Poll: Which is more important to high-resolution audio, sampling frequency or bit depth

[new_media]

Regarding 16/44 vs. 24/44 vs. 24/96 lossless PCM, which is more important in your opinion?

[wgscott]

Bit depth.

 

Where is the poll, btw?

[blownsi]

Bit depth

 

 

x2

[new_media]

As you probably know, you don't get to type in your poll options until you submit a new thread.

[wgscott]

Sorry.

[bdiament]

Hi new_media,

 

I voted both. Word length without the bandwidth is just as bad as bandwidth without the word length.

 

To my ears, 24/44, while fixing the low level issues of CD does nothing about its other issues, and 24/96 still sounds "very good" (which I consider a distortion -- I don't want the medium to "sound" at all; I want it to get out of the way).

 

I've worked with all sorts of analog and digital formats over the years, and it wasn't until I heard properly done 4x (i.e., 24/192) that, for the first time in my experience, I had a recording format that provided output I could not discern from the input.

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

[Jimmypowder]

No doubt bit depth

[YashN]

Quad DSD.

[audiventory]

Bit depth and sample rate is 2 parts of big number components recording-mixing-system.

 

Also

 

Information = BitDepth x SampleRate

 

DSD have only 1 bit, but it have quality (noise floor) better PCM 16 bit: -140 dB vs. -110...120 dB. DSD's features depend on used processing, of course.

 

DSD has big reserve by sample rate + noise shaping that allow get lower noise floor than 16 bit.

 

I.e. always need consider resolution audio in frame of used software and hardware.

[bdiament]

Hi new_media,

 

I voted both. Word length without the bandwidth is just as bad as bandwidth without the word length.

 

To my ears, 24/44, while fixing the low level issues of CD does nothing about its other issues, and 24/96 still sounds "very good" (which I consider a distortion -- I don't want the medium to "sound" at all; I want it to get out of the way).

 

I've worked with all sorts of analog and digital formats over the years, and it wasn't until I heard properly done 4x (i.e., 24/192) that, for the first time in my experience, I had a recording format that provided output I could not discern from the input.

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

 

To me, it is like asking whether horizontal resolution in video is more important than vertical resolution.

What good is great resolution in one, without the same in the other?

 

Happily, there is no need to choose. We *can* have both.

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

[Jabs1542]

There's really no good way to compare because no one offers higher resolution sample rates at lower bit depths (i.e. 16:96 or 16:192).

[Daudio]

I voted for 'Bit Depth', because over the years here on CA, I've seen many more people say they could more clearly tell the difference between 16 and 24 bit music file, rather then 44.1k versus any other higher rate.

[beetlemania]

There's really no good way to compare because no one offers higher resolution sample rates at lower bit depths (i.e. 16:96 or 16:192).

 

That's a good point but, in my trials, I can distinguish 24/44 from 16/44 but not so much 24/44 from 24/96 or 24/192. I hold Mr. Diament's opinion in high regard but my own ears do not seem very sensitive to higher sampling rates.

[krass]

bit depth, but I think it's because of availability of 44/24 and 48/24 and lack of high-bandwidth 16 bit material to compare.

 

compared to (say) 44/16, 44/24 offers "50% more bits" and I easily hear the difference.

Comparing (say) 48/24, 96/24 offers "100% more bits". I hear a difference but it's subjectively less. Trouble is, it's already 24 bit....

 

Problem is I don't have any high bandwidth material at 16 bits. To make a proper comparison I would need to hear 66/16 and compare it to 44/24, which of course doesn't exist.

 

I suspect that the same comparative improvement of either bit-depth OR bandwidth over 44/16 would have an equal effect; I just don't have the material to prove it so my emotion is out-voting my brain and currently convincing me that going 44/16 to 44/24 is the most significant.

[Tony Lauck]

bit depth, but I think it's because of availability of 44/24 and 48/24 and lack of high-bandwidth 16 bit material to compare.

 

compared to (say) 44/16, 44/24 offers "50% more bits" and I easily hear the difference.

Comparing (say) 48/24, 96/24 offers "100% more bits". I hear a difference but it's subjectively less. Trouble is, it's already 24 bit....

