How To: Process High Resolution Music Files
Jun 18, 2015 at 1:24 AM Thread Starter

#### HiResSndWizzard

Hi all,

I'm wanting to learn everything about the, how to, of oversampling, vs software upsampling to recover the more or less original sound of the music file. It is definately there, more or less, in every file I have processed to date.

I have been told, you never loose anything by downsampling a file, only when you compress the bit depth. Also, 16 bit is enough depth to capture all of the music, at least the part you really want to hear below 10khz, easily storing up to 22khz with minimal apparent losses.

Where the resolution is lost, (the fine detail), is in the compression from 24 bit in the master file down to 16 bit of CD & mp3.

Is this detail really worth saving? For the record, relating to the "Sampling Theorem", dynamic range decreased by 'log²(N). By removing 4 bits 256 times to form a 20 bit, & again removing another 4 bits 256 times to form a 16 bit.

Volume resolution suffers severe loss of the original dynamic quality by removing the bit depth. View this in a form of  steps from the basement to the attic.

The bottom steps are very close, You take them on the run, hundreds of them at a time, the middle steps are normal size, the top steps, say -6 dB to -4 dB, you need a ladder to get up to the next step, by -2 dB you need a helicopter, to clime up to the 0 dB ceiling breaker, you'd need a rocket ship to make the last upper step.

That is log²(N). 65,535 volume steps in 16 bit, 16.7 million volume steps in 24 bit. I think the count starts at -196 dB, so by -84 dB average noise floor, you have used most of your' aloted volume steps, to reach to the top of the file @ 0 dB

Going up in bit depth, can you dither? Well, yes. But you have already lost the sine resolution of the upper file frequency limit, say 22,050 to 24,000 frequencies become mostly triangle, & stay that way for the most part, after dithering from 16 bit to 24 bit.

By researching the subject, I came up with some interesting points of knowledge about this digital high res revolution.

First. I found proof of the above statement that nothing is lost in a sample clock down compression of equal bit depth.

The Nyquest theorem: It is found in the 'Bell Laboratories, "Sampling Theorem,"' of Harry Nyquest & Claude Shannon, where the 'aliasing of any frequencies present in the capture are only reproduced if the sample rate is at least twice the frequency of the highest captured frequency.

If there is not enough samples in playback, the highest frequencies in the bit depth, all of the frequencies above half the sample rate, are folded in reverse frequency progression into the spectrum of sound heard by your' ear.

Ie: (at 44.1 sample 23khz is aliased at 21khz, 24khz is aliased at 20khz, for a file that started as a 48k16bit & downsampled to 44.1k16bit). These frequencies originally captured in the bit depth at 16 bit, are then reproduced into the upper limits of your' audio, as a form of musical related noise.

The aliasing sound can be heard as mostly a loud room echoing, or an increase in the "Live" sound of the recording studio.

It is made up largely of ultrasonic sounds made audioible by the aliases of those sounds being sampled at a lower than the Nyquest frequency on playback.

IM Distortion in digital is called Aliasing. It degrades the reproduction quality, of the original stereo sound image produced by your' favorite speakers.

By upsampling alone, the aliases are erased by most software converters, if an alias filter is employed during upsampling. That is to say, you loose a small part of the resolution package hidden in the bit depth.

The Nyquest theorem: If the upsampling software does not apply aliasing filters, then the high frequency limit of the original recording can be reproduced by simply upsampling to at least twice the original sample capture frequency of the music file.

Under some conditions, I get a better result, if I upsample from 44.1 to 48 first, without the use of an aliasing filter during that upsample.

Then, I again upsample to 96k from 48k, this time using the aliasing filter to remove everything above 24khz, which is generally the original frequency bandwidth of most music.

Therefore, you don't want to add any aliases to the 96k upsample above 24khz original music data when upsampling to twice the original sample rate. Again, the Nyquest theorem..

When processed properly, you get a finished 96k sample file with data existing only below 24khz. All frequencies, including noise above 24khz, between 24khz & 48khz, is digitally suppressed down to around - 144dB or more

Does this all make a lot of money & cents, when it comes to selling or buying so called high res music files? When it is possible to recover that same file, plus or minus a possible noise factor formula of "SNR = sqrt(N)", from a lowly 16 bit mp3 file?

