Audio Bit Rates

[Sokel]

There are some crazy ones going around,that's one I downloaded some time ago from a pro gear website (I don't remember which,it was a pack,probably Korg?)

 

[danadam]

Bitrate is determined by the sampling rate and the bit depth.

Of uncompressed audio, yes.
(plus number of channels, of course)

When it comes to audio and bitrates, size does matter. The more kilobits per second the greater the quality of the sound. For many casual listeners, a bitrate of 320 kbps is acceptable. Clearly, CD-quality audio with 1,411kbps

First, it is confusing, because 320 kbps is presumably from a lossy codec and the bitrate of lossy codecs is determined by more things than just sample rate and bitdepth.

Second, as I already wrote once, comparing those two numbers doesn't make sense to me. The difference between 320 kbps lossy codec and 1411 kbps uncompressed file will be the same as between 320 kbps lossy coded and let's say 900 kbps lossless codec. So using 1411 is misleading, IMO.

MP3s are a very common form of a compressed audio file, with a maximum bitrate of 320 kbps at 16 bits

Actually internal format of MP3 is float.

 

[AudioStudies]

Actually internal format of MP3 is float.

I am not sure what you mean.

 

[voodooless]

I am not sure what you mean.

It means MP3 is not bound to 16 bits. You can feed the encoder perfectly fine with 24-bit audio. And you could also decode to 24-bit audio. The internal representation is nothing like this. In fact, even saying that internally it uses floats is not exactly accurate. It works mostly on the frequency and phase domain, not so much on sampled audio.

 

[AudioStudies]

aac at 128kbit.

This really seems to work out very well in my situation, AAC at 128 kbps (or higher) really rocks. I am just getting started with this and hope to come across more stations with Opus and OGG so I can evaluate those. MP3 doesn't seem too bad if it is at 320 kbps. I guess there is something called AAC+ also. I think the worst that I came across was MP3 at 32kbps and I (of course) didn't save that channel. Thanks to everyone who shared thoughts, it very much enhanced my understanding of this topic. For a good many years, I was just in the CD-quality only crowd and made no effort to dig much deeper into these other formats. I am very pleasantly surprised by AAC 128 and higher.

 

[AudioStudies]

First, it is confusing, because 320 kbps is presumably from a lossy codec and the bitrate of lossy codecs is determined by more things than just sample rate and bitdepth.

Perhaps in my next version of the document, I could make it clear that size matters with bitrate when comparing lossless audio files, and that with lossy audio bitrate is only one of many considerations. I think @Killingbeans was hitting on that also.

 

[fieldcar]

WAV files that are 16-bit 44.1 Kz (CD quality)

You should really use flacs unless you need extremely low overhead like in a DAW. FLAC is bit perfect, and you can convert back to WAV with not a single bit lost if you want. I love ez audio converter for it's batch processing. Totally worth it. They keep up with the latest licensing from fraunhoffer, and if you install itunes, you can use apple's AAC encoder without ever launching itunes.

Music Converter with Professional Audio Quality | Convert Music and Rip Audio CDs to/from xHE-AAC, AAC, FLAC, MP3, M4A, Opus, DSD, WAV, OGG, and more audio file formats

Music Converter converts music between FLAC, MP3, AAC, xHE-AAC, AAC, WAV, DSD, Opus and many more audio file formats. Securely rip audio CDs in bit-perfect audio quality. Burns audio CDs. Metadata editor for audio files.

www.poikosoft.com

I guess there is something called AAC+ also

xHE-AAC and AAC-exhale (open source and based on HE-AAC and AAC-LC, maxxes out at 192kbit) are lower bitrate focused codecs that kick ass. xHE-AAC is widely unsupported despite the compression being pretty amazing and maxxing out at 320kbit. Again, this is why it's worth it to just buy a license for ez audio converter to tinker with these codecs.

 

[Apesbrain]

Clearly, CD-quality audio with 1,411kbps will sound better, particularly to discerning listeners.

High Bitrate MP3 Internet Blind Test: Part 2 - RESULTS

A blog for audiophiles about more objective topics. Measurements of audio gear. Reasonable, realistic, no snakeoil assessment of sound, and equipment.

archimago.blogspot.com

 

[AudioStudies]

You should really use flacs

I have devices from the pro audio market that don't do FLAC. I use WAV to keep everything consistent, even so some of my devices can do FLAC.

 

[AudioStudies]

just buy a license for ez audio converter to tinker with these codecs.

