Some commonsense

[jabbr]

I think the mistake in the analysis here is that no one is directly listening to a “file”, rather the output of the electronics that translate the file into audio. 

 

With the exception of the rare 16/44 NOS DACs that perform no upsampling, the vast vast vast majority of DACs that accept 16/44 input perform some type of upsampling and/or SDM ... the benefit of higher bit rate input is that the DAC need not do extraneous processing best done by a CPU. Why not just use the resolution that the music was recorded at? Avoid extra filtering steps. 

[jabbr]

14 hours ago, Ajax said:

The point of the article is that mathematically 16/44.1 is adequate for music playback

 

That point has been made for decades. Nonetheless modern DACs upscale/convert 16/44 before the actual digital analogue conversion. 16/44 is effectively lossy compression. Arguably the lossy compression is not audible but nonetheless if a recording is made at 24/192 or DSD256 why in heavens would I want to compress it?

 

Just give me the closest to native format thank you. That’s common sense to me.

[jabbr]

1 hour ago, pkane2001 said:

 

Not arguing for unnecessary compression, but 24/192 is lossy, just like DSD256 is lossy, just like 16/44 is lossy. The question in all cases is precisely whether what's being lost is audible or not. Argument can be made for 24/96kHz being perfectly sufficient for human consumption. Dogs may need 24/192, bats may need 24/384 

 

24/192 is not lossy when it’s the recorded or mastered resolution. The only reason to down convert to 16/44 is for compression. Similarly when DSD256 is the recorded format (ok DXD is easier for processing)

 

Similarly I store my photos RAW even though jpeg is fine for a drivers license.

[jabbr]

Just now, pkane2001 said:

 

24/192 is lossy if you consider that you lose all the frequencies above 96kHz and probably a bit below due to filtering

 

Generally when we use the term “lossy compression” regarding digital data, we mean that bits are altered when round tripping between the compressed and uncompressed format. If the digital data starts out at 24/192 then it’s not compressed by that definition. Similarly for a photo which is also always bandwidth limited by the digital sensor (and lens). Lossy compression has a specific meaning: the difference between FLAC and JPEG. 

 

Of course one could always make a higher resolution recording just as one could use a higher resolution imaging sensor — but the term “lossy compression” means that bits are thrown out with respect to the digital recording at whatever resolution it was made.

[jabbr]

2 hours ago, tmtomh said:

So IMHO it doesn't make sense to call any digital sample rate "lossy" if it samples at more than 2X the upper limit of human hearing. For that reason, I would argue that it is plausible, but ultimately pointless, to call redbook/CD's 44.kHz sample rate lossy. 

 

You are missing the point. Whether a compression scheme is lossy or not has nothing to do with the sample rate per se. If a recording were made in 16/44 then it wouldn’t be lossy. Yet when a recording is made in 24/192 and mastered in 24/192 and then down converted to 16/44 or MPEG then its lossy. Simple. I’m not calling redbook/CD sample rate lossy, rather  CDs are themselves often lossy because the recordings often are made at a higher rate/depth.

[jabbr]

2 hours ago, tmtomh said:

The only respect in which 44.1kHz is lossy is that if it is generated by downsampling from a 96kHz (or 192kHz) source, the downsampling is non-integer (44.1 doesn't divide evenly into 96 or 192), and there is some alteration of the source there. But if you downsample, say, a 96k source to 48kHz, or a 176.4k source to 44.1kHz, I would argue that such downsampling is not lossy: because the sample rates divide evenly into each other, you are losing nothing - all audible-range frequencies are preserved.

 

This is the same logic that JPEG uses to provide lossy compression yet preserving “relevant” image detail. In any case lossy compression is lossy compression no matter how good it is — we can equally argue that MPEG 320 is audibly indistinguishable to a majority of humans from redbook CD.

[jabbr]

3 minutes ago, tmtomh said:

f perfect reconstruction (minus quantization error, of course) is the definition of lossless, then a 44.1k CD source made from a 176.4k original will be lossless in the audible range.