 

Problem is I don't have any high bandwidth material at 16 bits. To make a proper comparison I would need to hear 66/16 and compare it to 44/24, which of course doesn't exist.

 

I suspect that the same comparative improvement of either bit-depth OR bandwidth over 44/16 would have an equal effect; I just don't have the material to prove it so my emotion is out-voting my brain and currently convincing me that going 44/16 to 44/24 is the most significant.

 

I did some tests and I would say that 192/16 is definitely better than 44/16. Some time ago I tested 96/12 and found it to be better than 44/16, however getting that to work out required some tweaking of the dither settings. It would be easy to extrapolate that to 192/12 is better than 44/16.

 

Roughly speaking, once the sampling rate is well above the minimum required (which is about 48 kHz for acceptable quality) it's pretty much the total bits at play that count. That's why 1 bit DSD is significantly better than 44/16, all things being equal. (They aren't equal, but pretty much so with good implementations.) My guess is that 192/8 could even be made to sound better than 44/16, as there would be more than enough bits at play, but I didn't test this. (My comments about comparisons that I did were all validated by theory, measurements on sample rate conversions of test signals, measurements on the music in various formats and listening tests.)

 

If you substitute bit sampling rate for bit depth and push it too far, then the sound quality will not just depend on the format used, but it will also depend on the sensitivity the your preamplifier and amplifier to high frequency (dither) noise. In this regard the situation is similar to that with DSD64.

 

By the way, if subtractive dither is used and some extra noise is mixed in on playback, then even 8 bits at 44 kHz can sound musical. There won't be any distortion, but there will be a fair amount of "tape hiss". Nostalgia for me, as it reminds me of my youth, listening to pre-recorded 4 track tapes before there was decent stereo LP playback available. But then, not everyone can tolerate a little tape hiss. I can, and that's why I believe that cassette tapes (in good condition) can sound better than most CDs.

 

 

More Bits are Better.

[Miska]

Higher sampling rate can be used to achieve equivalent of having more bits (higher word length). Without extra effort, doubling sampling rate effectively gives equivalent of adding one extra bit to word length. With noise shaping, you can achieve 96 kHz bandwidth using 192 kHz sampling rate at 16-bit, while getting 24-bit worth of dynamic range in audio band.... For that, you only need to trade a bit over 10 dB of dynamic range at frequencies above 20 kHz.

 

So using higher sampling rate you can achieve both higher dynamic range (equivalent of adding bits) and better frequency/time domain performance at the same time. This is not possible by just adding more bits.

 

It is also easier to achieve higher real world linearity using higher sampling rates and noise shaping than it is by adding more bits to the conversion process.

 

For this to make sense, you need to look at entire end-to-end chain from analog to analog. Sticking purely to digital domain theory won't give you the right picture of end results...

[Jay-dub]

There's really no good way to compare because no one offers higher resolution sample rates at lower bit depths (i.e. 16:96 or 16:192).

 

You can always take 24/96 or 24/192 files, dither them to 16-bit and compare with the original. I haven't tried it, but there's more room for noise-shaping to work, so with the best choice you can expect less audible degradation than is the case when you dither to 16-bit at 44.1 kHz. You can try it, as Tony Lauck did. It doesn't require expensive software, although it will require experimenting with different kinds of noise-shaped dither. But that isn't the question here, which is about where 24/44.1 comes subjectively on the quality spectrum from 16/44.1 to 24/96. To me, both steps up are about equally audible, but the increase in bit depth is much more significant musically, especially in music with a wide dynamic range such as a solo piano, which can have thundering climaxes between long quiet passages where very subtle things are happening in the music. The difference between standard bandwidth properly done and extended bandwidth is similar in nature to, but subtler than the way a sound changes when you turn your head.

 

For me the biggest difference between 44.1 kHz sampling and 96 kHz is that 96 kHz sounds very good pretty much regardless of which antialiasing filter was used, whereas 44.1 and 48 need a perfectly chosen filter to sound very good. Barry Diament's standard by which the jump up from 96 to 192 is more significant than the step from 44.1 to 96 is curious, but ultimately irrelevant to me, since I've never heard a live microphone feed of good music in the studio. The difference between different filters at 96 kHz is subtle enough that I'm not really interested in higher rates.

