gremlins" happen randomly and only on occasion. I'll guess that it's 95% reliable.

I sent Matt a snipett of the problem and will see if he comes up with anything. If not, I think I'll just re-download the video with RealPlayer again and I'm almost certain the portfolio of gremlins will be different and hopefully, non-existant.

My thinking behind posting this problem is it may be possible to "filter" the audio file in some way, perhaps with a low-pass filter or similar. I also thought that someone here might have a contact at the makers of RealPlayer that could shed some light or at least put this on their bug list, assuming RealPlayer is the problem.

OK, I listened to the audio snippet Steve sent, and my equipment is quite good.

I do not hear any noise or pops or distortion etc., the sort of audio glitches I might have expected to hear. I DID hear an odd bit of unexpected audio, almost like someone in the background playing a rinky-tink piano. Is that what you call the 'gremlin' here?

Please give me a link to the YouTube audio and tell me the time stamp corresponding to the audio snippet you sent. I want to listen as carefully to the source.

First wild guess: does your PC have any other audio programs running that could have contributed to this, such as a streaming audio player like Pandora?

When one download application glitches another download application may not. +++ HERE +++ is a Windows YouTube to MP3 application that works well for me.

Like Dan I am also fond of using Audacity's "Record What You Hear" mode which works very well.

Dan, I use that method also! I also use it for recording records to MP3s.

I guess great minds do think alike!

Record "what you hear" sounds like the old Creative Soundblaster units. I actually miss that function.

I record whatever is playing audio by routing the soundcard output to Adobe Audition, a product similar to Audacity.

In the YouTube video, I hear the same rinky piano at that same place, in the left channel, exactly as in your audio snippet. If that were me as the audio engineer, I would suspect somebody left a fader up accidently for a second. Of course, I think engineers in the day used to leave those little easter eggs for folks to anguish over...

I think your software captured the audio correctly, at least as it is on YouTube (which does employ its own compression algorithms and who knows what could happen, but I don't think this). Either that, or you are hearing something I'm not.

Given that it is not terribly loud, you could perhaps use a plugin to put a Gate on the audio and hope it kills this without altering anything else. Touchy adjustment, though.

An interesting update.
I just re-downloaded the same video using RealPlayer; this time NO applications running except my browser. And once again, the same gremlin at the same timestamp is present.

This I did not expect.

I wonder if this is an intentional "watermark" embembeded into the audio file for some (unknown to me) purpose.

I do not know the tricks YouTube might use, but I could understand if they or the publisher or ?? desired to add this little snippet of sound as a digital watermark. The irony is that YouTube is the world's biggest offender.

You are absolutely right Matt . . . the YouTube video itself is flawed. Here I am assuming that an "official" video would always be clean, pure and error-free!

I think this might exonerate RealPlayer and we certaily can't blame Carol.

Leave it to me to uncover interesting problems . . .

Ya starting at 4:11 in the YouTube Video...,

"...winter spring summer or Fall, ...All you got to do IS CALL..."

I think I hear a bit of sonic disruption in the left speaker under the vocal IS CALL. This would be something only the recording engineer should catch.

OK folks, so I'm full circle back to my original question for you Professional mixing engineers.

Is there a way to isolate and surgically remove this bit of musical cancer without killing the patient?

So 3 guys walk into a bar. One named Ring Tone, another named Rinky Tink Piano and a third named Sonic Disruption.

Q. What do they all have in common?

A. They each ain’t got a friend! Ha Ha

As I mentioned, experiment using a gate. You set it to exclude sound below a certain volume, and hope it leaves most of the good stuff intact. If you employ this over the one or two-second snippet, it might not even be noticed.

On the other hand, it took some targeted attention to even hear it in the first place.

About why this is there, does anyone here actually have the original recording to see if this is on that?

Yeah, OK. But what color are they?

The phrase "...winter spring summer or Fall, All you got to do IS CALL..." is repeated starting at 1:22. A little cut and paste should allow you to eliminate the artifact.

I've been wrong before, but I'd be VERY surprised if this gremlin is on the original Tapestry album. No way could such a glitch make it past all the artists, recording engineers, mastering people, producers, marketing people and others involved on that album.

As a matter of fact, if I look hard enough in my music library I think I can find another example of this.