I appreciate the info. I don't really tinker, I just listen on internet radios that can either handle the codec or not.

 

[Jeromeof]

High Bitrate MP3 Internet Blind Test: Part 2 - RESULTS

A blog for audiophiles about more objective topics. Measurements of audio gear. Reasonable, realistic, no snakeoil assessment of sound, and equipment.

archimago.blogspot.com

I love seeing the result of these blind tests -

"As you can see, despite the confidence, most of the respondents thought that Set A (the original lossless audio) sounded worse than Set B (MP3)."

Its a pity we don't see more blind tests.

 

[AudioStudies]

From a comment in the link that @Apesbrain provided:

"The problem is that folk who are only exposed to lossy have that as an audio frame of reference."

I have long suspected that "frame of reference" is an important consideration in audio that I haven't seen discussed very much. I think the reason that some people cling to vinyl and tubes is because they have listened for so long that way, that has become their "frame of reference" -- that is they are convinced that is how music is supposed to sound. It would be interesting to have a study conducted with children growing up and have them exposed to various formats, and then change the format when they get to a certain age. I suspect they would cling to their frame of reference in many cases, even if they didn't start out with the scientifically best format.

 

[AudioStudies]

I love seeing the result of these blind tests -

"As you can see, despite the confidence, most of the respondents thought that Set A (the original lossless audio) sounded worse than Set B (MP3)."

Its a pity we don't see more blind tests.

I appreciate the efforts to conduct such a test, and I too would like to see more blind tests. I would be curious to know though, how this particular test would stack up against official protocols for testing by experts such as Floyd Toole and similar luminaries. One thing I picked up on is that some pretty crappy headphones were used for those that used headphones, or at least crappy by Amir's review standards. It doesn't surprise me that people found it difficult to tell the difference. But if one wants to make the claim that lossy is better -- I can see no theoretical scientific justification for that. How would we even speculate? Did the compression algorithm remove some things from the music that were undesirable?

 

[Killingbeans]

From a comment in the link that @Apesbrain provided:

"The problem is that folk who are only exposed to lossy have that as an audio frame of reference."

I have long suspected that "frame of reference" is an important consideration in audio that I haven't seen discussed very much. I think the reason that some people cling to vinyl and tubes is because they have listened for so long that way, that has become their "frame of reference" -- that is they are convinced that is how music is supposed to sound. It would be interesting to have a study conducted with children growing up and have them exposed to various formats, and then change the format when they get to a certain age. I suspect they would cling to their frame of reference in many cases, even if they didn't start out with the scientifically best format.

Absolutely!

I often wonder a bit when people talk about their "preference" in audio gear. Is it something that's more more suited to their personality and/or anatomy? Or is it simply something that feels familiar to them? I wouldn't be surprised if, in a lot of cases, it's the latter.

I think that's also one of the reasons why the high-end industry is so enamoured with the concept of "burn-in". They sell it as a physical phenomenon, but in reality it just gives the customer the time they need to get past the feeling of "This is unfamiliar to me, and therefore I don't like it" (in the cases with an actual audible difference).

 

[AudioStudies]

I think that's also one of the reasons why the high-end industry is so enamoured with the concept of "burn-in". They sell it as a physical phenomenon, but in reality it just gives the customer the time they need to get past the feeling of "This is unfamiliar to me, and therefore I don't like it" (in the cases with an actual audible difference).

Fantastic words.

 

[danadam]

WMA​

Up tp 96 kHz​

Up to 768 kbps​

Lossless​

No​

Yes​

You mixed here WMA (lossy) with WMA Lossless. The 768 kbps limit is for WMA (lossy).

I am not sure what you mean.

Just a technical nitpick. As voodooless already explained, the bit-depth concept does not really apply in codecs like MP3. Under some circumstances, i.e. when there is no strong signal to mask quiet signals, MP3 is able to preserve signals much lower than 16 bit. This usually doesn't matter because:

• it's so quiet that you won't hear it,
• decoders mostly decode to 16 bit anyway,

but it is possible to decode to float and recover it. For example the mp3 file in attachment was created from file containing tones at -80 and -140 dBFS:

lame decoder decodes to 16 bits, so the second signal is buried in noise:

but with ffmpeg we can decode to float wav and get it back:

ffmpeg -c:a mp3float -i "quiet.mp3" -c pcm_f32le  "ffmpeg.wav"

Of course once there is some strong signal there:

it would be unreasonable to spend bits trying to preserve the quiet one:

 

[bennetng]

WAV is a container format, while the popular use is to store uncompressed PCM data, the actual data format is defined by the format tag. For example, WAV can be used to store ADPCM, which is a lossy format with (usually) 4-bit, and decoded to 16-bit during playback. WAV header also contains a bit-depth tag, but in some cases this tag should be set to 0:

Some compression schemes cannot define a value for wBitsPerSample. In this case, set wBitsPerSample to zero.

