What is true lossless? For DIGITAL Audio Only.

chander

Well-Known Member
Joined
Nov 9, 2009
Messages
1,341
Points
113
Location
Goa
As the title states -
1 - What is true lossless?
2 - How do you know your files are truly lossless?
3 - Is it just the size?
4 - Is it possible to get truly lossless DIGITAL audio today? Only Qobuz comes to mind - maybe Apple music.
 
1. PCM is the pure and native form of digital audio.
2. Cannot make out - .Wav or .Flac are lossless containers of music data, designed to hold PCM data, but both can also be loaded with .mp3. So a .Flac or .Wav file by itself means nothing. But here the file size will show if it is holding a .mp3. But my next point is also important.
3. Used to be yes, but with modern remastering software like PGGB, size is no longer a arbiter of sound quality. We can today convert a 320 KBPS .MP3 to .Wav or .Flac with 16 bit / 44.1 khz resolution. So no way to find out.
4. Best bet is a original first recording audio CD. I dont know about Qobuz or apple music as I havent used both.

I bet my answer was the most useless one ever. Wrote a lot...but didnt say anything in the end 🤣
 
[/QUOTE]
I bet my answer was the most useless one ever. Wrote a lot...but didnt say anything in the end 🤣
[/QUOTE]

Point no.4 did mean a lot at the end:)
 
So what I understand is - a true lossless file is true to the source? If that is correct - if I create a FLAC file from an MP3 file - then "technically" that would be lossless too?
 
So what I understand is - a true lossless file is true to the source? If that is correct - if I create a FLAC file from an MP3 file - then "technically" that would be lossless too?
No, because, mp3 is already a lossy compression technique and info lost cannot be recovered. When you create flac from mp3, technically you aren’t losing anything but whatever is lost is lost.
 
No, because, mp3 is already a lossy compression technique and info lost cannot be recovered. When you create flac from mp3, technically you aren’t losing anything but whatever is lost is lost.
Exactly right....

When we choose to rip a disc, the software gives us the choice of the container to put it in. Which is .Wav for lossless and uncompressed PCM, but file size is big. And .Flac for lossless PCM but compressed this time for a slightly smaller file size. This encapsulated PCM, once converted into any of the lossy formats like .Mp3, will be throwing away bits of information in the process. And though there are remastering engines trying to use sophisticated algorithms to regenerate the lost bits, it is still guess work at best. And bits once lost, can never again be recovered.

So file size is never an indication of lossless or lossy audio. The best option for the OCD folks, is to make their own rips. And a lot of folks actually do that.
 
No, because, mp3 is already a lossy compression technique and info lost cannot be recovered. When you create flac from mp3, technically you aren’t losing anything but whatever is lost is lost.
Yup I understand that - & that is what I am trying to point out.

If I get some FLACs :) where I have no idea of the source material, I can never be sure of whether or not that is actually a "real lossless file" - created from a "real lossless source". Then - my FLAC pride is pure placebo.

So there is really no way to be sure whether or not the files you hold - "flac/lossless" are really lossless unless you have yourself done a CD rip? How many here have actually tested their FLAC collections for "lossless" authenticity?

Just curious, how many of us are just going by the ".flac" extension & actually listening to MP3s dressed as a super model?
 
Just curious, how many of us are just going by the ".flac" extension & actually listening to MP3s dressed as a super model?
True. You will have to undress (the file and not the model) to find that out. If it is mp3, then there will be lot of high end cut off. There will be a sharp cutoff at 20 kHz. Absolutely 0 content above 20 kHz. Lossless audio will extend above 20 kHz. You can use audacity to find that out. Also this utility here can help you figure that out - https://losslessaudiochecker.com/#detectionalgorithms

1711866206297.png
 
Here is another utility that helps you analyze the music. I have used this and it works on Linux, Mac and Windows.

 
So what I understand is - a true lossless file is true to the source?
No such thing as a "true" lossless file. For compressed formats it's black or white - ergo, lossy or lossless. "True" does not figure in the equation here.
In the case of a compressed audio format, if after decompression the resultant file is exactly the same as the source it was compressed from, it is lossless.
If that is correct - if I create a FLAC file from an MP3 file - then "technically" that would be lossless too?
Correct! I was wondering when someone would point that out.
 
Truly lossless? You know it is lossless only if you ripped the CD with no errors. Nothing else is guaranteed to be lossless.
someone burnt those bits into CD tracks and they should have saved the true digital master correct? (16 or 24 bits). i think THAT is true lossless. everything else is just an approximation. CD can't be that for sure, it is just an archival format like tape. even CD retrieval has so many errors, EAC is a thing right? but then who guarantees that CDs were burned with zero erros? can any music label guarantee a bit perfect consistency between all CDs from the same digital master?

is the digital master truly consistent with the analog master? that will be decided by the golden ears. but then we are not debating that here since we are only concerned about 'digital trueness' 😁

but FLAC is a great archival format because most modern computers can render them without auditory artefacts (aka. no observable delay/distortion, similar to lossless TIFF if you are into photography). wav is a truly perfect digital format because it is like RAW file from a camera.
 