FWIW, here is what my AI bot says on the subject.
"I couldn’t find any information about glitches or artifacts in the album Tapestry by Carole King. However, the album was certified 13× Platinum by RIAA and it is one of the best-selling albums of all time, with over 25 million copies worldwide. It received four Grammy Awards in 1972, including Album of the Year.

Is there anything else I can help you with?"

Maybe. But remember the state of audio quality back then.

When I pressed it a bit more I get this.

"It is possible that glitches or artifacts may appear on an “Official” video on YouTube. There are many reasons why this could happen, such as issues with the video’s encoding or compression, problems with the video player, or even issues with your internet connection."

So.... I listened closely, using my good ear buds. No one has gotten the correct answer yet.

As professional and particular as Carol King was with her music and the professional players on her sessions, it is doubtful that anything in there is a gremlin or an easter egg or whatever. If it's there, it is there on purpose. And yes, it's there, and it is actually supposed to be there. Listen to it again.... Carefully.

In the first place where "is call" happens (1:29) you can hear an acoustic guitar run occur in the left channel. It's very low in the mix
In the second instance where "is call" occurs (3:05) there is nothing going on with the acoustic guitar. I didn't hear it in the second one.
In the third instance at (4:13) you can clearly hear the acoustic guitar play the same/similar run as in the first instance however this time it is played in a more aggressive manner. There are more highs in the tone and that makes it sound like the "gremlin" you thought you were hearing.

In listening to this, the only thing I found that was "out of sorts" was the out of tune....slightly sharp.... bass. Not all the way through, but in some places. listen at 1:41. I tried to look it up but this sounds like it might have been an upright bass and that would be par for the course to not hit the notes dead on. Song character and yeah... fixing that would not improve the song itself in the least bit and most people will never notice it anyway.

So no..... it's not a glitch. Listen carefully and I think you will agree. It's an acoustic guitar playing what it's supposed to play.

I think I have the original on vinyl. I'll check it out and report back.

Check Back: I couldn't find that album in my collection. Doesn't mean it isn't there... I just haven't found it. I have to rearrange my vinyl into alphabetical order by artist and it's badly out of sync from my move to the studio 6 years ago. Yeah, I can absolutely be THAT lazy.

There's a disconnect that no one except I has addressed. What you showed does not convert video to mp3 because, as I stated correctly, it cannot be done. Telling us that you are doing it doesn't change this.

Try using Real Player to compress the video—per the settings you have checked—and save as a copy. Now go back to the original file but this time, check the box for Export as Audio only.

There are many apps that will compress this file to mp3. You want VBR (variable bit rate) these days. Now strip the audio from that compressed video and substitute your new mp3. See if the result is not more to your liking.

I am not up on the free and cheapware for Windows. A commercial app with telephone support that does everything you need is Adobe Premiere Elements. This link is to the Adobe site and you can download a 30 day evaluation. There are resellers who offer better pricing with or without Photoshop Elements bundle – if you can get EDU pricing, so much the better.

For three years during the pandemic, I averaged 2–5 videos a week for my church clients and every one had the audio processed separately. There are no subscriptions — use the permanent license till it just doesn't do the job anymore then buy the latest version. I used Premiere Elements 2018 till a few months ago when I updated to 2023.

I am wondering if Bass Thumper means extract the audio from a video and convert to a MP3? I do this all the time in Studio One Pro. I can import the video track and extract the audio to a track of it's own. Then simply export the audio to a MP3. Just a thought.

Nope. All 3 quotes I give above use the term "convert" or "converting" not "extract".
They use the term convert, so I use the term convert.
In any event, RealPlayer does this well.

Steve, I think Mike in his always helpful and positive way might have been more interested in making sure the terminology was the most important thing learned here, whereas the rest of us seemed to somehow understand what you meant and were primarily interested in helping you solve the problem.

Y'all are arguing over technical issues.

It's not a technical glitch.

It's an acoustic guitar you are hearing and it's supposed to be there.

Listen again...carefully.

I'm not arguing.

I listened carefully, gave an opinion, and then asked if anyone had the original recording.

For whatever reason, the OP apparently wanted it gone. I answered how he might accomplish that, and Dan offered another method.

From reading this thread and your complaints, I don't think so — or you don't know what you are doing — or you are not understanding what people are telling you as they try to help.

