Why some of the Audio quality of music is ruined by Dj X., page 1

IB1starvip Moving Biology on October 8th, 2015 / post 68410
Hi,

Quotes from https://xiph.org/~xiphmont/demo/neil-young.html

Quote: It's true enough that a properly encoded Ogg file (or MP3, or AAC file) will be indistinguishable from the original at a moderate bitrate.


BUT!

Quote: Lossless formats like FLAC avoid any possibility of damaging audio fidelity [23] with a poor quality lossy encoder, or even by a good lossy encoder used incorrectly.


And

Quote: A second reason to distribute lossless formats is to avoid generational loss. Each reencode or transcode loses more data; even if the first encoding is transparent, it's very possible the second will have audible artifacts. This matters to anyone who might want to remix or sample from downloads.


Thus Dj X downloads several (compressed) mp3's from,for example beat port/sound cloud etc (if they have been encoded correctly  they should be indistinguishable from the original. But then, Dj X producers a mix using the MP3's and then (saves) compresses the mix again into mp3.

Then some night club  will record the mix set of some Dj X (along with crowd noise) and again compress that crowd noise mix to Mp3. This mix,the distortion Mix,quite often finds it's way onto music sharing sites like tribal mixes.

Now some   Dj's may not have sufficient knowledge to understand the science behind the music they produce. Thus it is up to us, the listeners, to request that they improve their methods of delivering the music.

Quote: in the past few weeks, I've had conversations with intelligent, scientifically minded individuals who believe in 24/192 downloads and want to know how anyone could possibly disagree. They asked good questions that deserve detailed answers.

I was also interested in what motivated high-rate digital audio advocacy. Responses indicate that few people understand basic signal theory or the sampling theorem, which is hardly surprising


That is the genius of the scientific method. As it would be impossible for any individual to understand all scientific research, we can use a level of confidence to know that any particular scientific theory that has stood up to the rigours of the scientific method,I.E, Trying to disprove the theory( https://en.wikipedia.org/wiki/Reductionism#In_science) has a high level of being close to reality.

But of course, many individuals are not using science to make a decision. And this is where most of the confusion comes from with company’s trying to sell people music formats in "high resolution" stating that

Quote from many websites selling "high resolution music"

YOU WILL HEAR THE DIFFERENCE

When in reality the human ear can not perceive these frequencies (just like we can't see certain frequencies of light). This has also been proved many many times in double blind tests.

Quote: No one can see X-rays (or infrared, or ultraviolet, or microwaves). It doesn't matter how much a person believes they can. Retinas simply don't have the sensory hardware.



And some of these higher resolution formats may even be adding distortions.

Quote: 192kHz digital music files offer no benefits. They're not quite neutral either; practical fidelity is slightly worse. The ultrasonics are a liability during playback.







Music is physics brought to life with biology.
Seraiel Fanatic on October 8th, 2015 / post 68413
I also read that article, and some more, because I wanted to remaster a collection of mixes.

Even though its main theses, that the human ear cannot hear the difference from 16/44 to up to 32/196 for example, because the frequencies added are beyond 20 khz, so not hearable by humans, is correct, it's slightly untrue that i. e. 24 / 196 doesn't have any advantages afaik. It doesn't have any advantages for the person hearing the final version of a file, but in the studio where the music gets (re)mastered, certain effects, like i. e. a "compressor" don't have to be applied with the same degree of preciseness as they'd need to be in 16/44, so for working with those files, 24/196 can make sense.
Ofc., the original recording must be in 24 or 32 / 196 so this applies, taking a normal-quality file, then upsampling it and then resampling it down again can do harm to it. Also there is a discussion between scientists and musicians, where "audiophile" people argue, that it still can make sense, to sample music in 24 Bit and 48 kHz. A higher bitrate cannot cause any harm, and for some very few listeners might hear the difference, that is not true for the hordes of people claiming they were audiophile atm., and the argument is about 44 Hz or 48 kHz, not anything above 48 kHz, like marketing people, (mostly from Apple) , wanted to sell.
Seraiel Fanatic on October 8th, 2015 / post 68414
Learning about how to work with audio-files and what today's possibilites are, is absolutely fascinating. There i. e. is a big group of people get mad when somebody adjusts the loudness of file before it's fully mastered. They claim that it's possible to hear the difference in quality, similar to the people that claim themselves as being audiophile from before.

