Digital signals

[krabapple]

Yes and No. It IS true that the more samples we capture and the higher the bit depth used, the more accurate the digital copy will be.

No, that is NOT what is true-- more samples/ bits *don't* get you 'more accuracy' beyond a certain point...and that is the wonderousl, and rather counterintuitive, concept that is demonstrated by Shannon-Nyquist.

Please read Mr. Lavry's paper, that WmAx pointed to earlier. Digital audio and information theory do not always conform to 'common sense' -- some of its truths are counterintuitive.

 

[krabapple]

My manager is an audio engineer who has been in the audio field for 30 years. He has a boatload of certificates as well. Now we had a discussion a few days ago and he tells me that digital simply loses information whereas analog preserves that information. You get everything on the disk and you hear everything on the disk, as it were.

I bet he's one of those many 'audio engineers' who's failed to keep up with or fully comprehend the technical side of digital. *By far* the greatest 'loss of information' occurs during the electromechanical transduction steps of a recording: at the microphone, and at the speakers. Analog tape further loses information, substituting noise. Ditto stamping an LP, and playing it back. The 'loss' via digital is anything but simple and need not even be audible ('losing' frequencies over 25 kHz is highly unlikely to result in any audible artefact).

(The only technically 'super-savvy' audio engineer I've seen claim otherwise is Bob Katz...and I'm currently trying to find out from him what he *really* means by calling A/D and D/A 'lossy')

He also said that oversampling can not reproduce the signal without having problems of it's own. He was directing pot shots at oversampling and supersampling techniques. Now he is quite old, and he will definitely not post on this forum. I don't think he has ever posted on a forum ever in his life. I want to learn and understand this so that I can rebut his points correctly and so that I can understand this subject for myself.

References are available out there. Try Nika Aldrich's book, 'Digital Audio Explained for the Recording Engineer'. It starts all the way at the beginning, with basic acoustics and physiology of the ear, and progresses through all
of the importnat stuff, up to and including DSD.

 

[avnetguy]

No, that is NOT what is true-- more samples/ bits *don't* get you 'more accuracy' beyond a certain point...and that is the wonderousl, and rather counterintuitive, concept that is demonstrated by Shannon-Nyquist.

It certainly does if you are looking at it on a computer screen!

Steve

 

[MDS]

No, that is NOT what is true-- more samples/ bits *don't* get you 'more accuracy' beyond a certain point...and that is the wonderousl, and rather counterintuitive, concept that is demonstrated by Shannon-Nyquist.

The concept demonstrated by Nyquist that you are alluding to is sampling at twice the highest frequency, but saying you don't get any more accuracy with more bits or more samples is not logical and is not true. The premise behind that argument is that we cannot hear beyond 20 kHz so why bother sampling faster than 44.1 kHz and capturing the higher frequencies we cannot hear? I'm not convinced that the overtones present in some music that are beyond our hearing ability contribute in some way to our perception of the sounds we do hear, but nonetheless sampling faster captures them. So by that definition higher sample rates are more accurate and if you are in the same camp as Mr. Katz, you want those higher rates because you believe they do contribute positively to the result.

More bits equals more accuracy. I'm sure someone will throw out the rule of thumb that every extra bit gives you 6 dB of dynamic range. The reason for that is that you have a more accurate calculation of the amplitude at each point in time. If you measure with a ruler marked in inches, you are accurate to one inch - if the actual distance is 6.5 inches, you either choose the distance as 6 or 7 inches and are always off +/- .5 inches. If the sample value falls between two integral values, you have the same issue. Again, does it matter? Some people think so, I am not that golden eared.

The Lavry article simply states the theory and then shows the implementation issues with the theory and their particular way of working around the problems. Basically, 'we do it perfectly so buy our stuff'.

 

[jonnythan]

More bits equals more accuracy. I'm sure someone will throw out the rule of thumb that every extra bit gives you 6 dB of dynamic range.

Don't confuse *number of bits* with *number of samples*. A CD has 44,100 samples per second (per channel), and each sample is 16 bits. Increasing the *bits per sample* yields higher bitrates, but the same number of samples per second.

 

[MDS]

I'm not confused and I don't see where I said anything that would lead one to believe I was confusing the two.

 

[avnetguy]

The term that relates to accuracy is quantization error, the more bits you have (providing the A/D chip LSB error is the same) the less quantization error you will have.

Steve

 

[WmAx]

The term that relates to accuracy is quantization error, the more bits you have (providing the A/D chip LSB error is the same) the less quantization error you will have.

Steve

This translates directly to one thing: noise floor. So far as quantization: A properly dithered signal will avoid any detrimental quantization distortion(s). Any signal not dithered for a 16 bit system can not be considered to be proper.

