Article: Is It Time To Rethink Lossless?

[Luposian]

51 minutes ago, The Computer Audiophile said:

Looking further at this statement, it seems that you are totally fine if Spotify encodes its entire OGG Vorbis catalog into FLAC and calls it lossless. Is that what you said?

Lossless, as I have always understood it, is "the source and destination are identical" (no audio information was truncated/removed, to make the file smaller, regardless of compression used or not).  Therefore, what you hear in the resulting file should sound identical to the source.

 

My hearing is pretty bad (I hear almost nothing in my right ear), yet I will always choose a lossless format over a lossy format.  Why?  Because I want to know that everything I'm hearing is what I can hear, not what someone else thinks I won't hear/miss audibly.  I'm a purist that way, even if I can't appreciate it.

 

The FLAC file, in your example, is a lossless version of the .OGG file.  Nothing MORE was lost in translation.  But you can't say that it's a lossless file, because  audio information was removed in the source file.  So, even though you used a lossless format, that doesn't make it a lossless result.  The only way you can consider the resulting FLAC truly lossless is if the source recording was uncompressed and no audio information was removed.

 

CD (source) -> WAV -> FLAC (end result) = truly lossless

CD (source) -> .MP3 -> WAV -> FLAC (end result) = not lossless (even though two of those file formats ARE lossless)

 

KNOWING there is audio loss and HEARING audio loss are two separate things.  I may never HEAR a difference between an .MP3 of a CD track and a FLAC of an audio track, but KNOWING something was removed, to make the .MP3, will always bother me.

[Luposian]

1 hour ago, The Computer Audiophile said:

That’s a very simplistic approach and certainly one that makes thinking about it easy. 
 

However, in the real world we have to use real examples and compare the source to the end product. The term lossless was simple when we just ripped CDs. Now it’s used everywhere to describe something. 
 

Do you think the following should be described as lossless?

1. MQA in a FLAC container.

2. CD quality 16/44.1 streaming files in ALAC or FLAC where the master isn’t 16/44.1. 

3. A 24/96 WAV version of an album that has a 24/352.8 WAV master. 

 

 

If the source was unmodified (nothing was truncated/removed from the audio signal), then as long as no part of the end result modifies/removes audio information, I'd consider that lossless.  If MQA is a lossy file format, then you've removed audio information, thus a FLAC container is sheep's clothing... no different than making a FLAC of an .MP3... WHY?  Just keep the .MP3!

 

If you take a higher quality source and step it down to a lower quality (your 24/96 result conversion from a 24/352.8 master), you've truncated information, which would be a lossy translation.  The WAV format only preserves what it is given.  But the 24/96 result has lost some audio information, even though it's a WAV file.  You've MANUALLY killed off some audio signal, in that case.

 

If you "upsample" (translate lower quality to higher quality), you cannot consider that lossy, because you haven't removed audio information.  So, you could consider that lossless, because the definition of "lossless" is "no loss".  Adding audio information is the opposite of losing audio information.

 

The file format, alone, can not be the sole definition of "lossless".  "Lossless" must be carried out every step of the way.

[Luposian]

1 hour ago, The Computer Audiophile said:

I like your thinking on this. 
 

Too bad all the streaming services advertise music as lossless. It really doesn’t tell the consumer much. 

To that degree (in that situation), I just figure, if I like what I'm hearing, that's good enough for me.  I only consider "lossless" valid, if I can personally confirm that it truly is. 

[Luposian]

Of course, to get to the true grit of the term "lossless", we have to consider the original recording.  If I speak into a microphone and my voice is recorded at 16/44 WAV, then THAT is the source recording.  If that recording is then stepped down to a 16/22 WAV, you've REMOVED information, even though the WAV format is lossless.  So it's lossy.