My vote? All 3 are correct.

BTW, not only do I have the copy of Tapestry that I bought in high school when first released, I also have the MFSL half-speed vinyl. CDs weren't on the market yet for either release.

YouTube process audio when uploaded—even the official releases. This often adds artifacts and distortion. Further processing cannot make the signal better—ever—but can certainly make it worse. This is why I recommended getting ahold of an original copy and bouncing that down.

"You Got a Friend at 44.1kHz/16bit on the 2008 remastered CD is 54.7MB .wav file. It's 9.1MB as a VBR mp3 and 4.2 as 128k mp3. mp3 settings are Joint Stereo, best quality. I also have it as a 27.5MB m4a ALAC for iTunes. The sound is barely acceptable at 128. I bounced these down last year as I prepared the music for my 50th high school reunion so no new work here. I do all of these bounces in TwistedWave where I have control over every aspect if I need.

A digital capture of the "Official…" audio from YouTube reveals a 9.9MB m4a file as expected. It is bass heavy compared with other versions and there are a number of artifacts — but they match noises from James Taylor's guitar, percussion or damper noises from the piano. The 128k that I did has many of the same noises but they are more musical when bounced from a much cleaner recording; the bass is much better on the 128k, also.

Likewise, the 128k mp3 bounced from "Official…" YouTube m4a is crap.


I get that, because the tools I use cost a little money, you won't consider them. I, however, am not the one with the problem.

+1

Please go back and review the original request. He said "but it's bothersome enough to me". It does not matter what the sound is. It does not matter how it was recorded, uploaded, or converted to audio etc. He said he wants the sound gone. I gave a method to do that. So did Dan.

Mike, you are still not getting it.

It's not about the money. It's about your credibility. You cannot tell me that I can't convert video to audio using RealPlayer. Just like you cannot insist that 3+1=0. If you do you've lost credibility in the eyes of anyone with a properly functioning brain.

News Flash: There is something called "truth" and it's far bigger than you and me and everyone else on this planet.

Over the years, hundreds if not thousands of RealPlayer users worldwide have converted video to audio using RealPlayer. Dude, it's a no-brainer. Why are you having such difficulty grasping this and the above 3 quotes???

My time is too valuable to continue this anymore with you.
What are you going to tell me next, that the earth is flat or that photosynthesis is a hoax?

The gremlin lives somewhere between 5 and 10 kHz. The best I've been able to do is put a -24dB notch at 7kHz with a Q = 0.9
This doesn't entirely remove it but it's less bothersome.

Any significantly different center frequency or Q value steps on her vocals too much and the "cure" becomes worse than the "disease".

Very good. I was not approaching it by frequency; I was approaching it by volume. Since it’s so hard to hear, I figured the volume was low enough that you could just use a gate and pass everything above a certain volume. Did you try that?

But, whatever works.

I think the problem is terminology. I can convert water into ice. I can extract water from a grape. But I can't convert ice into a grape. Likewise you can not convert video into audio using Realplayer. You can convert a video file into another type of video file, say wav to MP4.

I don't blame you for using the words you have read on-line.

UPDATE - I just googled

https://customer.real.com/hc/en-us/articles/204295208-Convert-music-videos-to-audio-tracks

and saw what you are seeing. If you have time try this experiment for me. Take a silent video and convert it into audio. If it produces an audio file then I will be convinced and offer here an apology to you for the initial paragraphs. thanx

I get it that he thought it was a glitch and wanted to get rid of it. That's how he was hearing it.

However, it's not a glitch, and if Mike Halloran listens to his vinyl or wav from the CD version I'm sure he'll agree it's a guitar that is supposed to be there. I looked for the record but couldn't find it.

EQ notching isn't going to get it done. That's going to create other problems that you won't want in the song. If you want to remove the "glitch" yes...absolutely you can do it. Get a copy of MELODYNE and use it to show you the blobs for everything... simply find the blob for the guitar and delete it. Works like a charm removing such things. You don't have to do the entire song.... just that 3 seconds of the part.

Yes, that's a good suggestion if someone has Melodyne Editor.

Hi Mario, if you are implying that my terminology is wrong I have to disagree. I also disagree if you are implying that Mike was making the distinction between convert and extract. I have read his posts carefully and as of the time of this writing he has made no such distinction.

