[RRod]

http://dr.loudness-war.info/album/list?artist=Norah+Jones&album=The+Fall (I provided the third measurement down, btw).

Heh, "Limited edition". This industry, man.

 

[TheSonicTruth]

Have you ever heard The Rolling Stones' Beggar's Banquet? There are songs on there that are massively compressed and clipped deliberately.

Kiss? Frampton Comes Alive?

wow

Well in the case of 'Banquet' it doesn't need additional messing with for a so-called remaster. Why the heck add compression on top of what's already baked in to the stereo master tapes?? Just rip to 24/192, check for channel balance, make minor EQ *as needed!*, dither down to 16/44.1 and reissue!

And I don't want to hear about how 'originals won't translate properly on modern playback devices' - that's a load'a hooey. Those things have volume controls you know.

 

[bigshot]

They didn't remaster Beggar's Banquet. They remixed it and took all the life out of the deliberate compression and distortion choices on the original album. They ruined it by making it sound too clean.

 

[TheSonicTruth]

They didn't remaster Beggar's Banquet. They remixed it and took all the life out of the deliberate compression and distortion choices on the original album. They ruined it by making it sound too clean.

In plain terms, they FU___D IT UP. Whether by 'remastering' or remixing, that's all they did. And I couldn't imagine any of the Stones signing off on it. I'm liking the label or some bone-head producer for that disaster.

That said, people of your stripe always try to justify the post 2000 sausage-fest by pointing out the 'sins of the past' in mixing and mastering terms. That don't cut the muster with people of my stripe.

 

[amirm]

1. It doesn't and that's my point! There is no precise definition of dynamic range and therefore there cannot be a precise measurement of it.

Then why on earth did you say you support the statement Bigshot made that there is no more than 50 db of signal to noise ratio in music? You should have said the above to him as a minimum.

Instead you went on to say in your experience that is true. What experience I asked. DId you ever do the simple test of converting your music to 50 db of effective dynamic range like I did? Have you done so now that I showed how audible the noise level is in that regard?

You hadn't done any of that, right? If a subjectivists says they can hear the sound of coasters under their equipment, we immediately demand controlled listening tests and scientific references to back the same. Yet when it comes to our claims, we want a free pass to take things at face value. "Oh you don't even need more than 10 bits." When I ask for data to back that, the answer is "that is my experience." Heck that is what a subjectivist would say about his experience!

And no, the math you did in the OP doesn't account for that seeing how I showed direct, peer reviewed research that completely opposes it. You can't substitute your assumptions about audio science as the real deal. You need to confirm it.

So once again, neither Bigshot or you are remotely correct that there is only 50 db, 10-bit of information, etc. in our music. It is a repeated myth by forum objectivists that has no data whatsoever to back it. It is false and misleading and we shouldn't keep repeating it.

Determining the effect bit depth of music in the face of noise requires statistical analysis. As it turns out, such work was just done and reported in this write-up against MQA: https://www.xivero.com/downloads/MQA-Technical_Analysis-Hypotheses-Paper.pdf

In there you see them analyzing 24-bit audio and finding that in some cases 23 bits are valid music content!

"Most uncooled physical systems are able to reach a thermal limited noise floor at around -120dB below full scale. Nevertheless, a theoretical 24Bit system is able to push the quantization noise down to -144dBFS (without dithering) and therefore allows for 24dB = 4Bits of Headroom to place the information of the upper frequency bands. BUT! We did statistical investigations into existing high resolution recordings and we can confirm that sometimes even the 2nd LSB [Least Significant Bit] already holds information, most likely due to applied dithering processes..."​

That translates in 138 dB of signal to noise ratio in the recording! We could de-rate that by a good bit and still be nowhere near 16 bits being the limit let alone 8 bits that was advocated.

So I ask again, what have you done to confirm your assumptions that 50 db is enough? What listening tests? What references you can quote? Because if there is none, then the thesis of this thread is disproven and we are done.

 

[reginalb]

Heh, "Limited edition". This industry, man.

Now, are you referring to the Limited edition 2CD set, or the limited edition SACD box set, or the limited edition vinyl box set? You have to be specific because there are a lot of limited editions of that album. You might also mean the imported version on Amazon. Not sure where it's imported from, but it's imported. Then there is the most rare of all. The regular version.

 

[RRod]

@amirm I'm missing the part where human beings have to be aware of the informational content. That statistical analysis can show a non-uniform noise spectrum is unsurprising.

 

[amirm]

1. I'm not sure if my system will output 120dBSPL at my listening position, I've never tried and would never want to!

Again, how did you determine that your system isn't capable of that and that you have never listened to that level? No, you can't use your dumb SPL meter for any of that analysis. What is important here is instantaneous levels not slow average.

At LA audio show, someone asked Andrew Jones what the SPL levels were that his ELAC speakers were producing in that setting. He first asked the listeners to give numbers. People were like you, throwing small numbers like 80 and 90 db around. His answer was that the peaks were hitting in the neighborhood of 115 db! The music was dynamic and maybe "loud" by some standards but not at all what you are assuming.

In these discussions people take these SPL numbers as if we are sitting there listening to continuous tone at 120 SPL. We are not remotely doing that. We are talking momentary peaks that may last just a few milliseconds.