-Chris

 

[WmAx]

WmAx,
I did read the article but it doesn't touch on the amplitude issue I mentioned. While his illustrated examples are nice, an actual DSO capture showing the amplitude in and out of the sine wave would tell the story perfectly. I wish I still had the equipment around to test this myself but I got out of that game many years ago, maybe I can figure something out with the hardware I have laying around now. Again, let me state that I don't believe this is really needed as our ears probably can't detect the small amplitude differences at those frequencies, I just don't see it being a so called "perfect" reconstruction if one were to view the output on a scope.

Steve

The article did explain, in terms of the sinc function and the anti-alias filter(which is the applied equivalent), along with mathetmatically based plots showing the reconstruction. Re-construction includes restoration of amplitude. Note the graph I posted earlier. Adobe Audition used a similar function to reconstruct the waveform(including the correct amplitude) from the sampled data points; a perfect illustration of reconstruction. Note the graph I provided is a display of of a 20kHz sine wave that I recorded with the sound card(not a signal generated internally with Audition). The sound card/software only had the limited sample points and amplitude reference of those points, but it was able to perfectly reconstruct the signal.

-Chris

 

[Vaughan Odendaa]

One argument against all of this, according to my manager, is that their really isn't any reference point to judge accurate reproduction of the signal because there is no way of knowing what the recording sounded like in the first place.

On a unrelated point, he always tells me that you shouldn't use hifi speakers for hometheater applications because the digital signals are really, really quick and these effects will not be reproduced properly on hifi speakers.

What do you guys think ?

--Sincerely,

 

[MDS]

One argument against all of this, according to my manager, is that their really isn't any reference point to judge accurate reproduction of the signal because there is no way of knowing what the recording sounded like in the first place.

Sure, if you sample an analog signal and then play it back there is no way to *listen* to it and tell if it is identical. This is the same as what I was saying to refute the notion that analog is 'better' than digital. If you record something to analog, there is no way to determine that it is perfectly accurate compared to what you heard because for one thing, the microphone(s) may not have captured it perfectly.

We were talking about sampling the analog waveform once you have it and whether or not sampling can reproduce that waveform accurately.

On a unrelated point, he always tells me that you shouldn't use hifi speakers for hometheater applications because the digital signals are really, really quick and these effects will not be reproduced properly on hifi speakers.

That makes no sense whatsoever. Calling speakers 'hi-fi' implies that they are highly accurate and will reproduce what they are sent nearly perfectly. If the speakers are perfect (no such thing) then they would play the analog waveform exactly no matter how quickly it changes.

Digital signals are not 'quick' - is he associating the speed of sampling with the rapid changes in the music? Transients in music can be quick and whether or not you can accurately capture them in all cases is what the debate is all about.

I assume this idea is that because there is no headroom with digital (you cannot exceed 0 dB) that some transients could get cut off. With analog tape, if the record level is too high the tape will saturate. Kind of like a rubber band stretched too far will be slightly deformed but not break. This leads to the 'warmth' of analog that people claim is better than the harsh sound of digital clipping. With digital, if the level would exceed 0 dB, it gets set to 0 dB. That is how you get digital clipping - multiple samples in a row at the max amplitude. That problem is easily solved by setting the recording level properly or using a higher bit depth.

 

[Vaughan Odendaa]

What he means by "quick" is the digital signals used for film effects. Like explosions and other digital effects. He always tells me that home theater speakers will handle these effects far better because they were designed to reproduce them (designed to handle the very fast signal changes for HT), whereas hifi speakers are too slow to reproduce them.

So in essence he is saying that home theater speakers are "quicker". What do you think about this ? Refutations ?

BTW, thank you for discussing this even though it isn't related to the OP.

--Sincerely,

 

[WmAx]

So in essence he is saying that home theater speakers are "quicker". What do you think about this ? Refutations ?

BTW, thank you for discussing this even though it isn't related to the OP.

--Sincerely,

There is nothing worth refuting. Based upon the comments from said person that you have posted, that person is not very knowledgeable in the respective subjects.

-Chris

 

[Vaughan Odendaa]

Can you explain to me why ? Not that I am disagreeing with you but some theory on this would be appreciated. Because when we have clients come in and they want to use hifi speakers in a home theater role, he always mentions that for HT, simulated digital effects are very fast and the drivers used in HT speakers are designed to reproduce them. Speakers designed for music are not.

And I never really understood why. I never quite agreed with his opinions and now that I'm posting in forums like this I find that my manager is actually talking nonsense on the subject. Which I find to be shocking because of his credentials.

Comments would be appreciated.

