Tech V&R Discussion

[Quoth]

Current DSP based radios (maybe phones & WiFi now) deliberately have RF input many times the sample rate. The input bandwidth at RF then must be less than half the actual sample rate* and then the ADC also acts like a mixer due to the aliasing. The reason to do this because it's easier to have more bits (=more dynamic range) at lower sample rates.

[* With I Q inputs or multiple ADCs the actual sample rate can be twice to 16x the clock rate]

[jbjb]

Originally Posted by Quoth

Sampling at many times the final sample rate allows for a simpler analogue pre-filter as there must be no input above 1/2 the sample frequency. This also allows more content up to the sample frequency as an analogue filter can't be like a cliff. It's not a sharp cut off. Then the input can be digitally filtered to 1/2 the final sample rate.

Most recording studios work this way - much higher sampling rate for recording and mixing, then a decimation filter to reduce the sampling rate.

Because offline digital filters don't have the be causal, it's easier to get a sharp cutoff without nasty phase distortion.

[ratinox]

Originally Posted by jbjb

Most recording studios work this way - much higher sampling rate for recording and mixing, then a decimation filter to reduce the sampling rate.

Not really "much" that I'm aware of: 24 bits at 48KHz. Provides plenty of room to accommodate any mistakes or errors in the original recordings. Unless a customer insists on something absurd like 32-bit/96KHz or higher -- because "jargon hi-res audio jargon jargon." But maybe that's more rant than discussion.

[jbjb]

Originally Posted by ratinox

Not really "much" that I'm aware of: 24 bits at 48KHz. Provides plenty of room to accommodate any mistakes or errors in the original recordings. Unless a customer insists on something absurd like 32-bit/96KHz or higher -- because "jargon hi-res audio jargon jargon." But maybe that's more rant than discussion.

You're right in that most mixing seems to be at 48kHz. From what I've seen, though, the input (recording) stage would often sample at a higher rate (with a correspondingly easier analog anti-aliasing filter with a gentle roll-off, and then a non-causal digital decimation filter with a sharper roll-off would produce the final 48kHz "recorded" signal.

32-bit/96kHz is clearly pointless for the final product, but if you're doing a lot of manipulation and processing to produce that, doing that processing at higher resolution stops rounding errors from accumulating.

[JSWolf]

Originally Posted by jbjb

You're right in that most mixing seems to be at 48kHz. From what I've seen, though, the input (recording) stage would often sample at a higher rate (with a correspondingly easier analog anti-aliasing filter with a gentle roll-off, and then a non-causal digital decimation filter with a sharper roll-off would produce the final 48kHz "recorded" signal.

32-bit/96kHz is clearly pointless for the final product, but if you're doing a lot of manipulation and processing to produce that, doing that processing at higher resolution stops rounding errors from accumulating.

24-bit/96KHz is the best format for digital audio. It sounds better then 16-bit/44.1KHz.

[ratinox]

Originally Posted by JSWolf

24-bit/96KHz is the best format for digital audio. It sounds better then 16-bit/44.1KHz.

This is a well-debunked myth. Watch Monty's video. That's how sampling and reconstruction actually work.

16 bits is 96dB of dynamic range; 24 bits is 144dB of dynamic range. Sustained exposure to sound levels above about 80dB will cause permanent hearing loss, so if you are listening to music that loud? You won't be listening to anything in short order. The only "practical" difference for consumers is how much faster you'll deafen yourself listening to 120+dB sounds.

44.1Khz sample rate is 22.05KHz of analog signal. Human hearing tops off at about 20KHz. You cannot hear anything above this. No, you can't. You're a human, not a cat or a dog. If you've been listening to "too loud" sounds for most of your life then you probably can't hear anything above around 14-15KHz, if not less. See above.

Recording studios use 24-bit/48KHz and higher for working with audio tracks because these provide more "room" to work in, like avoiding clipping because the sound level on a track is set too high.