I have achieved some amazing results to say the least, using several methods of resolution recovery in just about any available file of 16 bit or better.

How has the tech world been doing it?: Can it really be done well by anyone, considering the fact that what is recovered by any means of upsampling [as well as the original master] is subject to having up to 16 dB more noise by doubling the record input bandwidth from 20khz to 40 khz, if those frequencies are not suppressed during the original file capture.

That refers to making the original master music file capture at 96khz sample, & having live capture of frequencies that bats would have trouble hearing, everything from 20 Hz to 48kHz.

That is pretty much saying, "what you recover in resolution from the bit depth of mp3's, as well as the market proclaimed Hi Res file, is, pretty much as they say, not much more than musical noise, your' pet bat might enjoy!

Oh! Don't forget the random ultrasonic noises picked up by the microphones around the recording studio, like antique ultrasonic denture cleaners, & other sounds, one would normally never imagine existed in a sound recording studio.

So I am focusing on recovery of the original 8 or so octives of the music found below 10 khz as my goal of improving the resolution of that segment of the Hi Res upgrade to 96k, 24 or 32 bit. 'C', Octive 0, starts @ 16Hz. octive 8 is up there around 4.5kHz. to 9kHz.

Sorry to say this, but the greatest leap in quality of recorded music, has to come from the sound studio engineers that ultimatly create the sound contained in the music file.

What is your take on the best way, "to take what you already bought in your' music collection", & improving it with various sampling methods, rather than going out to the cleaners and re-purchasing all of your' favorite misic tracks with Pono downloads?

Sorry Neil for the slap down of your' new compression file. I'm not sure any high res file sold at any price is worth any extra cost markup above what the average mp3 is worth.

Give me a rundown of how 'you' apply the conversion, & using what hardware, software, & OS?, in any answers to this thread

Thanks for your' thoughts.

Jun 18, 2015 at 7:03 PM
Doing any processing or resampling of your audio is not going to add or recover details. After reading your entire post I would say that you don't have a good understanding of the concepts you are discussing. There is a lot of false information in there.

Jun 18, 2015 at 7:15 PM
16-bit / 44.1 kHz is the highest resolution you will ever benefit from. There is no loss in the sense of anything that is audible.

Since you asked about upsampling, you may be interested in HQPlayer. It does oversampling, dithering, noise shaping, modulation, and other DSP. To me, it sounded much better than other players, but it's expensive and the interface is awful.

Jun 20, 2015 at 5:57 AM
Ok, Tell me exactly what incorrect information you are pointing to, that differs from my understanding of the Nyquest Theorem? If you can convince me that my college professors didn't teach me the correct information on the subject of signal losses in digital, then I will go back to school & demand my money back.

And you say what?

Do you know anything about the world of digital sampling? What happens within the file, when it's container is changed, (codec')? What is an Alias, & what happens to the file if that Alias is filtered from the compression? Why is it, some files can be upsampled to original resolution, while others are stuck at the resolution they were compressed into?

I know more than you think, but to tell the truth, I am more interested in what the average public really knows about high res music files. Call it a fishing expedition!

You see, it is this way. The recording industry has on one hand, turned out many productions down sampled to 44.1 CD & left all of the digital artifacts intact in the file to allow the file to be returned to it's original studio quality resolution, while other smarter music companies have been digitally filtering out these artifacts, blocking, or limiting the recovery of the original quality of the file.

This filtering deal has been going on for a period of time now, ever since some key music distributors, like I-tunes, started looking at the world market demands for better sounding music files to play on their home stereo systems.

At some point, you could take a tin can and turn it into battle ship armor plate. The little ole MP3, at one time had everything necessary to return it to a perfect 48k24bit sound track. Now the mp3 that are sold on I-tunes, you can hardly drag near FM broadcast quality out of it @ 16khz audio spectrum.

You have any thing to say about these trends, & whether this is going to make the music companies rich all over again on what they already rubed us the original quality from the files they already sold us, to create a new market for those files we already have in our home music catalogue stripped of their digital artifacts?