[jabbr]

Sampling frequency is more important for the fact that with a high sample frequency you can effectively add bit depth e.g. DSD. So a bit depth of 1 is effective at a high enough sample frequency. The converse is not true. 48 bits at 22khz doesn't cut it.

[audiventory]

22 kHz sample rate anyway don't suitable for audible range 0 ... 20 kHz for any bit depth

 

High sample rate give more freedom for developer. But demand additional price (in wide sense, not only money) too.

[adolf512]

sanpling frequency by experience, good luck finding a recording that even take advantage of 14 bits. Also i have not found noise to be that harmfull(people seams tl enjoy vinyl).

[electricity]

a good read on the subject

 

24/192 Music Downloads

...and why they make no sense

 

'Unfortunately, there is no point to distributing music in 24-bit/192kHz format. Its playback fidelity is slightly inferior to 16/44.1 or 16/48, and it takes up 6 times the space.'

 

'192kHz digital music files offer no benefits. They're not quite neutral either; practical fidelity is slightly worse. The ultrasonics are a liability during playback.

 

Neither audio transducers nor power amplifiers are free of distortion, and distortion tends to increase rapidly at the lowest and highest frequencies. If the same transducer reproduces ultrasonics along with audible content, any nonlinearity will shift some of the ultrasonic content down into the audible range as an uncontrolled spray of intermodulation distortion products covering the entire audible spectrum. Nonlinearity in a power amplifier will produce the same effect. The effect is very slight, but listening tests have confirmed that both effects can be audible.'

 

'In summary, it's not certain that intermodulation from ultrasonics will be audible on a given system. The added distortion could be insignificant or it could be noticable. Either way, ultrasonic content is never a benefit, and on plenty of systems it will audibly hurt fidelity. On the systems it doesn't hurt, the cost and complexity of handling ultrasonics could have been saved, or spent on improved audible range performance instead.'

 

'Sampling rates over 48kHz are irrelevant to high fidelity audio data, but they are internally essential to several modern digital audio techniques. Oversampling is the most relevant example.

 

Oversampling is simple and clever. You may recall from my A Digital Media Primer for Geeks that high sampling rates provide a great deal more space between the highest frequency audio we care about (20kHz) and the Nyquist frequency (half the sampling rate). This allows for simpler, smoother, more reliable analog anti-aliasing filters, and thus higher fidelity. This extra space between 20kHz and the Nyquist frequency is essentially just spectral padding for the analog filter.

 

That's only half the story. Because digital filters have few of the practical limitations of an analog filter, we can complete the anti-aliasing process with greater efficiency and precision digitally. The very high rate raw digital signal passes through a digital anti-aliasing filter, which has no trouble fitting a transition band into a tight space. After this further digital anti-aliasing, the extra padding samples are simply thrown away. Oversampled playback approximately works in reverse.

 

This means we can use low rate 44.1kHz or 48kHz audio with all the fidelity benefits of 192kHz or higher sampling (smooth frequency response, low aliasing) and none of the drawbacks (ultrasonics that cause intermodulation distortion, wasted space). Nearly all of today's analog-to-digital converters (ADCs) and digital-to-analog converters (DACs) oversample at very high rates. Few people realize this is happening because it's completely automatic and hidden.'

 

'None of that is relevant to playback; here 24 bit audio is as useless as 192kHz sampling. The good news is that at least 24 bit depth doesn't harm fidelity. It just doesn't help, and also wastes space.'

 

'Why push back against 24/192? Because it's a solution to a problem that doesn't exist, a business model based on willful ignorance and scamming people. The more that pseudoscience goes unchecked in the world at large, the harder it is for truth to overcome truthiness... even if this is a small and relatively insignificant example.'

[Tony Lauck]

a good read on the subject

 

24/192 Music Downloads

...and why they make no sense

 

'Unfortunately, there is no point to distributing music in 24-bit/192kHz format. Its playback fidelity is slightly inferior to 16/44.1 or 16/48, and it takes up 6 times the space.'