 

The real point however is the mechanism by which the 24/192 data stream is converted to 16/44 ... using a sample rate converter! Are these all the same? Are they all of the same quality?

 

Hell no!

 

Please don't suggest that all sample rate converters produce the exact same result....

 

1 hour ago, mansr said:

PCM is lossless in the sense that it captures perfectly any signal within the well-defined bounds determined by sample rate and bit depth. This is in stark contrast to perceptual coding systems that discard a little bit here and a little bit there according to their psychoacoustic models of what is or isn't audible.

 

See above. PCM *can* be lossless when use to hold the original data. As you very well know, it's the process that converts one PCM format to another that is lossy. Perhaps by y'all definition a perfect sample rate converter would produce lossless CD but I still prefer to have the source audio in as close to the mastering format as I am able to obtain, thank you.

[jabbr]

39 minutes ago, tmtomh said:

In this regard I'd remind you that the subject of this thread is about common sense, and your argument here is not a common-sense argument - as @mansr so aptly put it just above, "Insisting that anything with less than infinite bandwidth and precision should be called lossy only serves to muddle the distinction between on the one hand plain sampling, the accuracy of which is known upfront, and on the other hand perceptual coding where the accuracy varies wildly depending on the signal content."

 

Here is a common sense position: I cried a little inside when I read the nytimes article that states that the original Coltrane Impulse masters were destroyed in the Universal fire.

 

 

You apparently couldn't care less because you consider the CDs definitive. Compared to the masters, I consider the CDs (that I have) lossy. The fire was a huge loss.

[jabbr]

20 hours ago, mansr said:

PCM is lossless in the sense that it captures perfectly any signal within the well-defined bounds determined by sample rate and bit depth. This is in stark contrast to perceptual coding systems that discard a little bit here and a little bit there according to their psychoacoustic models of what is or isn't audible.

 

Again, its not PCM which is itself lossy, rather the process used to convert one PCM format (bitdepth-samplerate) into another. A perfect infinite length stream could be Fourier transformed and a perfect brickwall filter applied but no stream is infinite and no brickwall filter is perfect.

 

So ... the decision to use 16/44.1kHz is based on a "psychoacoustic" model of human hearing that claims that since the cochlea does not respond to tones above ~20 kHz, than bandlimiting the signal to 22 kHz will capture everything that is heard. This is an assumption folks. No need to argue whether it is a correct assumption but nonetheless it is based on a model of human hearing. 

 

Arguably the cochlea and human hearing system is better modelled with wavelets and so wavelet compression has been implemented. Is that "psychoacoustic" because different math is used (e.g. wavelet vs fourier?)

 

The point is that all of these formats whether 16/44 PCM or MPEG or AAC are based on human hearing assumptions (admittedly some assumptions e.g. 20 kHz upper limit, are more substantiated than others) and all of these formats are based on assumptions about what is necessary to represent audio.

 

My definition of losseless conversion between format A and format B is very precise:

There must exist a pair of transforms such that given a file f, when transformed and then inverse transformed, results in a bit identical result.   e.g.

f = T-1(T(f)) where

fB = T(fA) and

fA = T-1(fB)

 

So of course each format is lossy in a different way, and some are too lossy, and some losses aren't audible. Sure. Nonetheless downconverting a master to redbook/CD is typically lossy except where the source data was 16/44.1.

[jabbr]

20 hours ago, mansr said:

I really don't understand your insistence over this. The term "lossless" has a widely accepted meaning. You are attempting to redefine it. Why?

 

True or false:

a) Downconversion from 24/192 to 16/44.1 is frequently lossy?

vs

b) Downconversion from 24/192 to 16/44.1 is always lossless?

 

My position, according to the common definition of “lossy” is that (a) is true and (b) is false. 

[jabbr]

11 hours ago, mansr said:

Why are you even asking that question? Sample rate conversion isn't considered a compression algorithm (unless your name is Bob Stuart).