Last edited:
but then who guarantees that CDs were burned with zero erros? can any music label guarantee a bit perfect consistency between all CDs from the same digital master?
Errors may creep in:

1. When the CD is pressed.

2. As the CD ages and the reflective layer deteriorates / flakes off

3. The plastic CD disc is scratched and the laser cannot properly focus and read the data pits on the metalised layer.

A "Good' CD could have over 1 million bits of error!

CDs have rather robust error correction built into the way the data is written. Roughly 25% of the data written onto a CD is dedicated for Error Correction, to yield "Perfect" playback of the music.

CDs deploy Cross-interleaved Reed-Solomon Code.... Often referred to simply as Reed-Solomon Error Correction, named after the mathematicians than developed it.

Hence it not necessary to have a perfect CD to get 'perfect' sound. :)
 
Errors may creep in:

1. When the CD is pressed.

2. As the CD ages and the reflective layer deteriorates / flakes off

3. The plastic CD disc is scratched and the laser cannot properly focus and read the data pits on the metalised layer.

A "Good' CD could have over 1 million bits of error!

CDs have rather robust error correction built into the way the data is written. Roughly 25% of the data written onto a CD is dedicated for Error Correction, to yield "Perfect" playback of the music.

CDs deploy Cross-interleaved Reed-Solomon Code.... Often referred to simply as Reed-Solomon Error Correction, named after the mathematicians than developed it.

Hence it not necessary to have a perfect CD to get 'perfect' sound. :)
I'd rather hold onto my CD player then 🤣
 
Errors may creep in:

1. When the CD is pressed.

2. As the CD ages and the reflective layer deteriorates / flakes off

3. The plastic CD disc is scratched and the laser cannot properly focus and read the data pits on the metalised layer.

A "Good' CD could have over 1 million bits of error!

CDs have rather robust error correction built into the way the data is written. Roughly 25% of the data written onto a CD is dedicated for Error Correction, to yield "Perfect" playback of the music.

CDs deploy Cross-interleaved Reed-Solomon Code.... Often referred to simply as Reed-Solomon Error Correction, named after the mathematicians than developed it.

Hence it not necessary to have a perfect CD to get 'perfect' sound. :)
so why we have dedicated ripping software like EAC if the system is so robust? i am here to learn. does it mean we need specialized hardware?
 
As the title states -
1 - What is true lossless?
2 - How do you know your files are truly lossless?
3 - Is it just the size?
4 - Is it possible to get truly lossless DIGITAL audio today? Only Qobuz comes to mind - maybe Apple music.
Strictly, Lossless (compression) is a process, based on a particular algorithm.

Lossless is a process that creates a lossless file that is smaller in size (for more economical transmission or storage), which, can be processed later to yield the original file.

To test if the process or if the file is genuinely lossless, you MUST have the original file to compare with, bit-by-bit .... and yes, the original file could be any (audio) file, subjected to audio lossless compression.

A .wav file is the closest, un-compressed Audio File format to the CDDA bit arrangement on a 16/44 (Redbook) CD.

Some pointers I would like to highlight:

1. HOW LOSSY

Just a pointer on how lossy mp3 is :eek: .... a 128 kbps Constant Bit Rate (lossy compressed) mp3 file, Throws Away (never to be recovered again) audio data bits & creates a mp3 file that is 90% smaller. A 40 MB .wav file is reduced to a 4 MB 128 kbps mp3 file.

a 320 kbps CBR mp3 file (about the best mp3 you can create, THROWS AWAY (Lossy Compression) audio data bits & creates a mp3 file that is 74% smaller than the original. :eek:

2. LOSSLESS is NOT NECESSARILY INAUDIBLE
A Lossless file format like FLAC (40% to 50% compression), will yield back all the original bits when decompressed.... But that DOES NOT prove that a FLAC file will sound as good / same as a wav (uncompressed) file.

The decompressing process, requires computational horse power and the additional strain on the PC / Processor decompressing the FLAC in real time, i.e. while simultaneously playing out digital music, has led many to state that as a reason why some listeners are convinced (includes me) that wav files sound better than FLAC ...

My entire 16/44 library or self ripped CDs is in .wav... I have deleted all (earlier) downloaded 16/44 Flac albums.
 
Last edited:
so why we have dedicated ripping software like EAC if the system is so robust? i am here to learn. does it mean we need specialized hardware?
The Reed Solomon Error correction is VERY Robust in principle.... Mathematically, I think it can correct for a drop out of upto 2.5 mm along the CD track. However in principle, there are other non perfections in the system...