 

24/150 orig. recording (source) -> .MP3 = loss of information (obviously)

24/150 orig. recording (source) -> WAV (assuming it is translated at the same 24/150 rate) = lossless

24/150 orig. recording (source) -> WAV (if it can't translate (or is opted to not translate) at the same rate (say 24/96)) = lossy

 

If ANY step of translation removes information from the original source recording, then you've lost information, and the result can no longer be considered lossless.  Seems rather straightforward to me.  "Lossless" is not JUST the file format.  It's how much of the ORIGINAL recording is preserved in the end result.

[Confused]

It looks to me that the definition of "lossless" is perfectly clear.

 

When we add the question "lossless to what", very little is clear. The problem being that we almost never know the specific details of the actual master.

 

When purchasing downloads or similar, I have the same philosophy to an earlier poster, I would like to purchase in whatever format and rate that was used for creation of the master. This is fine, but the required data relating to the master is almost never available.

 

Furthermore, I very much doubt that the above situation will change anytime soon.

 

Meanwhile, I also think that the actual quality of the master matters far more than whatever format the recording is actually available in for consumption by the masses. So for now, just enjoy whatever sounds good, no need to worry too much.

[R1200CL]

What is SACD originated from ?

Same master as the CD ?

Can SACD be called hi-res ?

[The Computer Audiophile]

1 hour ago, R1200CL said:

What is SACD originated from ?

Same master as the CD ?

Can SACD be called hi-res ?

Its a giant bag of unknowns :~)

 

It’s why I like to focus on the people who create the music and I like to read what other people say about the recordings after they listen to them. 

[maxijazz]

On 11/8/2023 at 1:24 PM, The Computer Audiophile said:

 

I’m not asking to be confrontational, just asking to see where you loosen the definition from its strict meaning. All with good intentions and interest in your answers. 

There are no loose definition in engineering and world of physics.

You are mixing 4 (maybe even more) different worlds into one, only because the same english word could (could it?) be used to describe them:

- data compression

- sample theorem

- data conversion algorithms

- marketing mumbling and related arguing among uneducated audiophiles  

The first 3 have strict, well defined meaning.

You might want to focus on marketing meaning, but please do not try or suggest the definition of the first 3 should be changed/extended.

 

[The Computer Audiophile]

7 minutes ago, maxijazz said:

There are no loose definition in engineering and world of physics.

You are mixing 4 (maybe even more) different worlds into one, only because the same english word could (could it?) be used to describe them:

- data compression

- sample theorem

- data conversion algorithms

- marketing mumbling and related arguing among uneducated audiophiles  

The first 3 have strict, well defined meaning.

You might want to focus on marketing meaning, but please do not try or suggest the definition of the first 3 should be changed/extended.

 

The black and white world of engineering is very simple, but never translates to the outside world well. For all five people who still rip CDs or convert one format to another, the strict definition of lossless is critical. 
 

In the larger world of purchased downloads and streaming music, the term lossless is used daily without any meaning. This is the area on which I mainly focused in the article.

 

I know engineers think the world will end and the war on truth will go nuclear when people talk about rethinking concepts such as lossless, but I have no intent to wage such war. 

[GWBasic]

@Luposian

 

Quote

Lossless is lossless.  Full Stop.  Period.

 

Err, not quite.

 

The issue is that the term "lossless" is thrown around as a marketing term like "natural."

 

For example: The old DTS CDs from the late 1990s and early 2000s claimed "lossless" 5.1 sound at 44.1khz, and 20-bit audio. They were encoded at 1234.8kbps; yet a 6 channels of 20-bit audio at 44.1khz is 5292kbps. (DTS CDs use the lower 14 bits per sample to try to avoid speaker damage if they are played in a conventional CD player.)

 

That's a 4.2x compression ratio! (Basically, DTS CDs achieve a 23% smaller file size.) It's mathematically impossible to compress an audio signal that much: With FLAC, we typically get files that are 60% smaller than the original .WAV file.

 

So, even though DTS CDs were called "lossless," there were changes to the signal in order to make it fit on the CD.

 

(The same can be said for codecs like AptX, which supposedly allow for "lossless" Bluetooth audio.)