Tell me sir, What do you know about this subject? Do you know the difference between 3 pack 32 bit or state 1, 4 pack 32 bit? Do you  know that when converting a file depth, you remove 4 bits, 256 times to move from 24 bit to 22 bit? Do you know which bits to remove & still preserve the resolution in the bit depth? Do you know how this is applied in variable bit rate files?

Rather than telling me what I don't know, how about you start telling me what you do know about this discussion? I am here to learn from you and anyone else, but I wasn't born yesterday.

I have a nice day, every day!

Jun 20, 2015 at 9:09 AM

Oh yes, I have read about the HQPlayer, & it's pricy projected currency setback. & without reading the file you suggested about the resolution, I fully agree with you & the text file, when it comes to what is really better than what we might think we already have at the 44.1/16 level. If only the music manufacturing industry would have left a good thing, a good thing, & not skim the cream from the top before selling us that CD. Read below for an explanation.

When it comes to oversampling. That in itself covers a lot of digital sins, when it comes to cleaning up what you do hear. The main reason behind this approach to listening to your' current music catalogue is: it unlocks the high frequency alias from within the file bit depth, & returns it to it's original capture frequency. 24k is played 24k & not 20.1k.

Windows 7 & later, with a sound card that supports the resolution, has a property setting in the sound, so you can default play all files from your' computer, at an oversample of 96,000s 16bit.

Sorry, but that is the best qualty default, 16 bit. Just the 96k oversample alone, makes all of the old CD's & mp3's leap to their feet, & come to life, without all of that alias noise racket in the high frequency portion of the sound.

Since most average music files start in the studio at 48k 24bit & is 24 khz in bandwidth, the alias created when downsampling, 1/2 the Nyquest frequency of 44.1 or in terms of frequency, is 22,050hz, all of the sounds captured between 22,050, & 24000 hz are still there, laid down in reverse in the bit depth of the sample, highest frequency in the lowest slot, & vice versa.

Call it intermodulation distortion, for want of a better word you may have heard of, only in digital terms it is called an alias. It states in Nyquest Theorem, the alias for 24000 hz would exist in the 44.1 downsampled file at or close to  20,100 hz along with everything else that is suppose to be there at that frequency, & at the same time, creating a form of noise as a result when played back at the default 44.1ksample. Well, I'm sure you know what that sounds like.

All of these aliases of the original resolution at 24khz lie between 20 & 22khz in the new 44.1 file. This is the hard sound associated with downsampled CD & mp3 formats. Oversampling unlocks that alias so when listen to, you hear the original frequency band without the alias distortion above 20khz. You actually hear all of the original 24khz spectrum of the original studio recording without doing nothing but turning on the oversample in windows 7 properties.... COOL! REALLY?

The music marketing industry, including most of the record companies were well aware of this long time ago. There is even CD players with over sample capabilities.

So what did they start doing, & it continues to this day? They started using a digital filter to remove the information above 20khz before they put the file on a CD or sell it as an mp3 on I-tunes. They claimed it improved the sound quality, but I know better than that, & so do all of you out there. The old switch. Upgrade to remove what they don't want you to have any more.

The other method of madness, to create a new high resolution market place, was to on the fly down-sample the file to 32 k sample first to kill everything above 16 khz, [known as the brick wall of sound death], then still while on the fly creating the mp3, they run an alias filter over the 32 k sample file to remove that artifact, then & only then is the file up sampled back to the mp3 file container.

Needless to say, that mp3 isn't worth the hard drive space it's stored on, let alone pay I-tunes for it. Need I say? They also got to all of the CD files currently sold to us today, with the digital alias filter, to remove the artifacts of the 24 khz file it came from. You will note there is little or nothing existing above 20khz on that type of CD. Don't take my word for it, get a spectrum viewer and take a look yourself.

All, so they could create a new market place out of the whole milk, into dribs & drabs of cream & 2%. Oh & lets not forget all of the cheese they make out of the whole milk when it gets too old to sell. That part of the irony, is retracted, when I say I love Cheese, but not files purposely stripped of the high res alias data, so they can sell me the missing alias, with a new HD file download.