If he had made this distinction, I would have replied with what is below and this conversation would have taken a much different path.

OK, I’ll meet you part way on this. If you can convince my 3 quoted sources to use “extract” instead of “convert”, then I will concede and use “extract”. On 2nd thought, I’ll compromise even more. You can forget about Wikipedia and the AI bot and focus on the app developer RealNetworks. I believe they are based in Seattle 1-206-674-2700. I’m sure they will understand ice cubes and grapes But when you contact them to request your terminology change they may very well tell you this:

We appreciate your inputs, however the engineers and programmers of our award-winning RealPlayer software have stated that a proprietary, state-of-the-art algorithmic conversion code is used to convert video files to audio files. Therefore, our management team has carefully and deliberately chosen to use the terms “convert” and “converting” in our marketing, public relations and user documentation materials because it most accurately represents what is being executed by the program.

Like I say, if you or anyone can convince RealNetworks to “see the folly of their ways” then I’ll adopt what they say. Until then I have to hitch my vocabulary wagon to the RealNetworks tractor . . . after all, it is they who are the experts on this matter.

I have no doubt your suggestions are solid. I tried several effects such as Gate, Expander and Multiband Dynamics and I just can't remove the sonic disturbance. I'm just turning knobs and pushing buttons and don't have the necessary audio engineering skills. I'm not sure if anyone here does.

That said, this is not a major problem. I'm using the audio file of this song (converted by RealPlayer) only to play to and learn the bass line. And gremlin or no gremlin, the MP3 file has met that need.

I'm also appreciative to the helpful replies to this request for help. I learned something from this exercise.

No. You're not getting it.

Mike is correct and so is Mario.

It doesn't matter what the marketing department or whoever at Real Player says. You can't convert video to audio. They are wrong and misleading.

Words mean things. Sloppy,lazy, haphazard use of language leads to confusion. Being sloppy about the difference between Newton Meters and Pound Feet can damage parts or get someone killed.

There certainly is a thing called truth, and a large part of it is using language correctly.

OK, might I ask what your credentials are to make such a claim?

Were you or are you a member of the RealPlayer development team? If so, supply evidence.

Have you ever written software to process video and audio? If so, supply evidence.

On what basis do you claim a video file cannot be converted into an audio file? Supply evidence.

Extraordinary claims require extraordinary evidence.

I've given you my evidence from no less than 3 reputable sources.

I just record in real time what I’m hearing, but staying in digital. I guess I’d call that copying.

Bass Thumper's original wording was about "converting the video file to an MP3 file".
Note the use of the word file.

What Mike initially wrote was "convert a video file to MP3" then later "convert video to mp3".

The file conversion from YouTube video to mp3 audio is not a lossless process ... it will, of course, lose/discard the video part, but the file is converted.

I agree wholeheartedly about the "Newton Meters and Pound Feet" observation, but I also note that the same damage can be done by reinterpreting the text into something it never was.

The extraordinary claim is the one that video can be "converted" to audio.

Your "evidence" is regurgitated ad copy. None come from sources that would pass muster even at the Junior high school level for a paper. At least not when I was in school.

I dare you to take a college class and cite website marketing drivel and Wikipedia as your sources.

I'll bet the development team rolled their eyes at such nonsense and resisted.The marketing guys should have known better and very likely did, but felt they had to dumb it down for a scientifically ignorant and only quasi literate public that has little interest in such niceties as using language accurately.

In these fights, the marketing department always wins.

Frankly, I don't know why you have to go all monkey brain on this rather than just accept that what you are really doing is extracting the audio from a video file.

A video file (such as.mp4) is a container that holds two different streams of data: video and audio. (There can be other data such as subtitles). The two data streams are separate; the container is what synchronizes them.