* Like CD mechanical wobble while spinning at a high speed,

* Imperfect laser focusing (amongst other reasons... due to laser power fading with age)

* Scratched CD surface etc

In practice, most, well designed, new CD players will yield upto 1 mm drop outs (not 2.5 mm) as in audible.

CD Ripping Software has the luxury of time to go back and re read a CD segment that the Reed Solomon code is unable to correct. The Ripping software also has to accommodate the imperfect characteristics of the CD Ripping Drive like CRC (Cyclic Redundancy Check) error.


I am personally convinced that each ripping software, yields its own distinctive sonic signature.
I find EAC rips have more 'flow' while dB PowerAmp rips have more PRAT. ( YMMV ;) )

To answer your question: "does it mean we need specialized hardware?"

YES!


Better hardware, for example, a better Ripping drive will yield better rips.

A better CD Drive (eg a Philips CDM Pro) provides audibly better sound, in a CD Player.

... and Yes, different CD Drives have their particular sonic signatures.... Most CEC Belt Drive CD transport users will swear that their CD transport sounds more 'analog' than other transports......

Of course, every thing I have said in the above 2 posts is my opinion and unlikely to be universally acceptable.

I am only sharing my take take on the questions posed, and since you mentioned " i am here to learn." :)
 
Last edited:
The Reed Solomon Error correction is VERY Robust in principle.... Mathematically, I think it can correct for a drop out of upto 2.5 mm along the CD track. However in principle, there are other non perfections in the system...

* Like CD mechanical wobble while spinning at a high speed,

* Imperfect laser focusing (amongst other reasons... due to laser power fading with age)

* Scratched CD surface etc

In practice, most, well designed, new CD players will yield upto 1 mm drop outs (not 2.5 mm) as in audible.

CD Ripping Software has the luxury of time to go back and re read a CD segment that the Reed Solomon code is unable to correct. The Ripping software also has to accommodate the imperfect characteristics of the CD Ripping Drive like CRC (Cyclic Redundancy Check) error.


I am personally convinced that each ripping software, yields its own distinctive sonic signature.
I find EAC rips have more 'flow' while dB PowerAmp rips have more PRAT. ( YMMV ;) )

To answer your question: "does it mean we need specialized hardware?"

YES!


Better hardware, for example, a better Ripping drive will yield better rips.

A better CD Drive (eg a Philips CDM Pro) provides audibly better sound, in a CD Player.

... and Yes, different CD Drives have their particular sonic signatures.... Most CEC Belt Drive CD transport users will swear that their CD transport sounds more 'analog' than other transports......

Of course, every thing I have said in the above 2 posts is my opinion and unlikely to be universally acceptable.

I am only sharing my take take on the questions posed, and since you mentioned " i am here to learn." :)

The Reed Solomon Error correction is VERY Robust in principle.... Mathematically, I think it can correct for a drop out of upto 2.5 mm along the CD track. However in principle, there are other non perfections in the system...

* Like CD mechanical wobble while spinning at a high speed,

* Imperfect laser focusing (amongst other reasons... due to laser power fading with age)

* Scratched CD surface etc

In practice, most, well designed, new CD players will yield upto 1 mm drop outs (not 2.5 mm) as in audible.

CD Ripping Software has the luxury of time to go back and re read a CD segment that the Reed Solomon code is unable to correct. The Ripping software also has to accommodate the imperfect characteristics of the CD Ripping Drive like CRC (Cyclic Redundancy Check) error.


I am personally convinced that each ripping software, yields its own distinctive sonic signature.
I find EAC rips have more 'flow' while dB PowerAmp rips have more PRAT. ( YMMV ;) )

To answer your question: "does it mean we need specialized hardware?"

YES!


Better hardware, for example, a better Ripping drive will yield better rips.

A better CD Drive (eg a Philips CDM Pro) provides audibly better sound, in a CD Player.

... and Yes, different CD Drives have their particular sonic signatures.... Most CEC Belt Drive CD transport users will swear that their CD transport sounds more 'analog' than other transports......

Of course, every thing I have said in the above 2 posts is my opinion and unlikely to be universally acceptable.

I am only sharing my take take on the questions posed, and since you mentioned " i am here to learn." :)
i didn't know, thanks for sharing!

also, do the labels verify for the correct 'burning' of the CDs? are all the CDs from Taylor Swift burned identially? we are talking only hypothetical things which may not make any difference in the output but still it is interesting to know... will the ECC algorithms in the playback device correct for burning errors as well, not just for retrieval errors? in other words, does the original CDA file contain ECC data necessary for the original digital master? i didn't think it had but happy to be corrected. cheers
 
Last edited:
Follow HiFiMART on Instagram for offers, deals and FREE giveaways!
Back
Top