 

I am asking the question because when you assert that downconversion from 24/192 to 16/44.1 is lossless then you are asserting that the redbook/CD distribution is equivalent to the digital Master. 

 

Likewise with your example, you are asserting that one can recover a 24/96 digital master by upconverting a 16/44.1 CD.

[jabbr]

12 hours ago, mansr said:

Why are you even asking that question? 

 

To expand: this issue is important to me because I firmly believe in the preservation of the recorded track in all its glory regardless of whether ultrasonics are audible. I would preserve the original raw delta-sigma ADC output tracks as well as the (hopefully) DXD master. Regardless of whether it is audible. The redbook/CD downconversion is not the same to me, regardless of whether the difference is always audible. I consider that downconversion to be a loss, hence my use of the term "lossy", regardless of whether the loss is always audible.

 

Why? The issue of ultrasonics is not settled in my mind. No doubt there are fantastic CDs. But preservation of every bit of the recording preserves the possibilities of future as yet unknown technologies. That's common sense to me.

[jabbr]

1 hour ago, mansr said:

What do you mean by original signal? Typical recordings are multi-tracked at 44.1/48 kHz with a substantial amount of processing and mixing before a CD release eventually pops out. Comparing any individual microphone feed to the final CD wouldn't make any sense. It's not supposed to be the same.

 

The statement above belies a fundamental difference between your and my desired music and philosophy. No doubt we will never agree. I’m fact it is exactly my desired audio experience to listen to vocals sounding as close to the mic feed as possible. For me it is supposed to be the same (or at least give me as close to that illusion as is possible)

 

 

[jabbr]

3 hours ago, mansr said:

It's not about what I or you want. Jud specifically said he was interested in "typical" recordings. In a typical music release, the final master is not intended to sound the same as the microphone feed. That's just how it is, like it or not.

 

Like it or not your “typical” recording of Lady Gaga is not the same as my “typical” recording and certainly not the same as a typical audiophile recording by any of the discussed audiophile labels. For mainstream releases I do like Analogue Productions as another example. 

 

But if you consider “typical” as average mainstream then the intention is as you say — indeed intended to sound best on earbuds via Spotify, or a car stereo or through a TV screen ... and no doubt highly processed as you say. Again my position is that if I was recorded and mastered at a certain nitrate, then that’s the rate I want my delivery.

 

Again your typical is vastly different than mine so there we are.

[jabbr]

31 minutes ago, Jud said:

 

Mmm, nitrates - hot dogs and bacon!

🤣 didn’t even notice that meaning ... I was thinking of the nit-rate as the frequency that @mansr would nitpick a post. 

[jabbr]

43 minutes ago, mansr said:

I have never listened to a Lady Gaga track. I understood Jud's "typical" to mean representative of music production in general, not whatever any one individual chooses to listen to.

 

Ah, so you mean typical as in mainstream mastering engineers like Barry Diament?

[jabbr]

6 minutes ago, Ralf11 said:

 

@jabbr - while I appreciate your caution in preservation of everything, does your desired goal exclude all processing, even processing that is clearly euphonic?

 

e.g. Dave's Picks (and some of the later Dick's Picks) uses processing to give a more euphonic sound

 

 

I defer to the judgement of the mastering engineer, regarding processing and resolution. As I said: I want a copy of the Master when possible. 

 

[jabbr]

1 hour ago, mansr said:

I mean typical as in the bloody dictionary definition of the word.

The definition in the bloody dictionary was smudged.

 

In any case since we are unable to agree on the definition of mathematically precise terms such as lossy and lossless — which themselves are also defined in dictionaries — then your own interpretation of the dictionary definition of typical (here I’ve defined the group as audiophile music given our presence on an audiophile site) will likely be at odds with the English language definitions as I use them. 

[jabbr]

41 minutes ago, Ralf11 said:

 

I think we need to clearly distinguish recording from delivery.

 

One of my main desires in delivered music, is to allow me to close my eyes and feel like I’m at a live performance. To promote the hallucination of “being there”.