I've not read through more than the 1st three posts so may be regurgitating...
Lots of software available to rip audio, (strip the audio from the packet/codex/container), from a video is available & free. I use one all the time.
I always rip to .wav so I don't introduce much in the way of artifacts in my work flow.
MANY Utub videos are uploaded with heavily compressed audio and video...most users take the default options when rendering a video to upload and that means 192 MP3s...the noise, if not in the original recording, starts there. Reaper, and other DAWS, convert the MP3 to a preferred format behind the scenes, much like BIAB converts WMV to .wav for use.
Which MP3 algorithm used is an issue as well...all converters have a finger print and some leave larger smudges than others.
As to removing the noise...that's really problematic. Is it heard in the video form? Have you tried other rippers/converters? Can you find it on an section of otherwise sound free ripped file? If it can be isolated then it can, more or less, be filtered out.
I occasionally have this problem when working with a friend who insists on sending me his parts as MP3. I use a freq analysis VST to work out where the worst of the sound is focused and then automate some EQ to address that freq. only when it happens. IF you have a GOOD de-esser, (the sort that is part compressor and part EQ), you may be able to dial that in to address the issue.
IF there's no gremlin in the video then you know it's part of your processing and you can address it as mentioned way above.
Good luck.

Settle in for an interesting ride and read the whole thread, Ray.

Jim, just to let you know over the past two days Audacity downloads of Youtube audio has failed me completely. I don't know what has changed, but this is familar history for me when it comes to Youtube downloads. So I tried your recommendation and it worked perfect for 3 out of 3 attempts (ya I have a new project in mind). Just have to figure out now if I need to pay something or if it is free?

PS, tried RealPlayer and was not sucessful in several attempts.

Thanks

I did just that and the analysis in Reaper/Melodyn went well.

Here is my current conclusion:

Yes, it is not and sonic artifact but I now clearly hear an acostic Guitar riff,
Yes it is predominately, if not exclusively, on the Left Channel
Yes Melodyn picks the Guitar out and isolates the individual notes.
But to ID and remove just the guitar notes would take a lot of "listening".
By the time I got to this point, I fully understood what the guitar was doing and then I actually grew to like the little lick.

Here is a quick picture. This picture shows the vocals at ~ 4.12 minutes and only the phrase, "..All you have to do is call (Guitar - 16th notes, ditty, ditty, ditty ditty)".

Leave it in or hide it or take it out. That is up to you as the producer of this cover.

Yes I can because it can't be done.

An m4a contains a video and an audio file. They are processed separately as others have pointed out.

Had you actually read my posts and tried to understand them, you would have learned something. But nooooo...

I don't care what you think of my credibility. I have been processing video since the late 1960s at AMPEX. I know how these things work—had you paid attention, you would, too.

I tried Matt,
reading passive aggressive text hurts my brain though so I typed instead.
I ought to have mentioned Melodyne as well. It recognises noise as noise.

Mpeg-4 files of the various formats contain video, audio, subtitles, and maybe other data streams within the same single file. A stream is structured data within that file, not a seperate file.


... or data can simply be discarded if one doesn't want, or cannot allow, that particular data in the output.

If one software calls it extract, and another calls it convert .. who cares?

You quickly learn what it does and go from there.
The technical side of it is kind of mundane. Yes the audio is a separate part of the 'file bundle' at a certain level, but some users don't really need to know/understand that

'Extracting' the audio part and 'converting' it to a new format could make both terms correct.
Just my thoughts
/it likely runs thru a convertor of some sort along the way, even though you are extracting just the audio

I largely agree. Only those writing the software likely need to know the details.

What I care about is that someone is being addressed very rudely for allegedly saying something he did not say.

The person said "convert a [YouTube] video file to an mp3 file."
That is perfectly correct and reasonable.

If the person had said "convert video to audio", as alleged, that might reasonably be considered illogical, though frankly forgivable as "a video" normally has associated audio. If we watch "a video" we normally expect that it will have sound.

Your questions are totally irrelevant. They are a red herring. That's why I'm not going to answer them.

I not only was not part of the design team, I have never worked for Ford Motor Company in any capacity yet I have detailed technical knowledge of how my car works.

You tell me what logical inference can be made from this fact.

Again, I dare you to take a college class and cite Wikipedia in your bibliography as "proof" of anything. Hell, Wikipedia themselves tell you to tread carefully:
Reliability of Wikipedia
Wikipedia is Not a Reliable Source

Drop a video into your NLE and tell us what you see. Better: do as Mario suggested and take a silent video, "convert" it to audio and tell us what happens.

Fact: In your NLE you can separate the video from the audio. You can get rid of one or the other, make changes to one independently of the other and/or move them or parts of them in relation to one another.

