Testing audiophile claims and myths

[KeithEmo]

I'm jumping in late here... but it seems to me that the ENTIRE point has been missed.

If you want to make an analogy between images and sound then the color range of a JPG image is NOT a good choice. The main issue with lossy compression is NOT the EQ or frequency range of the music - although some compression ALSO affects that. The JPG format has plenty of color and contrast range to store any image that can be printed. The issue is that, when you use JPG compression, which is lossy, the resulting loss of information results in various digital artifacts; and, because these artifacts are the result of information that was present in the original image being discarded, you cannot "get back the quality of the original image" by readjusting it. And, likewise, since they are somewhat "non-deterministic", there is no way to precisely reverse them and get back the exact original image.

If you take a very sharp scan of text or line art, or a vector original of line art, and convert it to JPG, you will get "little echoes and ghosts" around some of the sharp lines, especially vertical and horizontal lines. You will also find that certain gradations between colors, and transitions between colors, have been altered or lost. And you may even find entirely added artifacts - including the well-known "JPG blockies" often seen in low quality JPG images. This can be minimized if you use the maximum quality setting available but cannot be entirely avoided. This is a pretty good analogy for the detail information that is discarded when a lossless file is compressed using lossy compression. And, just as there is no "dial or adjustment" that can turn a JPG image with lots of digital artifacts back into a sharp clean vector image, there is no adjustment that can "put back" the information that has been discarded during lossy compression.

(It's worth noting that the process of printing an image to newsprint is itself lossy. If you look with a magnifying glass you will find that the sharp lines present in the original line art have been approximated using a series of small dots of various colors and sizes. Therefore the concise sharpness of the original lines has been permanently lost. Then, when you scan that image, you have two choices. You can either blur the result, making it even less sharp, but hiding both the dots in the image, and any artifacts created by your scanner. Or you can make a scan sharp enough that you can actually see the size, color, and shape, of each individual print-litho dot. If you choose to blur the image then it will end up being less sharp... but will "look OK". If you choose to use maximal sharpness, and your scanner is good enough, you will get an exact copy of the printed image, but it will take up a very large amount of storage space... and will still be limited by the sharpness of the lithographic dot process used to print the original paper print.

In neither case will you be able to get back the sharpness of the original pen-and-ink drawing or digital vector image. And, in either case, if you attempt to produce a highly detailed print copy of the original artwork, you will find it virtually impossible to produce a copy that looks as good as, and exactly like, the original.

 

[Powermankw]

It just seems it can get inevitable, as current digital formats have "hi-res" for audio (really less to do for reproduction as audio post processing) vs "UHD" and "HDR" with visual formats (which are still relevant for reproduction, and more so post processing).

Not really. It's a 4 page argument about photography. The pissing match part in the sound science forum is the inevitable part. But if the point can't be made, or refuses to be acknowledged, after 4 pages, maybe it's time to move on. Just seems like we are missing out on more interesting discussions on cables.

 

[bigshot]

Tag, you're it.

 

[gregorio]

It just seems it can get inevitable, as current digital formats have "hi-res" for audio (really less to do for reproduction as audio post processing) vs "UHD" and "HDR" with visual formats (which are still relevant for reproduction, and more so post processing).

No, “hi-res” for audio doesn’t have anything to do with audio post processing either! There are some very specific professional applications for “hi-res” audio but not with music. The analogy is false.

G

 

[Davesrose]

No, “hi-res” for audio doesn’t have anything to do with audio post processing either! There are some very specific professional applications for “hi-res” audio but not with music. The analogy is false.

G

So source audio files are always 44.1k? I thought with them being 24/32-bit (which I know also deals with potential of dynamic range), there was a choice of 48k/96k/192k. The analogy was that you use uncompressed source files that will have higher DR and resolution than what the intended output is. The same way you don't use 320kb mp3s for audio mixing, video editors are using RAW formats for the dynamic range and potentially better resolution (IE cinema resolutions being different than consumer UHD).

 

[gregorio]

So source audio files are always 44.1k?

No, I just said there are applications for high sample rates. The standard rate for audio is 48kHz, except for music which is 44.1k due to the requirements of the CD format. Source files for music are not uncommonly 88.2k or 96k, more rarely 176.4k or 192k and very rarely as high as 384k but this is due to marketing requirements.

We also don’t use 24bit or 32bit for dynamic range, 16bit is more than enough dynamic range. We’ve had this discussion before but you don’t seem to have understood it!

G

 

[Ryokan]

I downloaded a track that was 48khz and it sounded better than 44.1 was that because it wasn't volume matched?

 

[Davesrose]

No, I just said there are applications for high sample rates. The standard rate for audio is 48kHz, except for music which is 44.1k due to the requirements of the CD format. Source files for music are not uncommonly 88.2k or 96k, more rarely 176.4k or 192k and very rarely as high as 384k but this is due to marketing requirements.

We also don’t use 24bit or 32bit for dynamic range, 16bit is more than enough dynamic range. We’ve had this discussion before but you don’t seem to have understood it!

