Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: I thought lossless bitrate = quality and lower rate meant less quality (Read 28010 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

I thought lossless bitrate = quality and lower rate meant less quality

Hi there,

I've recently purchased a lot of music from various online stores. Bought a few tracks off Beatport, Digital Tunes, SURUS and Juno.

I've noticed that the lossless versions of tunes that they sell vary wildly in sample rate.

Surus, for example, Blocks & Escher's - Broken is sampled at 1002kbps, sold as a Flac, which is acceptable. Whereas DJ Madd's - New Reality is only 457 kbps, (also sold as FLAC). Which is pretty pointless since I might as well have just bought the MP3 instead. (50pence less)

I'd assume if these places are selling lossless quality versions of tunes that they'd be CD quality, 1441kbps or thereabouts.

Who's to say that the 1441kbps file I downloaded from Beatport, DJ Shadow - Blood on the Motorway, was just converted from a lossy version of the track? This might be me becoming a bit paranoid though.

Is this not the store's fault? Is this just down to the labels sending over lossy copies of their tunes?


It'd be great if anyone could enlighten me on the matter.

Cheers

I thought lossless bitrate = quality and lower rate meant less quality

Reply #1
Before jumping to conclusions perhaps you should educate yourself on what lossless compression is and how it works.  HA or Wikipedia is your friend.
Creature of habit.

I thought lossless bitrate = quality and lower rate meant less quality

Reply #2
Yeah, really, how could it be compression if it did not compress the file? Bitrate does not equal quality in lossless formats; it only equals file size—but, in doing so, it DOES equal their entire purpose: to reduce file size without loss.

Besides, you said “sample rates”, which are something else entirely. Short of locking this thread outright, I’m at least going to rename it to something that isn’t misleading as the result of a (fundamental) misunderstanding.

I thought lossless bitrate = quality and lower rate meant less quality

Reply #3
I've noticed that the lossless versions of tunes that they sell vary wildly in sample rate.


It's not the sample rate that varies. What you're seeing is the result of different degrees of compression achieved with different files. This is normal for lossless encoding. The 'bitrate' you see is that of the compressed file, not the original material. Lossless codecs will achieve different amounts of compression with different audio source material - some audio compresses easily and some does not, so it will vary.

Uncompressed audio from standard redbook CD has a fixed bitrate. It is stereo (two channels), 16 bits per sample, 44,100 (44.1 kHz) samples per second.

2 x 16 x 44,100 = 1,411,200 bits per second (bps), or approximately 1411 kbps

But when compressed using FLAC or another lossless codec, the bitrate will be less. A particular track that compresses 30% will have a bitrate of 70% of 1411 kbps

30% compression -> (1 - 0.30) x 1411 = 988 kbps

and so on...

40% compressed -> 847 kbps
50% compressed -> 706 kbps
60% compressed -> 564 kbps
65% compressed -> 494 kbps
70% compressed -> 423 kbps

In my own FLAC library I have files that have compressed as little as 21% and some as much as 76%. So, yes, the bitrate can be expected to vary wildly.

I thought lossless bitrate = quality and lower rate meant less quality

Reply #4
The DJ Madd track is mono (or very nearly mono). It will compress relatively well. That's why its FLAC bitrate is much lower than usual.

 

I thought lossless bitrate = quality and lower rate meant less quality

Reply #5
What you're seeing is the result of different degrees of compression achieved with different files. This is normal for lossless encoding. The 'bitrate' you see is that of the compressed file, not the original material. Lossless codecs will achieve different amounts of compression with different audio source material - some audio compresses easily and some does not, so it will vary.
[…]
30% compression -> (1 - 0.30) x 1411 = 988 kbps

and so on...

In addition to this, note that if you only see one bitrate displayed, it’s the mean (often called average, but that’s a less specific term) bitrate of the entire file. The actual bitrate at any given time will vary with the complexity of the music during that period (a small amount of time called a frame, which has its own bitrate). Lossless compression has—and, by definition, must have—a variable bitrate.

I thought lossless bitrate = quality and lower rate meant less quality

Reply #6
So are you saying that the song with a bitrate of 1002kbps is of the same audio quality as the tune that has bitrate of 457kbps?


I thought lossless bitrate = quality and lower rate meant less quality

Reply #7
yes. within the context of this thread - differing bit rate of compressed lossless file = same lossless quality

Quote
The 'bitrate' you see is that of the compressed file, not the original material. Lossless codecs will achieve different amounts of compression with different audio source material - some audio compresses easily and some does not, so it will vary.



I thought lossless bitrate = quality and lower rate meant less quality

Reply #8
A mono track obviously only needs half the amount of data, compared to stereo.

I thought lossless bitrate = quality and lower rate meant less quality

Reply #9
Yeah, really, how could it be compression if it did not compress the file? Bitrate does not equal quality in lossless formats; it only equals file size—but, in doing so, it DOES equal their entire purpose: to reduce file size without loss.

So are you saying that the song with a bitrate of 1002kbps is of the same audio quality as the tune that has bitrate of 457kbps?

It might be hard to wrap your head around it, after being drilled for years by online shops and advertisers that bitrate is a quality metric: To save as much disk space or bandwidth as possible, in fact you'd want the lowest bitrate possible. For lossy encoding this means an encode at the lowest bitrate that still achieves transparency. For lossless audio this is of course quite obvious; since no audio information is lost, you'd want the resulting file to be as small as possible.
It's only audiophile if it's inconvenient.

