Tech V&R Discussion

[Renate]

Although I've done 6802, 6805, 68008, 68020, ATmega328, ATmega32u4, SAMD21, I just bought my first "Blue Pill", STM32F103C8T6 for $2+.
That chip's design is already 20 years old!
I just got a blinky going using my own (from zero) ld file, startup (vectors, clock, segments), blinky and CMSIS-DAP programmer software (including RAM stub).
It was mostly curiosity. I won't waste more time on this. I'll look into the STM32F4--.

And since this is the rant channel, I'll add:
In the stone age, microprocessors just started up instantly.
Now some boards PLL off 32 kHz xtals and can take up to 2 seconds before they blink.
Give me a high speed xtal, I can stand the couple of milliamperes.

[jbjb]

Originally Posted by 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.

The use of 24 bits doesn't mean you're just encoding louder sounds - it can be used to capture the same sounds at higher resolution, leading to lower levels of quantisation noise.

Let's say you're encoding a signal which has a maximum amplitude of 1V peak-to-peak (in the range of -0.5V to 0.5V). CD 16-bit audio gives a range of values from -32768 to 32767 (8000 to 7FFF in hex). If we're talking about signals which are balanced around zero, which we usually are for audio, we'll ignore the extra negative value at -32768 and represent -0.5V with -32767 (8001 hex) and 0.5V with 32767 (7FFF hex). So we have 65535 possible sample values, and are splitting the 1V peak to peak into 65534 steps. This means that the maximum error (quantisation noise) on each sample will be +/-(1/(2*65534))V, or about +/-7.6uV (ignoring dithering)

With 24-bit audio in this scenario, we'd encode -0.5V as 800001 hex and 0.5V as 7FFFFF, dividing the 1V into 16777214 steps and leading to an error range of +/- 0.03uV - a quantisation noise floor 256 times better in terms of voltage (65536 times better in terms of power as power goes with the square of voltage).

So, the question isn'i whether or not 24 bits can encode audio more accurately, with much better SNR and much better dynamic range (it definitely can), it's whether or not this improvement is actually detectable when using real-world amplifiers and speakers etc., and when listened to by human ears. I've yet to see any convincing tests which indicate that people can tell the difference, so 24-bit seems pointless as a distribution format.

It does still make sense to do manipulation and mixing at 24-bit, however, to stop the errors at each stage accumulating.

[JSWolf]

One thing that can make 24-bit/95KHz better then 16-Bit/44.1KHz are the chips that decode the digital signal do a better job at the higher values.

[nana77]

Originally Posted by JSWolf

One thing that can make 24-bit/95KHz better then 16-Bit/44.1KHz are the chips that decode the digital signal do a better job at the higher values.

Which, on a based mobile-world, lands to a choice of devices were very few phones have been build within those specs (LG V20, thought out of flag as updates iirc).

Spoiler:

https://en.wikipedia.org/wiki/LG_V20

Maybe there is a Fiio phone also, or very few others... also thought that that lack did build a market of DAC devices that costs like jewelries, and (rant note) that lately become more Android based, for the most.

Thanks for all the info you wrote, really interesting !
Ps: thought that all of those DAC's cable adapters (like iBasso DC05) might allow only more volume, if the file isn't previously encoded, yet from the source, within those specs' reqs.
I'm glad there are still manufacturers that allows a simple radio FM on a phone, despite not always that function is available within the default settings.

[ratinox]

Originally Posted by JSWolf

One thing that can make 24-bit/95KHz better then 16-Bit/44.1KHz are the chips that decode the digital signal do a better job at the higher values.

This is a misunderstanding of an aspect of oversampling, a misunderstanding that snake oil grifters have capitalized on heavily. Have you ever seen a CD player with a label that says something like "96KHz Oversampling"?

When you reconstruct an analog signal you can get aliasing artifacts, high frequency noise, in the output stage. In principle, analog filters remove this noise, but these filters are relatively expensive compared to the rest of the system. What 96KHz oversampling does is tell the DAC stage to behave as if the sample rate is 96KHz instead of 44.1KHz. This pushes aliasing artifacts very far above the limits of human hearing. Then a low-pass filter at 22KHz cuts off all the artifacts without needing costly analog filters. Simple, effective, and much cheaper to manufacture than costly analog noise filters.

[ratinox]

Originally Posted by jbjb

So, the question isn'i whether or not 24 bits can encode audio more accurately, with much better SNR and much better dynamic range (it definitely can), it's whether or not this improvement is actually detectable when using real-world amplifiers and speakers etc., and when listened to by human ears. I've yet to see any convincing tests which indicate that people can tell the difference, so 24-bit seems pointless as a distribution format.

Absolutely. The extra headroom is invaluable in the studio. It's pointless for consumer media.

[jbjb]

Originally Posted by ratinox

Absolutely. The extra headroom is invaluable in the studio. It's pointless for consumer media.

