How many bits, how fast, just how much resolution is enough?

[BlueSkyy]

It all started way back in about 1982 with the arrival of CDs originally providing a resolution or data rate of 44.1khz/16 bit. Now we can obtain or generate music files at a resolution of 384khz/24 bit or higher.

 

I've personally gotten caught up in the rush to higher resolution and have upgraded several times to be able to play back 384khz/24bit files and also getting into DSD64 and now DSD128. I'm not sure

 

Advances in electronics will, probably, allow for almost limitless increases in resolution and storage space to house those huge files is exceedingly low but at what point does the human ear stop hearing any improvement? Where will it end?

[jjjjusta]

In theory, it ended with 44.1/16-bit.

 

Sent from my SAMSUNG-SM-G891A using Tapatalk

[esldude]

44.1/16, maybe 44.1/24 or 48/24. In extreme cases with exemplary gear and young terrific hearing 48/24 might be needed.

 

Some of the more knowledgeable people for the most extreme possibilities estimate you might need 65 khz/20 bit. As that isn't an available rate, 96/24 or 88/24 does it for you with room to spare. If one just wishes to absolutely end any possibility of any person in the world hearing something in less than max fidelity due to recording limitations then 96/24 is enough and that.....is.......simply......that.

 

Long winded version of jjjjusta's reply.

[Ralf11]

you could see if there are any listening tests comparing - but... it may really be about filter slopes

[gmgraves]

It all started way back in about 1982 with the arrival of CDs originally providing a resolution or data rate of 44.1khz/16 bit. Now we can obtain or generate music files at a resolution of 384khz/24 bit or higher.

 

I've personally gotten caught up in the rush to higher resolution and have upgraded several times to be able to play back 384khz/24bit files and also getting into DSD64 and now DSD128. I'm not sure

 

Advances in electronics will, probably, allow for almost limitless increases in resolution and storage space to house those huge files is exceedingly low but at what point does the human ear stop hearing any improvement? Where will it end?

 

Theoretically, we're there. 24-bit resolution already gives a noise floor that's so far below the threshold of human hearing, as to be totally irrelevant. Not only can humans not hear noise that far down, but it is also more quiet than any analog stage that would follow the D/A conversion. Resolution? 384 KHz sampling rate gives a frequency response flat out to 192 KHz! Even if we allow that some instruments produce sounds quite a bit in excess of 20KHz, none are likely to produce RF, which is what 192KHz is converging on. So, theoretically, noise levels that can't be heard, frequency response that's serious overkil, and vanishing levels of distortion, even in soft musical passages, how much more is necessary?

[r_w]

I can agree with that... in a lot of systems it's probably more about DACs and how they filter than data depth / rate (after Redbook).

 

Redbook sounds great to me, I don't personally see the need to go with anything higher in my system.

 

I've compared very few hi-res recordings vs Redbook but at those few times 16/44 sounded the minutest bit warmer and was preferred - but that could be the re-mastering and/or the DACs filtering effect.

 

It's not enough to concern me either way really, I'd happily play either.

 

(I record at 24/48 but that is a required broadcast industry standard, not a preference)

 

All the system tweaks I've done are allowing me to hear ridiculous levels of detail with Redbook (that I never thought possible) and to be honest I actually don't want to hear any deeper... the balance is right and the music flows - any further dissection might possibly come at the expense of the music = not worth the risk.

 

... but as always YMMV.

 

 

you could see if there are any listening tests comparing - but... it may really be about filter slopes

[BlueSkyy]

Thanks to everyone who responded and it appears that you all are level-headed when it comes to bit depth and sampling rates. At this point I am taking away the points are the infrastructure, i.e. input/output electronics that supports the DAC and the way the filters are designed are where any audible benefits are to be realized. Also, me money I could save by not purchasing hi-rez files with stratospheric bit rates is warranted.

[rando]

Does not one dig a hole bigger than needed before planting a tree? So that loose fertile soil may provide amply for what is to grow over time.

 

The only problem here is what to do with the rocks and soil free of nutrients that were evacuated. Or too much focus on making deep cuts while forgoing a pleasing natural arrangement.

[cjf]

The only problem here is what to do with the rocks and soil free of nutrients that were evacuated

 

No problem at all, just throw that stuff over the fence into your neighbors yard

[audiventory]

In theory, it ended with 44.1/16-bit.

 

It's only part of theory. Ideal case. But in reality we have non-ideal processing and hardware.

 

 

44.1/16, maybe 44.1/24 or 48/24. In extreme cases with exemplary gear and young terrific hearing 48/24 might be needed.