 

So if lossless is lossless, how does that account for the changes in order to make the files so small?

 

Here's an oversimplified example of how "lossless" codecs achieve lower bitrates than FLAC:

 

1: They make the quiet parts louder, and then lower the resolution using noise shaping. At playback, they then make the quiet parts loud again. (This trick works very, very well.)

 

2: They use companding algorithms that are similar to the old Dolby and DBX algorithms that were used on tape. Then, they lower the resolution using noise shaping. At playback, they undo the companding. (Almost all analog masters used systems like this to reduce tape hiss.)

 

Honestly, I find it frustrating to see the term "lossless" thrown around when it's clear that the original signal is modified in some way to achieve bitrates that are mathematically impossible without some form of resolution loss.

[actuel audio]

The AV, technology, and streaming industries have appropriated and developed terms like lossless and high resolution into marketing terms, and a means by which we feel compelled to feel inadequate and will spend money to fix those perceived inadequacies.

Until audiophiles get that, they are just being manipulated for profit. It is a deep chasm of denial to see across but, eventually, you do.

The alternative is perpetual misery and money wasted.

[Luposian]

1 hour ago, GWBasic said:

@Luposian

 

 

Err, not quite.

 

The issue is that the term "lossless" is thrown around as a marketing term like "natural."

 

For example: The old DTS CDs from the late 1990s and early 2000s claimed "lossless" 5.1 sound at 44.1khz, and 20-bit audio. They were encoded at 1234.8kbps; yet a 6 channels of 20-bit audio at 44.1khz is 5292kbps. (DTS CDs use the lower 14 bits per sample to try to avoid speaker damage if they are played in a conventional CD player.)

 

That's a 4.2x compression ratio! (Basically, DTS CDs achieve a 23% smaller file size.) It's mathematically impossible to compress an audio signal that much: With FLAC, we typically get files that are 60% smaller than the original .WAV file.

 

So, even though DTS CDs were called "lossless," there were changes to the signal in order to make it fit on the CD.

 

(The same can be said for codecs like AptX, which supposedly allow for "lossless" Bluetooth audio.)

 

So if lossless is lossless, how does that account for the changes in order to make the files so small?

 

Here's an oversimplified example of how "lossless" codecs achieve lower bitrates than FLAC:

 

1: They make the quiet parts louder, and then lower the resolution using noise shaping. At playback, they then make the quiet parts loud again. (This trick works very, very well.)

 

2: They use companding algorithms that are similar to the old Dolby and DBX algorithms that were used on tape. Then, they lower the resolution using noise shaping. At playback, they undo the companding. (Almost all analog masters used systems like this to reduce tape hiss.)

 

Honestly, I find it frustrating to see the term "lossless" thrown around when it's clear that the original signal is modified in some way to achieve bitrates that are mathematically impossible without some form of resolution loss.

If you restrict or reduce or alter the original (source) recording to any degree, regardless of method, you can no longer call it lossless.  That’s is what I mean by lossless is lossless. No loss. No change.

[bbosler]

49 minutes ago, Luposian said:

If you restrict or reduce or alter the original (source) recording to any degree, regardless of method, you can no longer call it lossless.  That’s is what I mean by lossless is lossless. No loss. No change.

 So define “original (source) recording”

 

unless direct (to tape, to disc, etc) every recording is altered in some way… EQ applied, mixed, mastered, compressed, and so forth. Even direct it can be altered as it is recorded. Recordings to analog tape are digitized then manipulated or manipulated then digitized then manipulated some more. Every vinyl record has pretty severe EQ applied when pressed then reverse EQ applied by the phono stage. 
 

it is very rare for a consumer to get an unaltered original recording , so pretty much nothing is lossless by your definition 

 

forgive me is this has already been stated, I didn’t read the whole discussion 

 

 

[Luposian]

1 hour ago, bbosler said:

 So define “original (source) recording”