Fact: when rendering your project or transcoding it you can independently choose the video and the audio codec.

You tell me what the obvious logical conclusion is

Time to put this one behind us.

I did. Steve and my definitions of convert and extract are different. So be it.
No harm, no foul.

Peace

+1
PM me if you'd like to see if a meeting of the minds is possible.
There's still much to be explored on this subject.

Dan, The link I provided is for the free version of the 4K YouTube to MP3 application. You can purchase a license to unlock additional features but I haven't found a need. +++ HERE +++ is a webpage that explains the benefits of purchasing a license and how much the various license plans cost.

I have nothing against AI, Wiki or RealNetworks.

I have everything against your attitude when you were presented with an opportunity to learn from people who do know their subjects. We posted in good faith with the best of intentions and only kept replying so that you might learn. Trust me, we have learned.

This, however, may be the funniest thing that I have ever read on these pages. Thank you for this laugh.



Any of us who have ever worked in sales, marketing or management know the accuracy of that.

If you think we're the ones who look bad as a result of this thread, that's just another of your many mistakes.

But you do have something against them . . . you won't except what they are saying about RealPlayer converting video files into audio files when clearly they know much, much more than you.

You seem like an otherwise rational guy so let me try this.

If you choose to be adamant about RealPlayer extracting audio then please explain how that is done.
Does their software skip every 3rd byte in the source file and . . . . pooof, audio appears?
Or does it skip every 13th bit and then take the square root of the remainder?
Or maybe the first and 3rd fifth of the file is video and everything else is audio?

I've given you my evidence . . . times 3. Where is yours?

Mike, with all due respect, this is not a kindergarten schoolyard where you can say "My daddy invented paper, walked on the moon and is the strongest man on earth."
Rather, this is the adult world where credibility and fact-based evidence matters; provide some.

It is extracted for one because most video does not interweave the structure of the audio and video so it's not about counting bits or multiplying. They are separate inside the file.
You extract the audio part of a file that is labeled a video file. (video file semantically means 'video' but it often contains audio too, go figger)

Conversely, the audio section of a video file is usually 24/48 in a given codec .. so if the application has an end result of something other than that, then it did indeed convert it (as it extracted it).

Semantics .. some of the worst arguments to get into or suffer through, especially when in some cases both sides can be correct!

It's OK to try to look at it from the other side and think "yeah, I see why he'd say that"
Sometimes it's even true from both sides.

I can see that when I (yet again) asked Mike to explain how the extraction is done that my tounge-in-cheek question was taken literally. So let me try again, this time being more obvious.

Does their software skip every 3rd byte in the source file only if you ask nicely, dance around the flag pole 4 times under a full moon in the rain while howling with lit candles in both hands and . . . . pooof, audio appears?

If you haven't noticed, I'm evidence-driven. Personal beliefs, heresay and opinions don't mean much to me.

Q. As for semantics, could it be that you are the one wrong about semantics/terminology and not RealNetworks?
A. Yes

Here is my evidence. But I am flexible. Get RealNetworks to change their "semantics" and I'll certainly change mine.

Why do I do this. I should leave this alone but...

BassThumper's original question related to changing a YouTube video file to an MP3 audio file.

Mike's response was to say one can't convert video to audio.

But that was not what Bass Thumper had said.

Below is an image of before and after media-info reports for a YouTube mp4 file changed to a mp3 audio file by the simply doing:
$ ffmpeg -i SavedYachts\ 140MB\ Medium.mp4 SavedYachts\ audio.mp3

You will see that it has indeed 'extracted' the audio, but it has then converted from the mp4's AAC LC to an mp3 audio stream, probably in a different coding and certainly at a different data rate, before repackaging in mp3 file format.

One could equally argue that the conversion 'extracted' the video and dumped it.

In either case the new mp3 file is of a different type to the mp4, with different packaging data, so the file has been converted.

'Extraction' may well be part of the process, but the whole process from start to end is a conversion.

A final binary compare shows the two files differ right from the very first byte, despite these both being mpeg files. The file types and the file header information are different.

Ya, your first thought was the best one...

Thank you Gordon.
Here is further evidence that corroborates this from an FFmpeg perspective.

This should now be established truth, settled science and confirmed fact.

Let's all move on. . . whewwww, this took a loooong time.