It's extra *resolution* that 24-bit provides, not extra headroom.

The headroom is defined by the range of voltages the ADC supports, usually set by the reference voltage provided to the ADC, within limits dependent on its design. The bit depth defines how many steps that input range is divided into, i.e. the resolution within that range.

Admittedly, more bit depth can allow you to have a bigger input range and maintain acceptable resolution, but it's the resolution that the bit depth controls, not the input range directly.

[ratinox]

Originally Posted by jbjb

It's extra *resolution* that 24-bit provides, not extra headroom.

The headroom is defined by the range of voltages the ADC supports, usually set by the reference voltage provided to the ADC, within limits dependent on its design. The bit depth defines how many steps that input range is divided into, i.e. the resolution within that range.

Got it. Apologies for the misusages. I know enough about this stuff to be dangerous and maybe a bit more but it isn't my expertise.

[JSWolf]

Originally Posted by ratinox

This is a misunderstanding of an aspect of oversampling, a misunderstanding that snake oil grifters have capitalized on heavily. Have you ever seen a CD player with a label that says something like "96KHz Oversampling"?

When you reconstruct an analog signal you can get aliasing artifacts, high frequency noise, in the output stage. In principle, analog filters remove this noise, but these filters are relatively expensive compared to the rest of the system. What 96KHz oversampling does is tell the DAC stage to behave as if the sample rate is 96KHz instead of 44.1KHz. This pushes aliasing artifacts very far above the limits of human hearing. Then a low-pass filter at 22KHz cuts off all the artifacts without needing costly analog filters. Simple, effective, and much cheaper to manufacture than costly analog noise filters.

But I'm not talking of oversampling. I'm talking of a native 24-Bit/96KHz.

[ratinox]

Originally Posted by JSWolf

But I'm not talking of oversampling. I'm talking of a native 24-Bit/96KHz.

Yes, you are, even if you don't realize it. For a listener, the only thing a 96KHz source does is make the DAC's oversampling to 96KHz step a no-op. Your 96KHz files are, effectively, pre-oversampled. Everything else is the same. You get aliasing artifacts in the reconstruction and the low-pass filter to cut off these artifacts whether the source is 44.1KHz and oversampled by the DAC or it's 96KHz pre-oversampled.

[Quoth]

Originally Posted by JSWolf

But I'm not talking of oversampling. I'm talking of a native 24-Bit/96KHz.

It's fantasy unless you are a guinea pig, cat or some other non-human and the program content is non-human and everything was at least 24 bit @ 96 kHz.

I've been doing digital recording since about 1993. Worked in AV & Telecom R&D. Did BBC courses.

Nothing for humans needs more than Red Book CD Audio. There is some value in initial sampling at 96 kHz (or even 192 kHz) to make analogue audio filters better/simpler. But you can immediately reduce to 44.1 kHz or 48 kHz. There might be some internal value in converting 16 bit to 24 bit (or even 32 bits) before mixing multitrack, but then 16 bits afterwards.

Distribution doesn't need more than 16bit 44.1 kHz or 48 kHz. Playback might use 96 or 192 kHz conversion simply to make the analogue filter cheaper.

[JSWolf]

Originally Posted by Quoth

Did BBC courses.

That's the problem right there. BBC have no clue what they are doing in terms of sound. Some of their programs are terrible. Muffled voices and/or effects/music too loud while the voices whisper or are not loud enough. And iPlayer is a disaster sometimes. 128k stereo and some of the downmixes from 5.1 to stereo are very poorly done.

They can't even get Glastonbury to sound good. I've heard some really bad sound from Glastonbury.

[BetterRed]

↑ ↑ ↑ ✔

Radio programs where the presenter might be heard to say - "As you can see."
Cheapstake simulated interviews over Whatsapp - or is it Wart's Up.
As for Glastonbury - has beens, and never should of beens, prancing around for the good and great.

[Renate]

Originally Posted by JSWolf

BBC have no clue what they are doing in terms of sound.

On BBC World Service they could benefit from using VU meters.
It's impossible to set your device to a volume that is compatible with an entire newscast.
The inserts are often louder than the presenter.

[JSWolf]

Originally Posted by Renate

On BBC World Service they could benefit from using VU meters.
It's impossible to set your device to a volume that is compatible with an entire newscast.
The inserts are often louder than the presenter.

Usually it's the other way round where the inserts are usually not as loud as the presenters.

The thing is, when I watch some BBC programs on Netflix such as some David Attenborough documentaries, the sound is actually good. It's 4K/Dolby Vision/Dolby Atmos.

I just put on Our Planet S01E01. David's voice is clear and the music/effects are not too loud. The Dolby Atmos mix is good. BBC mucks it up so the voice is not as prominent or as clear.

So taking a course on sound from BBC is going to get you very messed up. You won't know how to do sound mixes properly.