Again, scientists needed to help out, and the final statement from them was that: "The algorythm to make loudness adjustments is actually so precise, that it shouldn't be used more than 40-60 times on the same file, because then, artifacts could appear :D .

For me this ment, that even if the original record-recorder adjusted the loudness before and after mastering, if the DJ adjusted it again, and if the person who live recorded the DJ also adjusted the loudness, you could still adjust the loudness 35 times without problems :thumbsup: .

Btw., did you know, that with today's programs, it's absolutely possible to cut out the noise of a ringing mobile-phone while an interview is recorded, without the voices suffering so that one would hear it, if one didn't knew about it? There is one file I want to remaster that has crowd noise. This will take horribly long probably and I don't even know if it's possible, but theoretically, one should be able to get rid of a pain in the neck promoter that tries to animate the crowd by talking useless stuff to them so that only the music is left :) .

Watch https://tv.adobe.com/watch/no-stupid-questions-with-colin-smith/introduction-to-noise-reduction-in-audition-cs55/ from the 5th minute onwards to see real magic :wink: .
IB1starvip Moving Biology on October 14th, 2015 / post 68464
Hi Seraiel



quote Seraiel "it doesn't have any advantages for the person hearing the final version of a file, but in the studio where the music gets (re)mastered, certain effects, like i. e. a "compressor" don't have to be applied with the same degree of preciseness as they'd need to be in 16/44, so for working with those files, 24/196 can make sense."

Yes I've read that using higher bit rates/Hz is beneficial for production also.

quote Seraiel "to sample music in 24 Bit and 48 kHz. A higher bitrate cannot cause any harm, and for some very few listeners"

Sounds possible, though in double blind tests, comparing samples of CD quality  with  professionals and casual listeners the overall net result has been 50/50.i.e - guessing.

quote Seraiel "Watch https://tv.adobe.com/watch/no-stupid-questions-with-colin-smith/introduction-to-noise-reduction-in-audition-cs55/ from the 5th minute onwards to see real magic :wink: ."

or good I'm fed up with all this pretend make believe magic ;-) .
Music is physics brought to life with biology.
IB1starvip Moving Biology on October 14th, 2015 / post 68465
Essentially the evidence suggests that good quality mp3's,ogg's etc are indistinguishable from lossless material provided that they are well encoded from the original material (no a Dj mix using mp3's!)

But if your not convinced do the experiment for yourself. Be aware that you may want to hear a difference. Thus to eliminate the placebo effect use some software so you don't know which track is mp3 (for example) and flac.



You'll have to make sure though that your system is compatible to play high bit rate/freq sound.
Music is physics brought to life with biology.
Seraiel Fanatic on October 17th, 2015 / post 68489
Funny:

When I listened through all the tracks for the James Holden Megapack I am preparing, I accidentally found 3 files, 2 being 2h and 2h and 10 minutes, and the other one being 4h and 10 minutes. I guessed, that this was the same set, only one time in 2 files and one time as a whole, because the length was fitting so exactly.

I first jumped through the tracks quickly, and found out, that Part II was actually in the beginning of the 4h long set, and that Part I was from the middle to the end. There were breaks in the music at exactly the right times, so there was absolutely no way for me to tell, if the 2 2h files were in the correct order and that the set was originally recorded as 2 files, or if the 4h long sole version was the original one, that had been cut into 2 pieces and falsely labeled by some n00b :-D .

I had to find out:

Listened to about 5 different small parts of the track, I somehow "felt" there was a difference, but it was impossible for me to tell which version was really better. Then I picked a 2h minute long passage and listened to it about 50 times each :-O :

I first tried to listened naturally to the passages. I tried to "feel" , the 4h long version somehow felt 'wider' while the 2-part-version felt 'narrow' , I wasn't sure though, if 'narrow' was just 'clean' and 'non-clipping' and 'without extra reverb' , or if the 1-piece-version was really better. I noticed small "things" , like "a few drums that stood out" or "an echo of the voice" or a highhat that sounded like a "sshhhhhsssssssst" in the one version, and like a "tsch" in the other one.