Here's to you, Pono! Sorry for the jab, Neil.

I am not happy at all about what the big business has done to make life & the sound of music better, for all 'wee' (as in little) fokes, that don't sit on the board of directors, you know, the ones who forces the left over skin & bones, broken down music files, down our throat.

I want everyone out there to have the oppertunity to buy a first quality file the first time, & not for some crazy, dizzy, HD file priceing schedule either , albeit in a good codec compressed form, so later on, if I so choose, I can upsample, to my delight, & hear what "You make lovin' fun", originally sounded like through the speakers in Studio City, LA. {Fleetwood Mac would like that?}

We need to spread the word, & put a stop now to the wholesale ripping off of aprox 4khz of file stripped from the current download sale base, & on the other hand, left in the file to call it HD.

I don't think the music industry should be catagorizing one file capture in one studio, as 5 different qualities of sound, so they can charge more for a closer front seat, in front of your' living room stereo, & mine?

{{Don't forget one of the sound qualtites spoken of here, is the digital "phased simulated stereo" file made & marketed now a days for earphone only listening to. (Talk about Glorified Mono! Try downloading a sample file & analyzing it sometime)}}

This should be the bottom line: The industry should be looking more at the quality of the recording itself made in a quality studio by better quality performers, in order to look for higher prices in what I would want to buy in the first place to play on my 1/4 million dollar stereo in my mansion in Beverly Hills. (Justin B wouldn't like that)

For all of the rest of us, \$0.99 is too big a price to pay for the junk files we are being lead to believe is standard definition, vs HD Music!

And the HDPlayer? Put it on the next flight to Mars. All technical wonders will wind up on Mars some day anyway.

I admit to one thing. I don't want to hurt the feelings of anyone, especially ascap & riaa, but it doesn't seem right for the industry to try to make us believe something that is not true.

To start with, The old time 44.1 music CD that existed for a time before they were stripped of their alias data was just fine & made great HD music when oversampled correctly. The music industry just had to ruin it, with ideas of making more money again on the same HD file that started in most every recording studio.

I  have a better day, every day! You all, out there that took the time to read this war story, have a better day, too.

Yep, I'm a lead guitarist, & proud of it.  I make my own HD music.

Have a listen to this track: Reba McEntire - I'll Have What She's Having. (Now that is a happy girl.

It's even better in 3D with my HD Matrix Sound Wall Stereo! Ask me, I might get loose lipped & tell you how it works.

Jun 20, 2015 at 10:43 AM
Some of what you said doesn't make much sense. When you convert a 16-bit / 44.1 kHz (which is designed to handle all the dynamic range and frequencies we can hear) file to a 24-bit file, or vice versa, it does nothing to change the sound. Just convert it and hear for yourself. It's physically impossible to tell the difference, whether it's been converted from "HD" to Red Book or vice versa. Resolutions of 24-bit or higher only have use in studio settings for advanced computer processing when editing audio. It does nothing for playback. On the other hand, real-time oversampling is DSP and, combined with other DSP as in the case of HQPlayer, can change the sound.

Jun 20, 2015 at 11:00 AM
I take a different approach. If I don't like the sound quality of something, I remaster it. Now I know that doesn't technically increase the sound quality, but afterwards it sure is more pleasing to listen to.

Mostly I use transient enhancers and mid-side processing but I'll also use Waves noise reduction if there's excess hiss. I also like using Slate FG-X to help older recordings. Slate Virtual Mix rack is fun to work with also. Really good software EQ's are available as well ( I use Equalibrium)

Every track I remaster usually only takes a few minutes

Jun 20, 2015 at 11:22 AM
I take a different approach. If I don't like the sound quality of something, I remaster it. Now I know that doesn't technically increase the sound quality, but afterwards it sure is more pleasing to listen to.

Mostly I use transient enhancers and mid-side processing but I'll also use Waves noise reduction if there's excess hiss. I also like using Slate FG-X to help older recordings. Slate Virtual Mix rack is fun to work with also. Really good software EQ's are available as well ( I use Equalibrium)

Every track I remaster usually only takes a few minutes

Now that's good thinking! Wish I was good enough to remaster my music collection...but alas, I am but a humble multi-instrumentalist with no studio know-how.