 

Ultrasound playback is not target of high resolution audio.

[Fitzcaraldo215]

From comparative listening, I believe there are and always will be sharply diminishing returns with increased sampling rates and bit depths. There are several ways to view this question, and I think it is important to look at it in terms of the original recording format.

 

A. Conversion to digital from analog masters.

B. Digital recordings at various sampling rates.

 

For A., my opinion is added bit depth beyond 16 bits is essentially wasted. Increased sampling rates beyond 88/96k or DSD64 offer very little if any sonic improvement, but I think hi res offers sonic advantages over 44k, mainly freedom from audible filter artifacts in a-d and d-a. I do not think analog recordings at their best are as good as high rez digital recordings at their best, and remastering analog to digital does not change that, nor does direct playback in analog. I also think that many who have concluded that hi rez is a waste and RBCD is sufficient have erred in their judgement by using source material natively recorded in analog as a basis.

 

For B., in addition to the above, 24 bit PCM is more than sufficient, but that is now the common standard for PCM, even if some, though not all, of the added bit depth beyond 16 bits is wasted. It allows added headroom in the recording process beyond 16 bits, and may often provide a small but noticeable sonic advantage as a result, mainly in low level detail. There are slight, but very subtle advantages to sampling rates above 88/96k or DSD64. Downsampling of hi rez recordings to 44k/16 bit usually produces a slightly but noticeably inferior sound quality, though such downconverted recordings may still be quite listenable musically. Only direct comparison to hi rez reveals the sonic tradeoffs. For native DSD recordings, conversion to PCM also causes a slight, noticeable downside due mainly to the inexactitude of that algorithmic conversion.

 

I have heard a number of such comparisons downloaded from trusted sources together with friends and the above is our consensus on a number of high quality systems. We are classical music listeners with large libraries, mainly SACD, which is DSD64.

 

We find the answer to "increased resolution" is via natively recorded, discrete multichannel, which offers 2.5 times more information than stereo recordings. That is quite obviously noticeable mainly in improved spatial presentation. Higher sampling rates, for which few native recordings exist commercially, have not sufficiently impressed us. We all convert DSD to PCM at 88 or 176k, as we feel that DSP Room EQ requiring PCM offers sonic advantages that far outweigh the slight sonic disadvantage of DSD-PCM conversion.

 

The best recording I have ever heard is a non-commercial classical recording done by an engineer friend of mine in DSD256 5.1 channels. Several Stereophile and one TAS reviewer I know have also heard it and they agree. However, the exotic mikes used and the care in engineering are undoubtedly part of that excellent result. I prefer listening to it with conversion to 176k PCM with my usual Dirac EQ applied. But, it also sounds very good in an almost spooky realistic way in native DSD256.

[plissken]

It all started way back in about 1982 with the arrival of CDs originally providing a resolution or data rate of 44.1khz/16 bit. Now we can obtain or generate music files at a resolution of 384khz/24 bit or higher.

 

I've personally gotten caught up in the rush to higher resolution and have upgraded several times to be able to play back 384khz/24bit files and also getting into DSD64 and now DSD128. I'm not sure

 

Advances in electronics will, probably, allow for almost limitless increases in resolution and storage space to house those huge files is exceedingly low but at what point does the human ear stop hearing any improvement? Where will it end?

 

16/44.1 wasn't a limitation of the technology back then. It was an understanding of where human hearing hits it's limits and they designed the technology around that.

 

16/44.1 steered the design process.

[YashN]

Where will it end?

 

Infinite everything with no noise.

[davide256]

16/44.1 wasn't a limitation of the technology back then. It was an understanding of where human hearing hits it's limits and they designed the technology around that.

 

16/44.1 steered the design process.

 

Pity the designers only had monaural sound in mind...

[plissken]

Pity the designers only had monaural sound in mind...

 

I've as of yet to find anything in our industry to impugn this presentation on why 16/44.1 is good as it gets for us.

 

24/192 Music Downloads are Very Silly Indeed

[audiventory]

16/44.1 wasn't a limitation of the technology back then. It was an understanding of where human hearing hits it's limits and they designed the technology around that.

 

Absolutelly. Base of audio design is 16-20 kHz.

 

16/44.1 was taken as minimal available that time (invention of 16/44).

 

All further increasing of sample rates was based on pragmatical technical reasons (real analog filtration abilities in DAC) on newer technologies.

 

Inceasing of bit depth give expanding of signal/noise ratio in listened dynamic range.