After about 20-30 times listening through the same passage, I got more and more sure, that the one-piece-version had the better quality. It's really very hard to explain this in any words, but it was like "a girl that's good looking and that you're friends with since elementary school" , or "that new, superhot fellow student, that goes to the same courses that you go to, and that sits somewhere in the front of you in the lecture, where you can see her well, making you totally miss out on what the professor is talking about" :wink: .

At about the 30th or 40th time when listening to the same 2-minute-passage, I started to realize, that the 1-piece-version had pieces of sound in it, that I completely not was able to hear in the 2-part-version, like a "click" that I "felt instead of hearing it", a baseline that moved "from right to left and back" instead of simply becoming more silent or more loud or like a sound-effect the DJ applied to the music, that was "outstanding" instead of going past me with me noticing that it was there, but without my attention getting drawn to it.

At the 50th time, I was very sure, that the 4h-long 1-piece-version had the better quality, not judging why, and that the 2h-long 2-piece-version was not simply "different" , but really "inferior" . I had to use headphones for this, even though I got a very good soundsystem I would never have realized the small "details" that made the 1-piece-version just a completely different piece of music after listening with such an enormous focus and to only such a small part. If comparing the two versions, it was similar to the experience I made when changing from a 200€ expensive very well rated HiFi system, to a 1000€ professional PA, those pieces of music weren't similar anymore, it was like one had twice the amount of accustic information and detail, than the other one had, one was "transparent" , while the other one really lacked details.

In all defense, both files were between 128 kbps and 192 kbps, so exactly where audio-professionals say, that that is the threshold, where listeners hear a difference, where music starts to sound "good" , even with having a lossy format (like .mp3, .m4a and all the like, so "non-FLACs" ) .
From the time where I got my 1000€ PA, I still know, that I mostly stopped listening to music below 192 kbps, and really began to value music encoded in at least 256 kbps, preferably 320 CBR or most preferably FLAC.
When I listened to the music in the video you linked, I had absolutely no problem of identifying the fitting piece, and I'm sure, that if I had repeated the test with that utility (which I don't have, which I why I didn't do it) , that I would have had at least 8 or 9 of 10. If one really listens, two pieces of music really aren't the same anymore, just think of the example of the girls from before :wink: .

It's really unfortunate, but some DJs are just so good, that I still download everything from them, that I can somehow get. When looking at the past year though, I mostly or even solely listen to the pieces that have the standards that my soundsystem can display, so 256 kbps, better 320 CBR and are really always special, I listen to those, when I really want to hear something special and outstanding. I'd appreciate very much, if more DJs would really pay attention to the way which they encode music with, but that also especially goes to the people recording it and publishing it on this page afterwards! I heard, that some idiots said, that music above 192 kbps is not distinguishable from higher quality like 320 kbps, but that is absolutely not true. I'm not audiophile, only trained and focussed, and I can hear the difference between 192 VBR and 320 CBR, and moreso I can feel it, otherwise I couldn't have explained it with the metaphorical language that I used! It's really very sad that people say such things, because it tells, that they actually don't listen to the music attentively, because if they did, they would hear the same differences that I hear.

Now I'm really astonished, at the wall of text I've written, but hope, that we'll start seeing less 128 kbps streams and more 320 CBR or FLACs, because disk-capacity is really the easiest thing to get nowadays :shutup: .
Seraiel Fanatic on October 24th, 2015 / post 68569
This ABX-comparison-feature is really great, thx for making me aware of it IB1.

I transcoded some .m4a's to .mp3's. I chose a way higher bitrate for the mp3's, because space didn't matter to me, I just wanted them in mp3-format, so i. e. 69 kb / s .m4a to 192 kb / s mp3.

The astonishing is though, that I could hear the difference between the original and the re-encoded version. I would never have suspected this, I thought that with choosing such a high bitrate, nothing could get lost, but the generation-loss-effect is really heavy, so:

Don't transcode / re-encode / re-compress unless you really have to!