G

No, I think this is another example of miscommunication. Like the previous time in which we had to have a long argument about my use of "32bit workflow" instead of 32bit recording workflow. Are you going to at least agree that 320kb mp3 files aren't used with mixing? My friends who record instruments, are using a high res format when they're doing the editing. But it seems you continue with ad ad hominems instead of understanding the original argument: which was "hi res" isn't needed for audio reproduction (the way a 16bit RAW from a 30MP camera is going to have larger features than a 4K TV display).

 

[gregorio]

I downloaded a track that was 48khz and it sounded better than 44.1 was that because it wasn't volume matched?

Possibly/Probably but it could also be because it’s a different version/master.

No, I think this is another example of miscommunication.

Very probably, as you don’t always seem to understand the terms you’re using but it’s also an example of a false analogy.

Are you going to at least agree that 320kb mp3 files aren't used with mixing?

No, because there are many situations where we have to do exactly that. However, given the choice we wouldn’t use a lossy compressed format because there are rare conditions where it can cause an issue.

My friends who record instruments, are using a high res format when they're doing the editing.

There are many amateurs/hobbyists who use hi-res because they don’t know the facts and like many audiophiles, think it must be better. Your argument was “for audio post processing”, not recording or editing.

it seems you continue with ad ad hominems instead of understanding the original argument: which was "hi res" isn't needed for audio reproduction

What ad hominem? And, I did not argue that hi-res isn’t needed for audio reproduction, I challenged your assertion about audio post processing!

G

 

[Davesrose]

What ad hominem? And, I did not argue that hi-res isn’t needed for audio reproduction, I challenged your assertion about audio post processing!

G

Saying that my post was false, and then saying I don't understand. Well at least you yourself admitted that 96k is a common file for music editing (#16,671). I said “post processing” because that is a common term for the editing stage in a file, photo editing, or video production (the context of the post was why compare audio with photography). I'll leave it to you to continue to misconstrue the claim that these files can be useful for editing, and then attack me for not being able to understand.

 

[bigshot]

Here we go again.

 

[Davesrose]

In neither case will you be able to get back the sharpness of the original pen-and-ink drawing or digital vector image. And, in either case, if you attempt to produce a highly detailed print copy of the original artwork, you will find it virtually impossible to produce a copy that looks as good as, and exactly like, the original.

Yeah, sorry, my graduate degree was medical illustration. When doing pen and ink drawing, we would draw on velum and do high scans in EPS which was what was printed (which for back then, was very high DPI). Then learned another workflow of trying to have pen styles with Adobe Illustrator and drawing with vector (which in itself has unlimited resolution). But also these arguments are irrelevant now as we have JPEG formats not being that compressed, and even consumer formats now are going to HEIC as opposed to JPEG (which introduces better compression and 10bit color space).

 

[gregorio]

Saying that my post was false, and then saying I don't understand.

That’s a statement of fact, not an ad hominem attack!

Well at least you yourself admitted that 96k is a common file for music editing (#16,671).

No I did not! I said that 96k is not uncommon for recording music (which does not mean it’s common) but not specifically not for editing. Music maybe recorded and distributed at 96k but either due to marketing (and therefore it’s requested by the client) or by amateurs/hobbyists due to ignorance. It makes no difference to editing.

I said “post processing” because that is a common term for the editing stage in a file …

You actually said “audio post processing”, which is not a common term because “audio post” indicates sound for film/TV as opposed to music recording and the vast majority, if not all, of the applied “processing” in music recordings occurs during the mixing or mastering stages, not the editing stage.

So, clearly you do not understand the terms you are using or the operating principles behind them, nor how, when and where audio processing is applied and therefore what the implications of using higher sampling freqs actually are.

It is therefore unwise to try and create analogies with the video/photo world because this lack of understanding of the audio world will likely and often does result in false analogies.

I'll leave it to you to continue to misconstrue the claim that these files can be useful for editing …

What am I misconstruing, let alone continuing to misconstrue? Your claim is clear and you’ve even repeated it “that these files can be useful for editing” - Which is false (with music)!

The “misconstruing” here is all on your part, falsely assuming enough equivalency in the terminology and principles of use in photography to make claims/analogies with audio. You then argue endlessly that your analogy or claim is true and that I am misconstruing your (false) claim, so around in circles we go again, time after time!

G

 

[Davesrose]

Within common terms of "post processing":

Yeah, audio is not at all within the same range as other disciplines that's a common term for "post processing"

 

[gregorio]

Within common terms of "post prossing":

Where does Wiki say it’s a “common term”? You just made that up!

Sure one *may refer* to post processing in terms of audio but generally one wouldn’t/shouldn’t, because of the reasons given and that the use of a “processor” has a specific meaning in audio creation and are used extensively in mixing and mastering but rarely or never in music editing.

You have unwittingly but very effectively proven my point! Your claim is false but your response was entirely based on the semantics of what may be referred to, which incidentally demonstrates you do not understand how the terminology is actually used in practice. Instead of arguing semantics, why don’t you support your claim by giving an example of where/when a 96k sampling rate would be useful when editing music?

G