 

In other situations the delivered music is the performance.

 

One of the most striking experiences I had at the Radiohead concert was seeing the complex electronic effects performed live in real-time. I had assumed these were multitrack studio creations. 

[jabbr]

34 minutes ago, Ralf11 said:

re " to close my eyes and feel like I’m at a live performance. To promote the hallucination of “being there”."

 

do this mean you are going Kal-ish?  multi-channel...

 

I often use headphones but there is great merit to multichannel & room correction. 

Quote

 

my (theoretical or conceptual) problem with relying on the mastering engineer is that the mastering may be be directed at a delivery system that differs from my delivery system (e.g. the egregious spectre of mastered for iPhooey, or mp3)

Likewise in photography we have raw capture. Input profiles/convolutions and output profiles/convolutions ie for each device. 

 

One goal to preserve the 24 bit headroom is that we can do more in room processing and/or even non linear ops — for example as you age and Hf hearing drops off perhaps a nonlinear op might give you the same perceptual hearing — or even transfer “air” at 24 kHz into “mist” at 18 kHz etc 😉

 

In any case I don’t need anyone telling me that I never want anything higher resolution than 16/44.1 because you never know what may come 😉

 

Quote

 

then there is the ability to lift veils in the future as in the Dick's Pxxx example above

 

I agree that the masters should be preserved.

 

😉

[jabbr]

I'd like to expand on the idea of overspec'ing

 

@esldude -- yes by all means get the best microphones possible and the microphone should have a range well beyond the highest frequency you need to capture,

 

similarly your interface: it should capture well outside the microphone range (as @Miska says),

 

similarly your power supply: it should handle well out side your maximum draw, same for your amplifier, it should handle way more than the maximum SPL you listen to. same for CPU etc etc etc

 

same for measuring equipment, we typically try to use equipment with 10x the resolution of the signal we are measuring

 

The point is that most every system starts degrading at the limit of its stated performance so you get the best results if the system can handle more than you expect to throw at it.

 

Overengineering is common sense.

 

Of course if you don't have $$$ then you cut corners and live with it.

[jabbr]

1 hour ago, marce said:

I agree with one caveat, capturing well outside the microphone range could add some noise outside the required frequency range  which may or may not cause problems with following circuitry. A fine balance is required. to much bandwidth can be as bad as to little sometimes.😀

 

This is not a trivial topic. My impression after just a little bit of investigation is the contrary — a well implemented, for example, DSD256 capture will have significantly less higher ultrasonic noise than a DSD64 or PCM44 signal which is upsampled to DSD256 (because of noise shaping). The caveat is that various filters might be applied, so hard to compare apples to apples but perhaps @Miska who I believe has one of  those sweet RME  ADI-2 ADCs can chime in with some real data. Since we know that most all DACs internally upconvert incoming PCM44 to hi-bitrate SDM, then I assume all DACs need to properly deal with ultrasonic noise (hence the output filter of course)

[jabbr]

9 minutes ago, PeterSt said:

Ah ... I didn't read the thread. So this was in the context of the sampling rate being 44.1Khz, right ?

 

The thread was putatively about common sense, so all bats are off...

[jabbr]

With all due respect to those who insist that 16/44.1 captures all that we could possibly hear, none have you have presented to me an audio reproduction system which sounds entirely realistic, to me, and in my eternal hope that future audio reproduction systems will improve on the current state of affairs, common sense tells me to preserve every bit of a recording --  holding out the real probability that we will need new types of recording, yet nonetheless.

 

Said more technically: y'all are entirely forgetting nonlinear mechanisms.

[jabbr]

1 hour ago, esldude said:

Hint: entirely realistic audio is not going to result from going to higher sample rate.

 

I’m not saying that higher sample rates lead to more realistic audio by itself.

 

Hint: you don’t know what does, nor does anyone else have a definitive answer. 

 

Im saying that I want the closest thing to a live feed from the mic, just in case.