WAVEFORMATEX (mmreg.h) - Win32 apps

The WAVEFORMATEX structure specifies the data format of a wave audio stream.

learn.microsoft.com

The later introduced extensible WAV format has the same file extension, but supports more features like multichannel mapping and so on.

WAVEFORMATEXTENSIBLE (ksmedia.h) - Windows drivers

The WAVEFORMATEXTENSIBLE structure specifies the format of an audio wave stream.

learn.microsoft.com

Here is an example showing the importance of the format tag, if it is wrong, the WAV file won't be recognized by other software:

Beta-test: DeltaWave Null Comparison software

Just added a page to DeltaWave website with data reprocessed from the Gearslutz AD/DA loopback test page. It's a subset of all the files. Since this was done automatically, by a script, there may be some things that are mislabeled. Let me know and I'll fix them...

www.audiosciencereview.com

 

[DVDdoug]

It means MP3 is not bound to 16 bits. You can feed the encoder perfectly fine with 24-bit audio. And you could also decode to 24-bit audio.

I have a bit-depth checker called Bitter that works with Audacity. If I check an MP3 ripped from CD it shows more than 32-bits. A simple volume change should also introduce rounding errors, filling all of the bits/ bytes with data and increasing the measured bit depth.

"As you can see, despite the confidence, most of the respondents thought that Set A (the original lossless audio) sounded worse than Set B (MP3)."

Not THAT surprising... Personal preferences vary, especially when there are small differences. But an ABX should really be done first to determine if the listener(s) can actually- reliably hear a difference before asking for a preference.

Its a pity we don't see more blind tests.

I agree. But it's not easy thing to do. To get useful "review" results you'd need a panel of listeners and the same panel every time.

Or you can do an ABX testy yourself to see if you can hear a difference. If you want to compare CODECS/formats, there is software for that. You can do it with hardware but you need somebody to do the blind switching and the levels have to match... If there is a volume difference between A & B, you'll identify X every time and the odds are that you'll prefer the louder one. .

I appreciate the efforts to conduct such a test, and I too would like to see more blind tests. I would be curious to know though, how this particular test would stack up against official protocols for testing by experts such as Floyd Toole and similar luminaries.

Generally, you can measure better than you can hear so a measurable difference obviously isn't always an audible difference.

Compression artifacts can be an exception. It's not easy to measure compression artifacts and they may not show-up in the "standard" measurements. People often measure the spectrum to show a loss of high frequencies with MP3 but when someone hears an MP3 compression artifact it's usually not the loss of highs that they hear.

One thing I picked up on is that some pretty crappy headphones were used for those that used headphones, or at least crappy by Amir's review standards.

Again, compression artifacts are somewhat of an "exception". The guys at HydrogenAudio who do ABX tests will tell you that headphones are better than speakers but high-end headphones usually don't make the defects easier to hear. It mostly depends on the listener and the program material.

 

[Steven Holt]

I want to thank AudioStudies and everyone else for this excellent discussion, I learned quite a bit. I would like to throw out a question : If the upper range of human hearing is 20Khz, and -- generally speaking -- hearing deteriorates with age, would it be a good idea for people over 50 to just rip their CD's to MP3 and save storage space? If you can't hear the difference, is there really a difference?

 

[voodooless]

I want to thank AudioStudies and everyone else for this excellent discussion, I learned quite a bit. I would like to throw out a question : If the upper range of human hearing is 20Khz, and -- generally speaking -- hearing deteriorates with age, would it be a good idea for people over 50 to just rip their CD's to MP3 and save storage space? If you can't hear the difference, is there really a difference?

MP3 isn’t just cutting > 16 kHz, so even people above 50 will hear differences if the bitrate is low enough.

MP3 is just legacy at this point, so it would not make senses to use it if you were to undertake this. Better use AAC, Opus or OGG.

But storage is cheap, and if you already have it stored lossless, you clearly have the space.

I’d say: waste of time.