Jun 20, 2015 at 3:10 PM
Two factors, 1 the frequency bandwidth is locked to the sample rate.

factor 2, the volume steps are locked to the bit depth.

starting somewhere down around -198 dB lets call that the start of the volume steps heading towards 0 dB. The formula is Dynamic range increases by log²(N) bits.

the number of voltage samples between -198 dB & 0 dB in a 16 bit file has only 65536 individual steps of volume. Remember this, the increase in the volume step is log², making every increase of every sound conform to this scale, with the step size increasing as you get closer to 0 dB by log².

at first you might think that the middle of -84 dB would have half of the bits above & half the bits below this value, in fact, the overwhelming majority of bits are below the -84 dB level.

When the volume starts to reach the upper volume level near 0 dB, each increase in volume/frequency audibly becomes much larger, causing a harsh sound at or near 0 dB.

actually the hard harshness sound starts around -6 dB an becomes intolarable by the time it gets above -3 dB.

One way around this harshness, is to keep the volume of the file below -6 dB, where the steps are not so harsh.

When you convert to 24 bit from 16 bit, you are replacing this voltage volume grid with a grid that has 16,700,000 steps, yeah, that's millions of steps.

The volume resolution is dramatically improved, however up near 0 dB the sound still suffers a little bit.

It is not until you move to 32 bit does the high level volume tend to be plesent to hear, near 0 dB.

32 bit has the capacity of 4.3 billion individual voltage volume steps between -196 dB & 0 dB.

So there is a dramatic improvement in the sound quality when moving up to a higher density bit depth. No dithering required.

I have a nice day, every day. I hope you do too.

Take a listen to George Strait - right or wrong, in high definition. Unbelievable You have to hear it for your self. What a good studio recording. We need more studio's around that sound that good.

Jun 20, 2015 at 4:01 PM
Two factors, 1 the frequency bandwidth is locked to the sample rate.

factor 2, the volume steps are locked to the bit depth.

starting somewhere down around -198 dB lets call that the start of the volume steps heading towards 0 dB. The formula is Dynamic range increases by log²(N) bits.

the number of voltage samples between -198 dB & 0 dB in a 16 bit file has only 65536 individual steps of volume. Remember this, the increase in the volume step is log², making every increase of every sound conform to this scale, with the step size increasing as you get closer to 0 dB by log².

at first you might think that the middle of -84 dB would have half of the bits above & half the bits below this value, in fact, the overwhelming majority of bits are below the -84 dB level.

When the volume starts to reach the upper volume level near 0 dB, each increase in volume/frequency audibly becomes much larger, causing a harsh sound at or near 0 dB.

actually the hard harshness sound starts around -6 dB an becomes intolarable by the time it gets above -3 dB.

One way around this harshness, is to keep the volume of the file below -6 dB, where the steps are not so harsh.

When you convert to 24 bit from 16 bit, you are replacing this voltage volume grid with a grid that has 16,700,000 steps, yeah, that's millions of steps.

The volume resolution is dramatically improved, however up near 0 dB the sound still suffers a little bit.

It is not until you move to 32 bit does the high level volume tend to be plesent to hear, near 0 dB.

32 bit has the capacity of 4.3 billion individual voltage volume steps between -196 dB & 0 dB.

So there is a dramatic improvement in the sound quality when moving up to a higher density bit depth. No dithering required.

I have a nice day, every day. I hope you do too.

Take a listen to George Strait - right or wrong, in high definition. Unbelievable You have to hear it for your self. What a good studio recording. We need more studio's around that sound that good.

Um...no. There is virtually zero difference in sound between a 16-bit file and a 16-bit file converted to a 24-bit file. You're just filling it with empty space. I get the feeling that you haven't actually done these conversions and listened to the files. They sound exactly the same.

Jun 20, 2015 at 7:46 PM
I have been told, you never loose anything by downsampling a file, only when you compress the bit depth. Also, 16 bit is enough depth to capture all of the music, at least the part you really want to hear below 10khz, easily storing up to 22khz with minimal apparent losses.

