If Bits are Bits

[Blumlein 88]

You'll need some damn good shielded cables for the electrons not to escape. Maybe then it's actually time to buy some of those audiophile ones.

Pure neutron shielding might help.

 

[raif71]

Some bits are actually bitter than the others

 

[kemmler3D]

can also call it histogram

No, that is not what a histogram is.

Anyway, showing constant flat values is correct in a sense, but misleading because it makes people think there's a stepped waveform being generated in the process somewhere, which there isn't.

 

[dorakeg]

I agree their presentation style is not optimal, for the sample value display they should have used pin plots (not connecting the dots) rather than bins one sample wide. But their meaning is explained fully correct:

"This diagram shows how an analogue sound wave can be represented with 24-bit 176,400 samples per second PCM encoding." (Underline mine).

That's what sample values do, they represent the voltage at the (infinitely short, theoretically) point in time the snapshot was made.
If you were to put out this value constantly for one sample period you have a NOS DAC "reconstruction" filter which is what you comment is probably about, but I don't see that in the context of the graphs and associated text. The sample stream is meant to be read as the abstract data it represents, not a time-domain actual output waveform of the DAC. Those dCS guys are serious engineers...

I would say they are all correct. Its just different ways of representation of values.

For dots, its real time data but not showing last known value. So, when the sample is taken at the start of the 22.7us block, you get a dot, after that there is nothing so no result till the next block...

For historgram, its showing last know value.So, you get a straight line instead of a dot.

 

[Blumlein 88]

Do we need to show the Monty video some more. Yes, I think we do.

 

[Rednaxela]

Can’t show Monty often enough.

 

[solderdude]

The ring DAC in essence is just a differently executed DS DAC with a really nice reconstruction filter.
It has the same linearity (low distortion) as DS DACs but less HF crap that needs less post filtering.
Seems like a well constructed DAC that, of course, does not have stairsteps and the inbetween sample values are just calculated values.

Every oversampling DAC and DS DAC does a similar thing... calculate 'sample values' in between the original values.

I was googling about more information regarding ring face and found this which I feel is very good explanation.

It does not explain what the OP found so is not a good explanation for that. It merely explains how the RING DAC (a form of DS DAC) works and some general info.

 

[dorakeg]

It does not explain what the OP found so is not a good explanation for that. It merely explains how the RING DAC (a form of DS DAC) works and some general info.

Part 5 of it did explain that it actually makes no difference whether source is from internet or USB. It did not specifically mention CD but it should not make any difference.

I am not sure if the streamer, CD player and DSD (not sure streaming or SACD) all is the same hardware. OP did not mention what is better/worse as well.

 

[Galliardist]

Part 5 of it did explain that it actually makes no difference whether source is from internet or USB. It did not specifically mention CD but it should not make any difference.

I am not sure if the streamer, CD player and DSD (not sure streaming or SACD) all is the same hardware. OP did not mention what is better/worse as well.

?

We hear from folks that bits are bits. Why does a CD redbook player sound better than a Streamer at 44.1/16, or even DSD.

and

Sorry folks, my first post. I had a 20+ year old Micromega Stage 6, Now a Bryston BCD-3, First DAC was a Halide HD, was better than Apple headphone jack, then upgraded to lindemann limetree network streamer. i prefer my Dacs to be in the same box as the CD/network device. I am just looking for reasons why the CD player is bang on, and the streamer/dac does not have the same great sound. Are they not both 44.1/16?? There are a few youtubers that love CD's over Streaming for the sound. There has to be a reason. Oh i am using Tidal 44.1 service. Maybe someone can science an answer. Hint intended Amir.

The OP told us pretty much everything.

And since you are having problems with understanding histograms, here's the WIkipedia definition (yes, I know...)

Histogram - Wikipedia

en.wikipedia.org

Clearly, the step diagram isn't a histogram. The video posted above tells you what it represents. I'm not going to put my own foot in it this time, go watch it.

 

[Takaya]

Sorry Folks, i am late to attend to page 8, I am hoping to acquire soon a bryston bp 17 with the DAC included, then a streamer with digital only outputs. Then i can compare the CD player and a streamer using the dac input of the bp17. Other than Lindemann Limetree Bridge II streamer, what do you folks recommend.

 

[antcollinet]

They are just plotting sample value over time using bar chart (can also call it histogram). So, the presentation is correct. It is supposed to be a straight line because the value does not change during that period.

If you were to use an equipment/computer to read the output values, and show them on a histogram, that would be exactly what you will get. Values will only change during beginning of each ~22.7 block, then remain constant throughout the block period.

But it is misleading in terms of what you get out of a (decent) dac. You don't get a stair stepped waveform, you get a perfectly smooth reconstructed wavfform, with the quantisation turned into a hiss like noise floor at -(n x 6)dB where n is the number of bits. (-96dB, and basically inaudible for redbook, and "gnat farting on the other side of the planet" levels for 24bit)

 

[dorakeg]

But it is misleading in terms of what you get out of a (decent) dac. You don't get a stair stepped waveform, you get a perfectly smooth reconstructed wavfform, with the quantisation turned into a hiss like noise floor at -(n x 6)dB where n is the number of bits. (-96dB, and basically inaudible for redbook, and "gnat farting on the other side of the planet" levels for 24bit)

Frankly speaking, I am not sure on this. I am sure there are ways to smooth out this steps. But how well they are done, I don't know.