I thought lossless bitrate = quality and lower rate meant less quality

Reply #10
So are you saying that the song with a bitrate of 1002kbps is of the same audio quality as the tune that has bitrate of 457kbps?


Go to random.org and generate a text file from random characters (10000 strings of 20 characters, using both numbers caps and minuscles) file. That gives you 10000 lines, each of 20 characters. Save it into a .txt file.
Compress it into a .zip file.

Then create a text file consisting of 20000 lines, each of 20 characters, where all characters are space. That is the same size as the previous .txt file.
Compress it into a .zip file.


Both zip files will be unzipped to their respective originals, yet the latter is much smaller. Is that because it is of lesser quality?

I thought lossless bitrate = quality and lower rate meant less quality

Reply #11
To put it another way: Is that because the compression of the second text document caused more loss? No: Neither caused any loss. The second document was just easier to compress because it was simpler.

Lossless audio compression works the same way as any other lossless compression, except it’s tuned especially for audio. Some audio streams are more easy to compress than others, and so the resulting losslessly compressed file will be smaller. Since bitrate in kbit/sec simply equals (size of file in kbit)÷(duration of file in sec) [excluding headers, metadata, etc. for the sake of clarity of illustration], a program will report the aforementioned smaller file as having a lower (mean) bitrate. In actuality, its bitrate will vary instantaneously with any given frame, depending upon the relative complexity of the audio in that frame.

In none of these cases is anything lost. Hence the name.

[Edit:] Also, to add to both this post and everyone’s earlier explanations: It’s important to know the difference between the bitrate of the compressed file and the bitrate of the actual (un-/de-compressed) audio stream. Whatever the bitrate—i.e. file size—of a lossless file, it will be decompressed to the same bitrate as the original source—i.e. the data-rate of the actual audio stream that will be sent to the hardware. For example, a WAV file ripped from an audio CD may be compressed from its native 1411 kbps to a lossless 987 kbps; for playing or decoding, this will be expanded (again, without loss) back to 1411 kbps uncompressed.

I thought lossless bitrate = quality and lower rate meant less quality

Reply #12
Both zip files will be unzipped to their respective originals, yet the latter is much smaller. Is that because it is of lesser quality?

Better yet, try it with the FLAC encoder itself:
Code: [Select]
flac --force-raw-format --channels=1 --bps=8 --sample-rate=44100 --sign=unsigned --endian=big -o test.flac test.txt

will happily encode a file of raw text and even give a bitrate for you.  Then decompress it with:
Code: [Select]
flac --force-raw-format --sign=unsigned --endian=big -d test.flac -o test2.txt

and compare the two text files.  They'll be identical, regardless of what bitrate the FLAC file winds up as.

I thought lossless bitrate = quality and lower rate meant less quality

Reply #13
So are you saying that the song with a bitrate of 1002kbps is of the same audio quality as the tune that has bitrate of 457kbps?

Well, yes and no. If you are using a lossless format, and the audio hasn't been transcoded, then yes, the quality is always perfect; it doesn't matter what the bitrate is. If you are using a lossy format like MP3, then the quality might be constrained by the bitrate, thus generally there's some correlation between bitrate and quality. But even then, it's possible for a high-bitrate MP3 to sound worse than a low-bitrate one, for various reasons...and you also have to define "quality", since the point at which you can't tell the difference is the point where the bitrate no longer matters.

In both MP3 and FLAC, quiet, simple, mono audio doesn't require as much bandwidth as louder, noisier, more complex passages. For music from CD, you'll see FLAC bitrates start around 125 for pure silence, maybe 300-600 for parts that are mono music, and roughly 800-1200 for loud, noisy parts with wide stereo. That's just what I see with my music, anyway.

We're always trying different analogies to explain this stuff. Here's one:

Imagine a large hardcover book and a compact paperback edition with half as many pages. The words are the same, but the paperback is printed smaller, with more words on each page... "losslessly compressed, like with FLAC," one might say. Both have the same number of words and take the same amount of time to read. If you express the reading rate in words per minute, it's the same for both editions. If you express the reading rate in pages per hour, though, the hardcover has double the rate of the paperback. Since nothing is missing, there can't be any quality difference between them. The fact that one has half the pages-per-hour reading rate of the other has no bearing on quality. This is basically what you're asking; pages per hour, kilobits per second... same thing.

Now imagine an abridged paperback, where the story has been trimmed so that there are actually fewer words, like MP3-style lossy compression. In this case, the page count might be cause for concern. Half as many pages might mean the story has been gutted in an aesthetically displeasing way. Or maybe it has been quantitatively gutted, but not in noticeable, qualitative ways. Or maybe only a few words were taken out, and they don't affect the story. Or maybe they do. So the relationship between perceived quality and the page count / reading rate could be strong, moderate, slight, or nil; it depends.

Make sense?

I thought lossless bitrate = quality and lower rate meant less quality

Reply #14
Both zip files will be unzipped to their respective originals, yet the latter is much smaller. Is that because it is of lesser quality?

Better yet, try it with the FLAC encoder itself:


I don't think that's my first suggestion to a beginner who needs to know the concept of compression, but I was curious enough to try.
Uncompressed:

219.980 strings0.txt
219.998 strings1.txt

7z:
232 strings0.7z
158.369 strings1.7z

FLAC
117.006 strings0.flac
204.762 strings1.flac


Obviously not optimized for this