Bit depth doesn't have anything to do with the frequency response, unless you are talking about noise shaped dither, which you never mentioned.

Is this detail really worth saving? For the record, relating to the "Sampling Theorem", dynamic range decreased by 'log²(N). By removing 4 bits 256 times to form a 20 bit, & again removing another 4 bits 256 times to form a 16 bit.

By removing 4 bits you lose ~24dB of dynamic range. That is 1/16th the number of steps, I don't know where you're getting 256 from.

The bottom steps are very close, You take them on the run, hundreds of them at a time, the middle steps are normal size, the top steps, say -6 dB to -4 dB, you need a ladder to get up to the next step, by -2 dB you need a helicopter, to clime up to the 0 dB ceiling breaker, you'd need a rocket ship to make the last upper step.

I've no idea what you are trying to say here.

That is log²(N). 65,535 volume steps in 16 bit, 16.7 million volume steps in 24 bit. I think the count starts at -196 dB, so by -84 dB average noise floor, you have used most of your' aloted volume steps, to reach to the top of the file @ 0 dB

Where does -196dB come from? 16bit audio has a theoretical noise floor of -96dB, and 24bit is -144dB.

Going up in bit depth, can you dither? Well, yes. But you have already lost the sine resolution of the upper file frequency limit, say 22,050 to 24,000 frequencies become mostly triangle, & stay that way for the most part, after dithering from 16 bit to 24 bit.

Sine resolution? Frequencies become mostly triangle? What does that mean?

The Nyquest theorem: It is found in the 'Bell Laboratories, "Sampling Theorem,"' of Harry Nyquest & Claude Shannon, where the 'aliasing of any frequencies present in the capture are only reproduced if the sample rate is at least twice the frequency of the highest captured frequency.

If there is not enough samples in playback, the highest frequencies in the bit depth, all of the frequencies above half the sample rate, are folded in reverse frequency progression into the spectrum of sound heard by your' ear.

Ie: (at 44.1 sample 23khz is aliased at 21khz, 24khz is aliased at 20khz, for a file that started as a 48k16bit & downsampled to 44.1k16bit). These frequencies originally captured in the bit depth at 16 bit, are then reproduced into the upper limits of your' audio, as a form of musical related noise.

The aliasing sound can be heard as mostly a loud room echoing, or an increase in the "Live" sound of the recording studio.

Once the aliasing is mixed in with the rest of the signal it can't be separated. You can't resample the audio to the original sample rate and restore those frequencies that were aliased. Aliasing does not sound like "room echo" or "live sound."

IM Distortion in digital is called Aliasing. It degrades the reproduction quality, of the original stereo sound image produced by your' favorite speakers.

Inter-modular distortion and aliasing are not the same.

By upsampling alone, the aliases are erased by most software converters, if an alias filter is employed during upsampling. That is to say, you loose a small part of the resolution package hidden in the bit depth.

The Nyquest theorem: If the upsampling software does not apply aliasing filters, then the high frequency limit of the original recording can be reproduced by simply upsampling to at least twice the original sample capture frequency of the music file.

Under some conditions, I get a better result, if I upsample from 44.1 to 48 first, without the use of an aliasing filter during that upsample.

Then, I again upsample to 96k from 48k, this time using the aliasing filter to remove everything above 24khz, which is generally the original frequency bandwidth of most music.

Therefore, you don't want to add any aliases to the 96k upsample above 24khz original music data when upsampling to twice the original sample rate. Again, the Nyquest theorem..

When processed properly, you get a finished 96k sample file with data existing only below 24khz. All frequencies, including noise above 24khz, between 24khz & 48khz, is digitally suppressed down to around - 144dB or more

Does this all make a lot of money & cents, when it comes to selling or buying so called high res music files? When it is possible to recover that same file, plus or minus a possible noise factor formula of "SNR = sqrt(N)", from a lowly 16 bit mp3 file?

Again, if there is aliasing in the file, upsampling won't remove it.

Jun 20, 2015 at 8:19 PM
Thank you for the quote.

well, I suppose everyone has a right to their opinion, so the story goes. The story continues with the enemy of my enemy is my friend, but somehow I don't see you as an alley.

