Music producer Rick Beato has an awesome YouTube channel if you’re into the composition, production, and history of rock music. Here’s one episode of his series ‘What Makes This Song Great?’ that now has 77 episodes and has covered a wide variety of acts like Blink 182, Tool, Led Zeppelin, The Police, The Rolling Stones…
Here’s his episde for Schism by Tool:
I encourage all you music buffs to check out Rick’s channel - something for everyone there.
Great channel! I found him when I wanted to see if someone else noticed that The Planets sounded like Star Wars…or the other way around I guess
I stick around for the air drumming
I’ve been following him for a while. He blew up from 500k subs to way over a million pretty quickly.
You get a good mix of music theory, anti-Apple banter, entertaining musical top 10 vids, and fun song analysis.
I always found this one interesting.
I mean the only real complaint I have with that video is that he has more capable speakers and instead uses m50x. I think that’s the real limiting factor with his lossless vs mp3 test. On any proper studio monitor the difference is pretty apparent
I’ve been doing a lot of thinking about this “HD” vs “SD” vs compressed audio. Rick’s video certainly makes you think and his providing of the engineer/producer’s point of view adds to the conversation. I like Rick’s channel, but I don’t think he’s given this issue its fair and full consideration (as many in the music/audio industry don’t appear to have done so either). This may get long so I’ll try to be as brief as possible with a an enumerated list:
Framing “SD” vs “HD” vs “MP3” as a question of better is misleading. They do sound different, but asking someone to pick “better” adds in the extra variable of personal taste. The NPR study does methodologically ask participants to pick which one is the wav file and which is the mp3 (at least it did in the multiple attempts I did), but the introductory text poses the question of which file is better - either explicitly or implicitly. Full disclosure: I took that NPR test myself at least a dozen times and never scored below 5/6, and sometimes got 6/6, using very modest gear at the time. The track(s) that threw me were hip-hop tracks. A Jay-Z cut in particular was VERY difficult to distinguish between the formats. That was some time ago and the track list in Rick’s video looks different than when I took the test.
We as humans are prone to all sorts of biases. A relevant one here is that we tend to let our first and or most common experience(s) with a thing shape our expectations of it. In audio, one of the reasons vinyl has never really died and why “boomerphiles” often laud it as the superior format is because that’s what they grew up listening to, so that’s what sounds right to them, despite all of the empirical evidence that many digital formats offer higher fidelity. In 2009, a Stanford professor did a study on the “iPod generation” and found that college students actually preferred the sound of 128kbps mp3/aac files (https://www.techhive.com/article/160791/mp3_fidelity_preferred.html). Why? In 2009 the iPod had established itself and most of the audio files available during the 2000’s were lower bitrate files from file sharing services or Apple music store.
I don’t think being able to distinguish between music file types is strictly or even mostly correlated with hearing acquity. Certainly you have to be able to hear reasonably well to have a chance on this test, but past a certain level of acquity I think memory and pattern-recognition have a bigger impact. How can you tell if thing B sounds (or looks/tastes/feels/smells) different/better if you don’t have a strong sense of what thing A sounds like? So far, I don’t think most blind A/B methodologies have appropriately accounted for memory and pattern-recognition variables.
This one really isn’t a response to Rick’s video, or this thread. I’ve seen in many places (including elsewhere on HiFiGuides) that some have removed all of the information contained in a 96KHz FLAC file above 20KHz, complete with displaying a graph just of the information between 20K and 96K with caption something to the effect of “See! It’s all just noise!”. My response to that is “What did you expect to see?” In an analogy, imagine taking a page from a Dickens or Hemingway novel and removing all the letters that occur in the alphabet after ‘w’. We don’t use x and z all that often, so we can probably understand what’s left over. But, what if we just put the x’s, y’s, and z’s on a page by themselves? Would we expect that page to make any sense at all? No. We wouldn’t. It would look like “noise”. In physical wave theory (feel free to make fun of my username here), we know that waves interfere with each other. You remove wave information of any frequency, you change the shape of the waveform that reaches the ear. So, scientifically speaking, removing data above 20KHz (or any other cutoff frequency) physcially changes what reaches the ear. That leaves the question of being able to detect the differences. And that circles back to points 1, 2, and 3.
OK, I’m done. Thanks to all who read this rant this far.
Pretty much agree with this lol, especially how some engineers pretty much cut out information near ultrasonic and infrasonic information. Sometimes it is needed to do that for some sounds in a mix, but a decent amount of engineers still run the entire mix through something that removes/rolls off that information. It’s pretty common for some modern music to not have information above 18k or even past 16k.
And I agree that once you get to understand what lossy compression does to music it’s pretty easy to see what’s lacking (also not a huge fan of blind abx if you are not very familiar with the material that’s being tested)
Also having extremely good hearing is very helpful for figuring out if a file actually has info over 20khz. According to a hearing test I went and got like 6ish months ago now I can hear reliably up to upper 22khz. Perhaps even higher but idk. I can’t really tell what goes on up there, but if it’s present i can tell and it feels weird lol