24/192 Music Downloads are Very Silly Indeed

I think this is a good article about the subject. I don't recall us discussing this article before, but if so, my apologies in advance.

24/192 Music Downloads are Very Silly Indeed

The upshot is that 44.1khz/16bit is all we really need for listening.

Sorry schreiber9, dirk wright is correct. The difference you hear is slightly different mix as these hi res companies are doing to literally foul you. I bet you can't hear anything above 15khz with -3db and noise of 24bit?? Yeah good luck ?

Well all I can say is that we all have different needs.

doug7489, correct. If you will mix this material you need 24 bit as each mix adds noise - reduce bit depth. 16 is no good as 3-4 mixes and you're down to 12 bits and yes you xan easly hear this.

I don't agree with you at all. I have a good ear, and I do hear the difference between the common CD-sounds and the better ones, for example the music on SACD, or on 24 bit /192 KHz. It's very easy to check out for everybody who has an SACD-player and some hybrid SACDs. The difference is very much audible and enormous. Even the remaster functions of the CD-players can't improve sufficiently the sound of the CD to the level of an SACD. This difference is audible more on the classical music CDs, or on acustic jazz ones. but the sound of the rock recordings are also better on SACDs.

Of course there is difference i hear it as well but it is not down container it is held in but mix itself.

Sebastian Gawlik When many years ago I bought my first wonderful Vivaldi SACD the seller (contrary that he was interested in selling me that) told me the same. But thanks to God, I didn't believe in him. So let me tell you, that you may know: a musical sound is not only about the frequencies, but harmonics, and many all that I don't even know about. I trust in myself, and my ears, my musical taste, so I do hear that the sound of CD is a muddy weak thing compared to the one of the SACD. And I also read about the difference between SACD and 24/192 PCM sounds, but I couldn't choose between them. Otherwise if I listen to test samples on my earphones I can even hear slight differences between 2.8 and 5.6 Mhz DSD sounds, or the 24/96 and 24/192 sounds. And let me try to convince you, I am not snob, and also not a HIFI-mad, (who listens to only "vinyl" or "analogue" tapes). I like and appreciate digital musical technology as did Herbert von Karayan and Frank Zappa.

schreiber9, do some double-blind tests, then get back to us about what you can 'hear'. I thought the same as you, and I was a classic high-end tweaker, until I actually sat down and did some proper DBTs, back in the early '90s.

victor4503 awesome

Reply to this post

haha golden rule from the article "Misinformation and superstition only serve charlatans". ?

Reply to this post

It's nice to have 24 bit when recording/mixing to give that extra headroom, but you're right 44.1/16 (or 48/16 for movies) are all that is needed.

yeah, he discusses that. For recording, 24 bits allows the recording engineer to make a rough guess as to needed headroom without running into problems later.

I still need to watch the video, I was just commenting on the headline?

It's kind of like shooting photos in raw before editing and converting to another format (like jpg)

I just realized that you are not my friend Dru Wright, who is another audio guy.?

mike6 no problemo

Finally read it - good article. I didn't know about that site and will bookmark it for further reading.

A book that I highly recommend for understanding this subject at a technical layman's/recording engineer's level is "Mastering Audio" by Bob Katz, which I think agrees with this article completely.

I got it, used. It seems like a textbook?

Honestly, I only find 24bit to be relevant when recording something into my audio interface. When mixing my DAW uses 32bit float internally, so I don't need to worry about it at that stage. I've never been in a situation where 16bits wasn't enough for a mixdown.

Reply to this post

Nice article, Dirk. Thanks! However, the devil is in the details... The theory is beyond question, but it makes all the difference in the world how exactly the stuff is built. It takes just one crappy component in the signal path to compromise the results, and conclusions...?

Reply to this post

It is necessary, but apperently mostly skipped, to understand the differences between recording and distribution formats.

Reply to this post

There is another weakness of the article. It had been written in 2004, and his main worry was the "errors" that occur during the bigger sampling rate. Now it's absolutely not a problem.

Absolutley correct.

Reply to this post

Guys, before you superimpose your personal experience over the theory in the article, do you honestly believe that all 16/44 converters are equal? How about opamps? Are they all the same? The list goes on and on.. Try to achieve a noise floor of -140dBu using carbon film resistors... Good luck! The theory and arguments put forward in the article are 100% valid.

yes, and in order to get 16 bit resolution you usually need a converter that has higher bits. For example, from what I understand, most 24 bit converters are only 18 bit resolution. So, you'd need something more than a 16 bit converter to achieve what the article is promoting. And, of course, you're correct, the analog section has to be very transparent in order to maintain the resolution.

Using a NOS 16/44 converter would be silly today, given the wonderful technology available and all the goodies that come with it, I am personally a big fan of asynchronous resampling, and in my experience 24/96 would be the practical limit of distribution media, exceeding by a wide margin what we can hear or what the actual electronics are capable of. If you get true 22 bits without supercooling or a Faraday cage for each and every stage, congratulations! Yes, it can be done, no doubt, but consider marginal returns..

Reply to this post

I had a chance to compare many formats in a controlled environment. All from a live source in a studio. Both for PCM and DSD the than top ADDA converters were used. There is no way in hell I would sattle for 44.1 16 bit. I have a large collection of material that is mastered from analog on different formats as well (SACD, CD, high res PCM LP) as well. You need 24 88. Higher sample rates are not much use but going from 16 bit to 24 huge. Going from 44.1 to 88 is big and DSD well thats often hard to distinguish from the soruce. So while I totally agree that 24/192 is silly 24/88 or DSD is not. But it all depends on the source (mixed) material and gear that was used to record with. Beck Sea Change is a recording I can recomend to anyone who wants to hear it themselves. It is availabel from the same master in many formats. There are also a remasters from teh analog masters for Pink Floyd etc where the lesser quality CD release came from the same master...

For waht is worth, REM Automatic for the people on DVDA has 192 and I had a lady frined who could tell 88 from 192. Me, never.

So, you're saying that you have a large collection of material mastered from analogue, which has never had more than 85 dB dynamic range, but for some reason you think that more than 16 bits is necessary to digitise it, when that has a dynamic range of about 93dB with proper dithering? Care to give a rationale for this bizarre claim?

Changes in SPL less than 0.5dB are perceived as changes in sound quality, not as changes in loudness. The quotes from test subjects, both trained and untrained listeners are: better clarity in the high register, more prominent and detailed midrange, better, punchier bass and increased soundstage/better imaging.

If everything got better, most likely it was sound level that changed. 99% of audiomyths are caused by this. Level matching matters.

stewart4831 If you read my post it says, All from a LIVE source in a studio. Both for PCM and DSD the than top ADDA converters were used. Then I go into analog sourced readily available media...

andor4 don't be disingenuous. After that first section, you went on to claim that 'you need 24/88' for anything mastered from analog, and that 'going from 16 bit to 24 huge'. I call bullshit on that claim.

andor4 So, it was level matched, instant switching blind test? Nothing else will do i'm afraid.

Reply to this post

To test converters: digitize a noisly lp, listen to the noise, the better the converter the less obvious it is, just as we evaluated systems back in the 1970'ties, the noisy lp test is still valid to determine if transient misbehaviour occurs.

interesting idea.

Reply to this post

I can hear the difference. I like good recoded cd even

Reply to this post

Reply to this thread

This site uses cookies and other tracking technologies to differentiate between individual computers, personalized service settings, analytical and statistical purposes, and customization of content and ad serving. This site may also contain third-party cookies. If you continue to use the site, we assume it matches the current settings, but you can change them at any time. More info here: Privacy and Cookie Policy