I accept the fact that there is going to be individuals who are color blind, as statistically there is one in 4 men & one in 10 women who can't tell the difference between the colors Red & Blue, the most common color blindness next to green/yellow blindness.

Off hand, I don't know the statistics of just how many individuals have tin ears, but you sound like a member of that elete group.

I want you to know, that I don't hold it against you. Your' not the first one on the block, to think a trumpet sounds like a bugle. And there are individuals that think a car backfiring sounds like a gun. Tell that to the inocent man & woman who were shot dead, some ungodly number of times by police, 49 times alone by one cop, why, because the cop thought the backfire was a gun. Even I can tell the difference between a gun and a car backfiring. The police officer surely either shot too many guns with out earmuffs, or was looking to kill somebody that day.

Which brings me back full circle to you. How did you loose your' hearing? Don't tell me, let me guess, LAUD 16 bit heavy metal machinery at work? That might explain why you can't split hairs between mp3 noise & real upsampled quality music.

I'll bet you thought 8 tracks & cassettes sounded the same? Well? Maybe that was a little before you lost your' hearing?

I know, your' car stereo is at fault, when you burned out some of the speakers at full volume.

Then again, your' hygene might be so bad, you have potatos growing in your' ears. Which reminds me? Seriously, when was the last time you cleaned your' ears?

There is one more possibility? You grew up listening to your' next door neighbor's stereo through tin cans & a string. Yeah, I hear you, it was a hard life growing up not knowing what real music should sound like.

You had better not buy anything like an HDPlayer, or heavens forbid, buy a SACD. They would be a total waste of money for you to do that. Come back anytime, after you regain your' hearing.

I have a nice day, every day, inspite of your' indifference to hundreds of millions of music lovers all around the world, upsampling their music files, purely because they think it helps, & you can quote me again, the deaf, dumb, blind, & the brain dead are all welcome here anytime. I won't hold it against ya'.

Hold that flight to Mars. Alchmist wants to get that HDPlayer on board, to use while he/she jogs. Alchmist was told the HDPlayer sounds better, but he doesn't believe it any more than I do.
Yeah, I told him the same files used in the HDPlayer is the same quality as upsampling a file can be. He didn't believe that either.........

...........

Jun 20, 2015 at 8:41 PM
HD audio is a gimmick. Like I have already explained, 16-bit / 44.1 kHz is the highest quality available. Everything else provides no audible benefit whatsoever. Some of the other formats only sound different because they came from a different master. When you convert the files yourself, it is physically impossible to distinguish between them. Read the article I linked you to.

Jun 20, 2015 at 9:32 PM
You are absolutely incorrect. 16-bit has more than enough dynamic range to handle all the recordings out there. 24-bit just adds more dynamic range, but there's literally nothing to add in terms of anything that would be audible. And 44.1 kHz is designed to play all the frequencies we can hear. Anything above that is inaudible. This is proven indisputable science here. 24-bit and higher is used in studios for computer processing purposes and has no relevancy to playback.

Here's an idea: since you insist that you can hear a difference between these things, would you be willing to prove it? I can facilitate this, if you so desire.

Jun 20, 2015 at 10:54 PM

Well, thank you for the reply.

I hope you don't take anything I said as an insult. But the quotes just might have worked for Saturday Night Live..... I bet even mom gave you the story of the potatos growing in your ears. That story is older than I am by far.

Your' point is noted. Somehow, I can't believe you really think one blank check fits all transactions.

Some transactions are all technical, while some transactions are purely experience.

Your' point that 44.1 16 bit can, & really does contain all that is needed for a good musical experience is only one transaction.

But, you are laying that burden at the door step of the music CD marketed for the first time 30 + years ago.

While the technology is the same unchanged, the personal preferences to the "all you need to hear is already contained in the 44.1 16 bit CD, has changed. I don't think any one would dispute that fact, at least on the surface of how the land lays.

Yes, files are captured every day @ 48k, 96k, 192k, 16bit, 24bit, 32bit, 48 bit, 64 bit, as well as some older dat machines that did 18 bit, 20 bit, 22 bit, et cetra. But the capture rates have nothing to do with editing at all.