However, I don't think it will be perfectly smooth, it will be very smooth, but as we zoom in on the wave, it will show tiny steps or variations like zagged lines.

 

[Blumlein 88]

Frankly speaking, I am not sure on this. I am sure there are ways to smooth out this steps. But how well they are done, I don't know.

However, I don't think it will be perfectly smooth, it will be very smooth, but as we zoom in on the wave, it will show tiny steps or variations like zagged lines.

Time to Cue up Monty again. Watch this. He will monitor the output using some of the very finest analog scopes ever built. You can see if there are "zagged lines." Stair steps are covered in the first few minutes. Don't stop there however. Very much worth watching the whole video. I posted it just after one of your earlier comments a couple weeks back. Really do watch it this time.

 

[antcollinet]

Frankly speaking, I am not sure on this. I am sure there are ways to smooth out this steps. But how well they are done, I don't know.

However, I don't think it will be perfectly smooth, it will be very smooth, but as we zoom in on the wave, it will show tiny steps or variations like zagged lines.

It really doesn't - watch the video.

 

[Cbdb2]

If 24bits has extra resolution, as in extra steps between samples, the ever present noise added to those steps probably makes them useless. Most recordings ( plus the repro electronics) have more noise than 16 bits so little point going greater bit depth.

 

[voodooless]

They are just plotting sample value over time using bar chart (can also call it histogram). So, the presentation is correct. It is supposed to be a straight line because the value does not change during that period.

As per Monty video: this is totally incorrect. The representation is wrong. The samples value is only correct for the exact moment in time it was measured at. The sample values before or after that sample moment are unknown. Therefore we cannot use staircases to represent digital audio samples.

Nor are these staircases generated by most DACs because the most successful designs are delta sigma DACs, that use oversampling and a single or multibit output that is filtered to create a smooth result. Some older DAC chips or the more obscure ladder DACs may actually output something resembling staircases, specifically in NOS mode, but generally these designs aren’t very good performing.

 

[RayDunzl]

Frankly speaking, I am not sure on this. I am sure there are ways to smooth out this steps. But how well they are done, I don't know.

However, I don't think it will be perfectly smooth, it will be very smooth, but as we zoom in on the wave, it will show tiny steps or variations like zagged lines.

To visualize the smallest steps, should they not be filtered away:

For 16 bits, stack 65,536 sheets of paper on your desk.

The middle of the stack would represent 0 volts, the top and bottom pages would be (typically) + and - 2 volts.

The individual steps would equal one sheet of paper.

If your DAC resamples (probably) to 24 bits, expand the stack to 16,777,216 pages.

---

A ream of 500 pages of copy paper is about two inches thick.

So, the 16 bit stack might be about 22 feet tall, and the 24 bit stack would be around 5,600 feet tall.

32 bits takes it to 271 miles.

 

[krabapple]

We hear from folks that bits are bits. Why does a CD redbook player sound better than a Streamer at 44.1/16, or even DSD.

GO DIRECTLY TO JAIL.
DO NOT PASS GO.
DO NOT COLLECT $200.

 

[Tangband]

If 24bits has extra resolution, as in extra steps between samples, the ever present noise added to those steps probably makes them useless. Most recordings ( plus the repro electronics) have more noise than 16 bits so little point going greater bit depth.

The extra bits are needed at the digital recording and digital mixing and mastering level . Different plugins like eq, peq, compression, exiters, reverb etc is eating resolution in the digital domain. You need the extra bits in the producer chain to have a chance to get the 16 bit resolution in the end . Its true though that microphones have much more noise than -144 dB.

 

[KSTR]

As per Monty video: this is totally incorrect. The representation is wrong. The samples value is only correct for the exact moment in time it was measured at. The sample values before or after that sample moment are unknown. Therefore we cannot use staircases to represent digital audio samples.

No, it is totally correct, albeit a bit misleading in representation. It's you guys are triggered by the staircase plot, blindly assuming without further checking (simple reading would have done the job) that the authors are claiming that this is were the actual waveform... which they didn't.

Quoting myself:
"I agree their presentation style is not optimal, for the sample value display they should have used pin plots (not connecting the dots) rather than bins one sample wide. But their meaning is explained fully correct:

"This diagram shows how an analogue sound wave can be represented with 24-bit 176,400 samples per second PCM encoding." (Underline mine).

That's what sample values do, they represent the voltage at the (infinitely short, theoretically) point in time the snapshot was made."

EDIT: As for the "unknown" thing, that isn't correct either, the analog waveform between sample points is known when proper sinc() filtering was used in the analog-to-digital conversion and thus can faithfully be reconstructed within the system bandwidth.

EDIT2: "Therefore we cannot use staircases to represent digital audio samples." Again: the staircase plot is a perfectly valid graphic representation of a digital sample stream. It does not represent the corresponding proper analog waveform, though.