I informed myself a little bit on the current audio-codecs, and it seems, that m4a is currently the best codec. Anyhow, the differences are so minor, that converting everything to m4a would be totally idiotic. m4a mainly has an advantage wen re-encoding files multiple times, there .mp3 fails horribly and has a terrible degradation of sound quality while .m4a stays ok. In listening tests from 2013, .m4a, .mpc and .mp3 were rated almost completely equal. The only thing, that really costs quality, is using an outdated converter or choosing a too small bitrate, so when you record, aim for 192-256 kb / s minimum, because that is where sound becomes transparent. 128 kb / s is really not enough, I promise you, I can hear the difference between 128 kb / s and 192 kb / s 20/20 times.
Seraiel Fanatic on October 25th, 2015 / post 68575
Today, I found out that there are even bigger idiots, than I thought. Upscaling 128 kb / s to 320 kp / s, doesn't make any sense, but at least it only does little harm to the file. On Mixcloud however, there are people that transcode beautiful 256 kb / s MP3's down to 69 kb / s m4a's. ARGH. It's even better: They rename those files, give them a new date, upload them, and everybody thinks that this was a real performance of James.
IB1starvip Moving Biology on November 2nd, 2015 / post 68719
Hi Seraiel ,

Good work  :-) -  testing it out and coming to your own conclusions.

Another free ABX comparison tool  https://lacinato.com/cm/software/othersoft/abx   ( cross platform)

And some useful info on performing your own ABX comparisons.

https://www.homebrewedmusic.com/2014/07/30/a-new-abx-tool/
Music is physics brought to life with biology.
Seraiel Fanatic on November 6th, 2015 / post 68749
Hey :) .

Thx :) .

I btw. just found an absolute genius tip on superuser.com : Sometimes I just know, that I want to hear something in at least xx kbps, but if I really like a DJ, I don't only download his high quality recordings, then I download everything, because there are also times, when I don't care that much about the quality of a recording, also, if a mix is really good, then I might accept, that it's in 128 kbps, so what I want to express with that, is, that having recordings in different bitrates is absolutely normal, but with a great soundsystem, it sometimes has to be 320 kbps. Now how do you see that?

1st way: Use a media library (bad) . Bad, because when you download to the same folders that belong to the media-library, your audio-player will update it, whie torrent writes files to it. This easily causes more load on a notebook harddrive, than the harddrive can manage. Also, when the media-library updates the info on a file, it gets used, resulting in a block of write access for the torrent process, so the torrent client cannot write towards a file that currently gets scanned, so the media-library can update. This resulted in a great number of I/O errors in my client lately. Also, using a media-library is generally superflous, because files and folders have names. Metadata is ok, but to find a file, a proper folder-structure like i. e. [music type] \\ [arist] \\ [year] \\   is more than sufficient.

So what to do? Very simple: Add the bitrate towards the culomns of the music-folder. Then open the folder-options dialog, and apply that setting to all subfolders:

https://superuser.com/questions/15087/windows-7-add-column-to-all-music-folders

Like that, the normal Windows folders can be adjusted to the personal wishes. One will instantly see how good the quality of a recording is. One can display any sort of metadata one wants, just like in a media-library, creating the possibility to completely disable the media-library and by that enable the possibility to directly download towards the right directory, without I/O errors, while still having the exact same functions, only then in the much faster Windows Explorer, instead of the resource-limited audio-player!

I know, much text for a simple tip, but I'm euphoric. Add the length metadata towards the columns, and you'll instantly see the name of the recording, the length and the bitrate! You know the artist, because you know which folder you opened. Display the year metadata, and you don't need subfolders for the different years, unless someone has produced so much over so many years, that the other option would be to have 300 files in 1 folder. Then you can decide "I now want to hear something from James Zabiela, I want it to be a long recording" , go Jambes Zabiela folder, sort by track length column, pick a recording with a good length from the top and choose by bitrate, then you got the longest recording with the highest quality with only 3 clicks!

Sry, I'll shutup now.
you cannot post in this forum.
click here to to create a user account to participate in our forum.