I can do everything to any file at any bit, with standard studio editing software. The sample has nothing to do with ease or need to use 24 bit over 16 bit for editing. So where do you get the idea that somehow the studio sample rates are tied to editing the file?

And as far as the bit depth is concerned. While the 16 bit, as I already explained, has a very limited number of volume steps to support the normalization of the file to 0 dB. I also explained the point, that to limit this rough sounding music peaks above -6 dB, simply by normalizing peaks to below -6 dB. It is just that simple. Most recording studios set the normalization in several ways for reasons involving the average/peak rm². Some RM² levels average at -12 dB, others average -15 dB, still other files with little or no audio compression add to the audio can have averages as low as -24 dB. All of these files have peaks as well.

Based on the level set of the lead singer @ -6 dB peak, this file can have total music peaks of -3 dB. Other files, the singer is still held at -6 dB, but the audio compression allows for peaks no higher than -5 dB.

All of these level sets are based on audio compression, or the lack there of, & the fact that no matter what bit depth, the harsh normalizations in a 16 bit file, above -6 & getting progressively worse up to 0 dB first of all do not in any way damage or is considered any defect in the file capture at any bit rate.

Furthermore, if you tracked just one frequency for it's varying volume output in the file, because of the non linearity nature of the low sample count of 16 bit normalization levels, results in a log² expansion of volume of all frequencies as it approaches 0 dB, this volume expansion is the reason for the harsh sounding high volume in the file. 24 bit suffers far less non linear volume expansion purely because there is so many more incremental points of normalizing volume, (16.7 million of them), right up to 0 dB these digital volume steps remain tolarable to listen to.

This fact has nothing to do with the studio using 24 bit for editing purposes. You may have the experience if you have opened hundreds of thousands of 16, 24, 32 bit files, like I have opened over the years, just to find that generally speaking, the studio engineer knows the noise level is so low in a digital file compaired to say vinyl or tape, that just right off the bat they just set peaks down close to -6 dB. They do this not for any other reason than to limit the harsh volume at the peaks of normalization, when the file ultimately ends up on a 16 bit CD.

So with those facts on the line that I can prove is true, how about we stop talking about how 16 bit captures everything you need to hear in the music. While this is true, there are certain draw backs to playing 44.1 16 bit, & just excepting the sound as is, good or bad. Do you get my pount?

I said studio engineers know about the harsh high volume sound from a 16 bit file, but I didn't say just how many know that fact. Too many sound engineers got the job because they were related to the studio owner, having no idea these drawbacks in digital sound even exist. There are tons of files in this world that are normalized to 0 dB @ 16 bit. Someone made that file, & didn't do their homework first.

I could compare them to you in that respect. I bet you would fire those vu meters to +3 while recording on a DAT?

There is no head room in digital!

No, I don't think you would do that, as most DAT machines don't have vu's that register over 0 dB.

Now, I get your point about hearing everything in a 16 bit CD. So I ask you to read carefully what I am saying, in order to get my point, which is most if not all files that start at hi bit rates, first, retain everything that was contained in the first capture bit rate, after it is compressed in this case, from 48k to 44.1k, but again the not a defect of the high frequency aliasing, that can be cured by simple oversampling the file in play, & the volume expansion explained here, both of which can be improved by upsampling the 44.1 file back to it's original sample rate, usually 48k, as well as improving the volume expansion by murging from 16 bit to at least 24 bit, which requires no dithering.

So, if you want proof to the podding, I can help you with that also.

Don't continue to think what is good enough for you, should be good enough for the rest of the world.

I happen to have fantastic luck in upsampling files that have not been damaged by any ways or means, including other file converter software with various hidden routines in their core that damages inportant compression artifact data.

The files I hae the best luck with are raw, right off the hot pressed CD. MP3's are for the most part a waste of time, but over the years I have found some of them that I turned a tin can into battleship armor plate. Believe me, these files still sound really good througn almost any speaker system or earphone terminal.

I wold stack a few of these choise sample renderings against the original DAT recording, & I bet, this time you, yes you, couldn't tell the difference.

I'm having a really nice day, how about you?