And no, you don't remotely damage your hearing because two things are needed for that: loudness and exposure time. Here is the recommendation from US workplace safety standard, OSHA: https://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=STANDARDS&p_id=9736

On the left is the SPL number, on the right are the duration in *hours*. Converting 0.125 for 120 db in minutes we get 7.5. In other words, you need to listen for 7.5 minutes to the same constant noise, not a few milliseconds as we have in music, to hurt your ears.

So please don't keep talking like these are unheard of numbers. Can't be done. We will go deaf, etc., etc. These are forum objectivists talking points we need to leave behind.

Yes, we are talking about reference level playback. If you listen to very modest level, your needs will be different. But again, in the context of what is audible we need to include the full population and their usage of technology.

 

[amirm]

@amirm I'm missing the part where human beings have to be aware of the informational content. That statistical analysis can show a non-uniform noise spectrum is unsurprising.

We are discussing what is in the content, not what is audible to humans there. The PR tactic is that if we can show music can never have more than 16 bits, then why talk about > 16 bits? This data disproves that assertion.

 

[amirm]

2. Really, all your hi-res content cost the same as the 16/44 versions? I don't think that matches the experience of most here.

No, that was in regards to *hardware* that plays it. We have already paid for the capability to play high-res content in the DACs and systems we own. So you are not saving anyone money by advocating that they should not need > CD capabilities. They have the hardware and if they choose to play high-res content, that is a decision they can make on an instant by instant basis.

As I mentioned earlier, my hope is that one day the recording industry gives us access to music prior to final mastering for CD. That content is already in high-resolution format and should it not be subject to evils of loudness compression, would be very well worth the extra premium charged for it.

 

[amirm]

No, it's not! Typically today and for a number of years it's created at 64bit float and before that it was typically created in 32bit float or 56bit fixed. I don't know of a time music was created in 24 bits but it must have been more than about 20 years ago.

No, that is the internal processing format for your DAW. IThe music was captured at 24-bits and the only output format that is playable by all of us is integer PCM at up to 24 bits or "32 bits float." That is the format that is handed to the mastering engineer who then processes it with EQ, limiting, etc. and then proceeds to also convert the sample rate down to 44.1 Khz and bit depth to 16 bits. I want this last step completely eliminated for the music I consume. I have no need for that "mastering" to formats I don't care about (CD, compressed AAC, MP3, etc.).

 

[amirm]

I notice you ignored every single one of the 4 questions I asked, which seems to be a trend in your responses!

I hope you are satisfied now. I don't feel the obligation to always answer your multi-part questions. My interest in engaging only go so far and I don't want to bore the membership with such detailed back and forths. So please don't read much into me not answering everything you say.

 

[TheSonicTruth]

No, that is the internal processing format for your DAW. IThe music was captured at 24-bits and the only output format that is playable by all of us is integer PCM at up to 24 bits or "32 bits float." That is the format that is handed to the mastering engineer who then processes it with EQ, limiting, etc. and then proceeds to also convert the sample rate down to 44.1 Khz and bit depth to 16 bits. I want this last step completely eliminated for the music I consume. I have no need for that "mastering" to formats I don't care about (CD, compressed AAC, MP3, etc.).

Why don't you care for CD?

 

[RRod]

We are discussing what is in the content, not what is audible to humans there. The PR tactic is that if we can show music can never have more than 16 bits, then why talk about > 16 bits? This data disproves that assertion.

You've got that backward. The PR tactic is to show information below 16 bits and then to say you need more than 16 bits in your normal living room.

 

[Cutestudio]

The problem is that it's impossible to quantify what is over-compression, a perfectly acceptable amount of compression for one song might be ridiculous over-compression in another.

Impossible? In reality it's one of the most trivial things to quantify. A simple histogram shows the distributions and reliably picks out the over compressed/clipped tracks, additionally an entire track with the same consistent level peaks regardless of RMS level or frequency (the classic 'brick' shape) is a classic tell. Clearly you have never studied the subject.

In addition to that one can easily see the flat tops, no 'perfectly acceptable' amount of compression squashes down a 6dB peak into a small range of around -40dB or less. If you'd actually bothered looking at some waveforms you'd not be making these daft statements. Indeed your statements lead me to conclude that you may have heard the phrase 'loudness war' but have no real idea or grasp of the problem; just an observation, because like many you simply have decided your opinions trump any research. I have nothing against ignorance, but it's not working for you here.

And incidentally, Mastered for iTunes does NOT stipulate appropriate amounts of compression or directly affects/combats the loudness war.

No specific stipulation no, but it's known for not suffering from clipping/over-compression. Because I've actually researched the Mastered by iTunes masters I knew this which is why I mentioned it as a good example of Apple 'doing something' about the shoddy mastering product record companies pass off, your casual implication that they are poorly mangled is false, sorry.

The number of egos on this forum supporting the total lack of QA by the record companies is sad to see in a HiFi forum and is perhaps the reason why the industry is in such decline. Bigshot even claimed that record companies aim for earbuds and car systems as an excuse for their dreadful output, despite that fact that Pink Floyd's 'The Wall' CD tracks play perfectly both in the car and on any MP3 player I've tried, whereas the modern loud continuous drone pumped out today sounds awful on any playback system.

Other industries have reduced product quality until people stopped buying it, I don't see the record industry bucking the trend. Justifying the destruction of a product may be justifiable to some but for the rest of us it's baffling as it's fraudulent for the buyer, hurts sales and re-sales and would be as easy to avoid as not pressing a few buttons. No one who likes music of sound can support this wilful vandalism.