--Sincerely,

 

[MDS]

Speakers are designed to reproduce the signal they receive as accurately as possible - the speaker has no idea whether it is playing an explosion from a movie soundtrack or the bass guitar from a song or music within a soundtrack of a movie.

The idea that some speakers are better for movies and some are better for music can be traced entirely to the frequency response of said speakers. You cannot generalize that 'Home Theater' speakers don't do music well or that 'Music' speakers don't do home theater well.

You need to get clarification on what is meant by 'simulated digital effects are fast'. The only notion to which 'fast' is applicable are transients and if the original analog signal was captured accurately and the speakers are accurate, they will reproduce it accurately.

Is the idea that a matrix decoder like PLII will process the signal and split out parts of it to be sent to the surround speakers and that happens so fast that the speakers cannot keep up? That's not how it works. The digital data must be converted to analog before it hits the speakers and the analog waveform is all the speakers will see. Speakers are either accurate over a given frequency range or they are not - regardless of whether the signal originated in an analog or digital format. Besides, the digital processing happens an order of magnitude or more faster than our ears could ever detect.

Maybe the guy just doesn't care for some of the effects produced by matrix decoders and like a typical 'audiophile' has assigned a vague and meaningless term to describe why.

 

[Vaughan Odendaa]

Simulated effects as in explosions. He has mentioned that on several occasions. So then what are the downsides to using hifi speakers for an AV role if high passed at 80 hz (compared to a HT speakers marketed as HT speakers ?

Thanks.

--Sincerely,

 

[MDS]

First of all, how do you know that a typical explosion was 'simulated'? Movie makers use all kinds of techniques to get the sounds they want. An explosion for the soundtrack could have been created by blowing something up and recording the sound or it could have been generated with an audio editor. It could be multiple different sounds mixed together, EQ'ed, compressed, reverbed, etc. From the perspective of a speaker, it is a waveform with a specific amplitude and frequency range.

So then what are the downsides to using hifi speakers for an AV role if high passed at 80 hz (compared to a HT speakers marketed as HT speakers)?

If the speakers have a reasonable frequency response from 80 Hz and up and a subwoofer is handling 80 Hz and below then there is no downside. Now if the speakers are only good from 80 Hz and above and you have no subwoofer to handle the low frequencies, then no they won't be good for HT (or music for that matter) because you'll miss the low frequency impact. That may be what he is getting at but that implies that 'hi-fi' speakers can't accurately reproduce low frequencies and 'HT' speakers can. A 'good' speaker is good for music AND home theater.

 

[mtrycrafts]

One argument against all of this, according to my manager, is that their really isn't any reference point to judge accurate reproduction of the signal because there is no way of knowing what the recording sounded like in the first place.

And there is with his analog recordings? LOL He is over his head on this.

On a unrelated point, he always tells me that you shouldn't use hifi speakers for hometheater applications because the digital signals are really, really quick and these effects will not be reproduced properly on hifi speakers.

What do you guys think ?

--Sincerely,


You need to listen to others, and just nod to him.

The signals is the same, no difference to speak of. Again, he is over his head, and just pushing mythology on you.

What speakers is he pushing then???

 

[mtrycrafts]

Can you explain to me why ? Not that I am disagreeing with you but some theory on this would be appreciated. Because when we have clients come in and they want to use hifi speakers in a home theater role, he always mentions that for HT, simulated digital effects are very fast and the drivers used in HT speakers are designed to reproduce them. Speakers designed for music are not.

And I never really understood why. I never quite agreed with his opinions and now that I'm posting in forums like this I find that my manager is actually talking nonsense on the subject. Which I find to be shocking because of his credentials.

Comments would be appreciated.

--Sincerely,

I bet he believes that since explosions are loud and of short duration perhaps, it must be quick? How about that kick drum? How fast is that strike and that first signal?
As was stated, he is blowing smoke.

 

[WmAx]

Can you explain to me why ? Not that I am disagreeing with you but some theory on this would be appreciated. Because when we have clients come in and they want to use hifi speakers in a home theater role, he always mentions that for HT, simulated digital effects are very fast and the drivers used in HT speakers are designed to reproduce them. Speakers designed for music are not.

I will have to have to hear specific technical claims in order to provide specific information.

And I never really understood why. I never quite agreed with his opinions and now that I'm posting in forums like this I find that my manager is actually talking nonsense on the subject. Which I find to be shocking because of his credentials.

As you are learning, credentials and/or experience are not proof that someone is correct. You will have to judge the reliability/accuracy of information from someone based upon their historical accuracy of information, which in turn requires you to independently research/verify samples of that information, as you appear to be currently attempting.

-Chris