All digital audio players including so-called "hi-res" players use a low-pass filter at between 20KHz and 25KHz. If they didn't then your headphones and speakers would damage themselves trying to produce high frequency aliasing noise coming out of the DAC stage. Even if you could hear sounds well above 20KHz, which you can't, they would be cut off by the low-pass filter.

If a 24/96 file of a song sounds better than a 16/44.1 file of the same song then it's because the 24/96 file was made from a cleaner mix or master or both.

[JSWolf]

Originally Posted by ratinox

This is a well-debunked myth. Watch Monty's video. That's how sampling and reconstruction actually work.

16 bits is 96dB of dynamic range; 24 bits is 144dB of dynamic range. Sustained exposure to sound levels above about 80dB will cause permanent hearing loss, so if you are listening to music that loud? You won't be listening to anything in short order. The only "practical" difference for consumers is how much faster you'll deafen yourself listening to 120+dB sounds.

44.1Khz sample rate is 22.05KHz of analog signal. Human hearing tops off at about 20KHz. You cannot hear anything above this. No, you can't. You're a human, not a cat or a dog. If you've been listening to "too loud" sounds for most of your life then you probably can't hear anything above around 14-15KHz, if not less. See above.

Recording studios use 24-bit/48KHz and higher for working with audio tracks because these provide more "room" to work in, like avoiding clipping because the sound level on a track is set too high.

All digital audio players including so-called "hi-res" players use a low-pass filter at between 20KHz and 25KHz. If they didn't then your headphones and speakers would damage themselves trying to produce high frequency aliasing noise coming out of the DAC stage. Even if you could hear sounds well above 20KHz, which you can't, they would be cut off by the low-pass filter.

If a 24/96 file of a song sounds better than a 16/44.1 file of the same song then it's because the 24/96 file was made from a cleaner mix or master or both.

[DNSB]

Real world intruding but a few years back, one friend of mine got into SACDs, HDCDs, 24bit/96KHz, 24bit/192KHz and quite a few other varieties of snake oil. He could easily afford the expense but in testing, none of us were able to tell a good CD from any of the "advanced" format CDs when listening over speakers and very few were able to agree on which format sounded better when using his headphone collection.

About 6 months back, he held a party and we tried the various tests from Think You Have Golden Ears? Take These 10 Audio Challenges and Find Out. It seemed none of us had the "golden ears" required to hear the differences in the real world. Perhaps a few decades back but our old ears are not what they used to be.

[ratinox]

"More detail"

That's not what bit depth is.

A sample is a value at a point in time. It is a number from 0 to whatever the maximum bit size which for 16 bits is 65,535. As previously explained, 16 bits covers a dynamic range of 96dB. If the sample is within 0dB and 96dB then this number is going to be between 0 and 65,535 (64K) with 64K being the loudest possible sound.

What happens if you increase this to 24 bits, for 144dB at the top? This number is going to be between 0 and 16,777,215, right?

Here's why that video is nonsense: 65,535 is the same value whether it's represented with 16 bits or with 24 bits. If the dynamic range of the input is within that 96dB "window", which it usually will be because human hearing, then you will get the EXACT SAME SAMPLES whether you use 16 bits or 24 bits.

That's the simplified explanation. You need to dig into dithering, oversampling, noise shaping, and how they affect signal-to-noise ratio. But the short version is that "more detail" and "less noise" from 24-bit files is snake oil. They don't sound better than 16-bit files unless they use superior quality masters.

Funnily enough, you'll hear the same kind of difference when comparing a 1990s CD of a 1970s album and a contemporary CD made from a contemporary remaster of that same album. The original recordings and mixes don't change, and the CD specs don't change, so what's different?

The master.

[ratinox]

Originally Posted by DNSB

Real world intruding but a few years back, one friend of mine got into SACDs, HDCDs, 24bit/96KHz, 24bit/192KHz and quite a few other varieties of snake oil. He could easily afford the expense but in testing, none of us were able to tell a good CD from any of the "advanced" format CDs when listening over speakers and very few were able to agree on which format sounded better when using his headphone collection.

I got two more links to share.