 

unless direct (to tape, to disc, etc) every recording is altered in some way… EQ applied, mixed, mastered, compressed, and so forth. Even direct it can be altered as it is recorded. Recordings to analog tape are digitized then manipulated or manipulated then digitized then manipulated some more. Every vinyl record has pretty severe EQ applied when pressed then reverse EQ applied by the phono stage. 
 

it is very rare for a consumer to get an unaltered original recording , so pretty much nothing is lossless by your definition 

 

forgive me is this has already been stated, I didn’t read the whole discussion 

 

 

Whatever the original recording is, is the master (reference).  If you recorded your voice to an .MP3 (ridiculous, but...), then if that is the first recording, then THAT is the original source recording.  As long as you don't alter THAT recording in any negative way (i.e. lossy), then every "version" (.WAV. FLAC, etc.) beyond that source can be considered a lossless copy.  The .MP3 example is a ridiculous one, but if that was the FIRST recording, then that is your original source.

[bbosler]

1 hour ago, Luposian said:

Whatever the original recording is, is the master (reference).  If you recorded your voice to an .MP3 (ridiculous, but...), then if that is the first recording, then THAT is the original source recording.  As long as you don't alter THAT recording in any negative way (i.e. lossy), then every "version" (.WAV. FLAC, etc.) beyond that source can be considered a lossless copy.  The .MP3 example is a ridiculous one, but if that was the FIRST recording, then that is your original source.

Define negative way

Define FIRST recording 

 

Most recordings are altered in some way before they are released. Many are released in many formats at once. Many consist of multiple tracks combined in all sorts of ways with different processing applied to different tracks so there really is no FIRST recording, They may be recorded at 24/192 or 32/384 or DSD then converted to PCM for processing or to tape then digitized for processing then mixed and mastered and EQ’d and compressed and other effects applied before released, those effects may be different for different media. 
 

 

[Luposian]

1 hour ago, bbosler said:

Define negative way

Define FIRST recording 

 

Most recordings are altered in some way before they are released. Many are released in many formats at once. Many consist of multiple tracks combined in all sorts of ways with different processing applied to different tracks so there really is no FIRST recording, They may be recorded at 24/192 or 32/384 or DSD then converted to PCM for processing or to tape then digitized for processing then mixed and mastered and EQ’d and compressed and other effects applied before released, those effects may be different for different media. 
 

 

Ex. You record your voice.  THAT is your FIRST recording.  As is, where is. Regardless of how it was recorded, what effects applied, etc., the end result of THAT recording is your ORIGINAL master.

 

As long as EVERY copy is 100% identical to THAT master, then those copies can be considered lossless.  You’ve “lost nothing” from the original recording.

 

But, of course, you have to know what the meaning of “is”, is! 🤣 

[Luposian]

1 hour ago, bbosler said:

Define negative way

Define FIRST recording 

 

Most recordings are altered in some way before they are released. Many are released in many formats at once. Many consist of multiple tracks combined in all sorts of ways with different processing applied to different tracks so there really is no FIRST recording, They may be recorded at 24/192 or 32/384 or DSD then converted to PCM for processing or to tape then digitized for processing then mixed and mastered and EQ’d and compressed and other effects applied before released, those effects may be different for different media. 
 

 

At some point the entire argument of “lossless” becomes pointless, if people are going to manipulate the original recording a bazillion different ways.

 

How about we just listen to the music/recordings and stop analyzing them, hmm?  Maybe good enough really should be good enough.  Lossless. Lossy. Who the (!!!) cares at some point.  Maybe we should all just shut up and  listen to the music.  Works for me. 

[The Computer Audiophile]

12 hours ago, actuel audio said:

Until audiophiles get that, they are just being manipulated for profit.