 

'192kHz digital music files offer no benefits. They're not quite neutral either; practical fidelity is slightly worse. The ultrasonics are a liability during playback.

 

Neither audio transducers nor power amplifiers are free of distortion, and distortion tends to increase rapidly at the lowest and highest frequencies. If the same transducer reproduces ultrasonics along with audible content, any nonlinearity will shift some of the ultrasonic content down into the audible range as an uncontrolled spray of intermodulation distortion products covering the entire audible spectrum. Nonlinearity in a power amplifier will produce the same effect. The effect is very slight, but listening tests have confirmed that both effects can be audible.'

 

'In summary, it's not certain that intermodulation from ultrasonics will be audible on a given system. The added distortion could be insignificant or it could be noticable. Either way, ultrasonic content is never a benefit, and on plenty of systems it will audibly hurt fidelity. On the systems it doesn't hurt, the cost and complexity of handling ultrasonics could have been saved, or spent on improved audible range performance instead.'

 

'Sampling rates over 48kHz are irrelevant to high fidelity audio data, but they are internally essential to several modern digital audio techniques. Oversampling is the most relevant example.

 

Oversampling is simple and clever. You may recall from my A Digital Media Primer for Geeks that high sampling rates provide a great deal more space between the highest frequency audio we care about (20kHz) and the Nyquist frequency (half the sampling rate). This allows for simpler, smoother, more reliable analog anti-aliasing filters, and thus higher fidelity. This extra space between 20kHz and the Nyquist frequency is essentially just spectral padding for the analog filter.

 

That's only half the story. Because digital filters have few of the practical limitations of an analog filter, we can complete the anti-aliasing process with greater efficiency and precision digitally. The very high rate raw digital signal passes through a digital anti-aliasing filter, which has no trouble fitting a transition band into a tight space. After this further digital anti-aliasing, the extra padding samples are simply thrown away. Oversampled playback approximately works in reverse.

 

This means we can use low rate 44.1kHz or 48kHz audio with all the fidelity benefits of 192kHz or higher sampling (smooth frequency response, low aliasing) and none of the drawbacks (ultrasonics that cause intermodulation distortion, wasted space). Nearly all of today's analog-to-digital converters (ADCs) and digital-to-analog converters (DACs) oversample at very high rates. Few people realize this is happening because it's completely automatic and hidden.'

 

'None of that is relevant to playback; here 24 bit audio is as useless as 192kHz sampling. The good news is that at least 24 bit depth doesn't harm fidelity. It just doesn't help, and also wastes space.'

 

'Why push back against 24/192? Because it's a solution to a problem that doesn't exist, a business model based on willful ignorance and scamming people. The more that pseudoscience goes unchecked in the world at large, the harder it is for truth to overcome truthiness... even if this is a small and relatively insignificant example.'

 

The world is full of deaf idiot assholes. The language is that of a second rate engineer who doesn't know what he doesn't know.

[Bill Lord]

a good read on the subject

 

24/192 Music Downloads

...and why they make no sense

 

 

Monty is a good guy. He and his team are behind the Free Lossless Audio Codec (FLAC). https://xiph.org/flac/

 

What he says at the end is what I like people to focus on every time this article comes up. Which is:

• Better Headphones
  
• Lossless Audio
  
• Better Masters
  
• Surround
  

We can have it all, 24 bit master versions,better equipment/lossless audio, and multichannel audio.

[YashN]

What he says at the end is what I like people to focus on every time this article comes up. Which is:

• Better Headphones
  
• Lossless Audio
  
• Better Masters
  
• Surround
  

We can have it all, 24 bit master versions,better equipment/lossless audio, and multichannel audio.

 

If instead, people also learned what the characteristics for good music reproduction are that are not related to frequencies, we would be winning even more.

 

Instead, we get people posting this crappy article time and time again.

[Bill Lord]

I disagree that the article is crappy; it's strong counterpoint. You do not have to like the points. Instead could you elaborate on the characteristics that you believe for good music reproduction. Specifically, which is more important to audio fidelity? Thanks