First is a recording of Ethan Winer's Audio Myths Workshop. Has some great information.
https://www.youtube.com/watch?v=BYTlN6wjcvQ

Second is a recent article from Tom's Hardware in which audiophiles couldn't tell the difference between music played through copper wire, a banana, or a pan of mud.
https://www.tomshardware.com/speaker...riment-creator

[DNSB]

Reminds me of one time in the 1980s when one audiophile doubter ran a test where a rotary switch was used to select between a high end solid state amplifier, a McIntosh tube amplifier, a Carver magnetic amplifier (actually not a true magnetic amplifier but used unusual power supply designs) and a mid range home quality receiver all hooked up to a high end CD player and pre-amp with the amplifier output feeding a bank of headphones.

Oddly, most of the golden ears crowd could "hear" and commented on the differences in sound between the four amplifiers. The humorous part was after the testing was done and the comments were made, it was revealed that the switch did nothing since all 4 throws were connected to one amplifier. You did get a momentary interruption as the switch was turned but that was all that happened.

Anyone else seen a CD Stop Light marker pen for a mere $24.99? You ran it around the edge of the CD to reduce reflections and improve the sound. An appeal to the crowd that couldn't tell the difference between digital and analog.

See The Greening of SACDs, (gulp) CDs and other digital madness if you don't mind the insanity.

[ratinox]

Originally Posted by DNSB

Reminds me of one time in the 1980s when one audiophile doubter ran a test where a rotary switch was used to select between a high end solid state amplifier, a McIntosh tube amplifier, a Carver magnetic amplifier (actually not a true magnetic amplifier but used unusual power supply designs) and a mid range home quality receiver all hooked up to a high end CD player and pre-amp with the amplifier output feeding a bank of headphones.

The Audio Myths Workshop has what I believe to be the original version. Starts about 3 minutes in. I'll let you watch it and not spoil the punch line.

Anyone else seen a CD Stop Light marker pen for a mere $24.99? You ran it around the edge of the CD to reduce reflections and improve the sound. An appeal to the crowd that couldn't tell the difference between digital and analog.

Oh, ghu, yes. That was so stupid, though not as stupid as people who spend a thousand dollars or more on a 1 meter Ethernet cable.

There WAS a marker thing that did actually do something. One of Sony's attempts at copy protecting audio CDs was to add a "security" track to the discs that would in principle prevent them from being played on computers. In practice it caused a lot of mechanical problems especially with Macintosh computers in which these discs would get permanently stuck and require disassembly of the drives to remove the discs.

The "crack" was to draw over the security track with a marker so that it couldn't be read.

[DNSB]

Originally Posted by ratinox

The Audio Myths Workshop has what I believe to be the original version. Starts about 3 minutes in. I'll let you watch it and not spoil the punch line.

Ah, I remember the Plastic Tiger. Etched the circuit board and built it while I was at university and used it for close to 2 decades. Then I built a color organ (think it was from Radio-Electronics) with handmade pulse transformers and I had some fun.

[Renate]

Originally Posted by DNSB

Ah, I remember the Plastic Tiger. Etched the circuit board and built it while I was at university and used it for close to 2 decades. Then I built a color organ (think it was from Radio-Electronics) with handmade pulse transformers and I had some fun.

Are you talking SWTPC?
In ~1978 I had a SWTPC Tigersaurus 250 (Watt) #210.
I had a fun generator, DVM, freq meter, RTL Microlab, Numitron...

[DNSB]

Originally Posted by Renate

Are you talking SWTPC?
In ~1978 I had a SWTPC Tigersaurus 250 (Watt) #210.
I had a fun generator, DVM, freq meter, RTL Microlab, Numitron...

Yes, that's them. The Plastic Tiger was the baby in their amplifier family with a 30W output. I built quite a few electronic projects from various magazines and companies back then. I used to dislike some of the British projects since I spent too much time locating the North American equivalents to their parts but in some ways, that was part of the fun.

The last item I owned from SWTPC was their 6800 computer though I sold it shortly after since I was using 8080/Z-80 chips and CP/M a lot more.