It's amazing many audiophiles make it to work every day without someone looking out for them 😳

[The Computer Audiophile]

7 hours ago, Luposian said:

But, of course, you have to know what the meaning of “is”, is! 🤣 

I'm always in for a meaning of is is comment :~)

[The Computer Audiophile]

6 hours ago, Luposian said:

How about we just listen to the music/recordings and stop analyzing them, hmm?  Maybe good enough really should be good enough.  Lossless. Lossy. Who the (!!!) cares at some point.  Maybe we should all just shut up and  listen to the music.  Works for me. 

 

This is kind of what I was getting at in the article. We should use terms that have meaning and use them when they fit and actually tell someone something useful. Lossless if perfct when talking abotu ripping a CD or using dBPowerAmp to convert one format to another. After that, it's really misleading.

[whell]

Interesting train of thought Chris.  I've often wondered the same thing: like maybe we need some kind of numbering system or something similar to identify how far the music we're listening to or purchasing is removed from the source.  Consider:

 

1 - 1 = mastered directly from the source.

1 - 2 = mastered from a copy of the source. 

1 - 2 - 2 = mastered from a copy of the source, and remixed.

1 - 3 - 16 = We've messed with it so much that it's damn near unrecognizable. 

[actuel audio]

10 minutes ago, The Computer Audiophile said:

It's amazing many audiophiles make it to work every day without someone looking out for them 😳

I am interpreting this to mean, despite the best efforts of the industry, audiophiles find a way to justify spending crap loads of money on snake oil? Eh, no, you can't mean that. What do you mean? It is not me looking out. I could give a fig. But, just wanted to toss that thought out that, well, it is an industry and many terms are just fancy ways of identifying a feature and inflating it into more than it actually provides in return. When often, non technically sexy solutions are a better approach or, just don't bother. Life is too short to be sucked into the vortex of the audio industry marketing technical minutia machine. :-)

[bbosler]

8 hours ago, Luposian said:

As long as EVERY copy is 100% identical to THAT master, then those copies can be considered lossless.  You’ve “lost nothing” from the original recording.

 

But, of course, you have to know what the meaning of “is”, is! 🤣 

 

7 hours ago, Luposian said:

At some point the entire argument of “lossless” becomes pointless, if people are going to manipulate the original recording a bazillion different ways.

 

How about we just listen to the music/recordings and stop analyzing them, hmm?

That's the point I'm trying to drive home. Trying to tie "lossless" back to the original recording is pointless. You also mention "master" but the master in today's world is usually not the original recording. It is the final mix after all the manipulations and mixing of various tracks is done, and they may make different masters for different formats. .. 2L offers their new releases in all the formats below. Which is the one we should consider as lossless?

 

So once and again and for the last time (I promise) your idea of going back to the original as being lossless makes no sense in today's world. 

 

 

 

Disc 1	Hybrid SACD MCH 5.1 DSD Stereo DSD RedBook PCM: MQA CD
Disc 2	Pure Audio Blu-ray 2.0 LPCM 192/24 5.1 DTS HD-MA 192/24 7.1.4 Auro-3D 96kHz 7.1.4 Dolby Atmos 48kHz mShuttle: MQA + FLAC + MP3 Region: ABC - worldwide

 

 

[Luposian]

2 hours ago, whell said:

Interesting train of thought Chris.  I've often wondered the same thing: like maybe we need some kind of numbering system or something similar to identify how far the music we're listening to or purchasing is removed from the source.  Consider:

 

1 - 1 = mastered directly from the source.

1 - 2 = mastered from a copy of the source. 

1 - 2 - 2 = mastered from a copy of the source, and remixed.

1 - 3 - 16 = We've messed with it so much that it's damn near unrecognizable. 

Anyone still remember when CD’s used the AAD, ADD, DDD nomenclature?  I was always happy when I bought a CD with the DDD. 😁

[The Computer Audiophile]

4 minutes ago, Luposian said:

Anyone still remember when CD’s used the AAD, ADD, DDD nomenclature?  I was always happy when I bought a CD with the DDD. 😁

Ah yes, the SPARS code https://en.wikipedia.org/wiki/SPARS_code

 

Distinctions without a difference.