 

DSD's sample rate is matter of signal/noise ratio in audible frequency range in complex with analog filters of DAC.

[r_w]

That was a quite a good read and it generally concurs with my experiences - though I'm no digital expert and I have no crusade either.

 

I've always tended to believe that either audible waves (20-20) at discernible dynamic range are accurately (enough) being encoded and decoded or they are not... aside from that I have always assumed how the ADC / DAC and subsequent circuitry (and filters) dealt/deals with the 'signal' decided/decides the audible and measurable differences in the analogue domain (across 20-20).

 

... I would be interested to hear evidence to the contrary (purely out of interest however, not argument).

 

:-)

 

I've as of yet to find anything in our industry to impugn this presentation on why 16/44.1 is good as it gets for us.

 

24/192 Music Downloads are Very Silly Indeed

[esldude]

Absolutelly. Base of audio design is 16-20 kHz.

 

16/44.1 was taken as minimal available that time (invention of 16/44).

 

All further increasing of sample rates was based on pragmatical technical reasons (real analog filtration abilities in DAC) on newer technologies.

 

Inceasing of bit depth give expanding of signal/noise ratio in listened dynamic range.

 

DSD's sample rate is matter of signal/noise ratio in audible frequency range in complex with analog filters of DAC.

 

Some revisionist history here I think. Before the CD became a standard various digital recording formats were around being experimented with for music. They ranged from 13 bits to 16 bits. Sample rates ran from 32 khz to 50 khz. Some used various pre and deemphasis schemes. When Philips and then Sony decided to back a new standard Philips had decided 14 bits was enough while Sony insisted it be 16 bits. Both agreed sample rate needed to exceed 40 khz by some small amount. As is commonly known we got 44.1 khz because it fit the Sony pro video recorders that were adapted to hold digital data. More than that would not fit the data into video frame efficiently. 48 khz was later adapted for video use due to how it fit with scan and frame rates of the existing broadcast video at the time. For one period of time pro video standards considered using 60 khz because it fit all the various frame rates nicely regardless of country. It was deemed wasteful and unnecessary.

 

So no, 44.1 wasn't chosen as some minimum available at the time. Basically anything above 40 khz would work and the other factors caused it to be 44.1 for CD.

[semente]

With this image I am triying to describe graphically the advantage of a 24 bit file in regards to resolution as it was explained to me by Barry Diament:

 

- the effective resolution of a 16 bit file at -60dB is 6 bit

 

- the effective resolution of a 24 bit file at -60dB is 14 bit

 

Whether this is audible would depend I guess of quiet your room is, of how loud you are playing, of how "transparent" your speakers are.

 

R

 

 

 

[r_w]

The article appears to address that as a misconception?

 

 

"Above: Sampled signals are often depicted as a rough stairstep (red) that seems a poor approximation of the original signal. However, the representation is mathematically exact and the signal recovers the exact smooth shape of the original (blue) when converted back to analog"

 

With this image I am triying to describe graphically the advantage of a 24 bit file in regards to resolution as it was explained to me by Barry Diament:

 

- the effective resolution of a 16 bit file at -60dB is 6 bit

 

- the effective resolution of a 24 bit file at -60dB is 14 bit

 

Whether this is audible would depend I guess of quiet your room is, of how loud you are playing, of how "transparent" your speakers are.

 

R

 

[ATTACH=CONFIG]31047[/ATTACH]

 

 

[ATTACH=CONFIG]31048[/ATTACH]

[semente]

The article appears to address that as a misconception?

 

[ATTACH=CONFIG]31053[/ATTACH]

 

"Above: Sampled signals are often depicted as a rough stairstep (red) that seems a poor approximation of the original signal. However, the representation is mathematically exact and the signal recovers the exact smooth shape of the original (blue) when converted back to analog"

If I'm not mistaken the points are joined using some form of interpolation and the more points you have/take the more accurate curve will be reconstructed.

 

 

R

[semente]

See here:

 

https://en.m.wikipedia.org/wiki/Quantization_(signal_processing)

[mansr]

If I'm not mistaken the points are joined using some form of interpolation and the more points you have/take the more accurate curve will be reconstructed.

 

 

R

 

You are mistaken. See Nyquist.

 

The bit depth matters, of course. Sample rate does not once it is sufficient for the input signal.

[semente]

You are mistaken. See Nyquist.

 

The bit depth matters, of course. Sample rate does not once it is sufficient for the input signal.

Where did I mention sample rate?

[mansr]

Where did I mention sample rate?

 

Your remark about interpolation using more points implied it.