This should get some debate going!
- Thread starter Johnny Vinyl
- Start date
Yes, at the input, just like j_j said (analog to digital conversion), and they end up to be audible, at the output, through your ears.
Don,
I think what J_J means with clip a DIGITAL signal, is just overloading the input of an ADC.
Really?
So the digital file we see is the result of ADC clipping which has aliased back into the audio band creating anharmonic distortion
Then when this digital file (with it's distortions) is used to create the vinyl file, its aliasing distortions disappear from within the audio band?
Is that what he's saying? How does this work then?
What can I say, Mark.
John, I'm sure j_j will answer your questions with precision.
What j_j is talking about is clipping that occurs after any antialiasing filtering at the input. e.g. in the ADC or due to accidental saturation in software etc. If this occurs, frequency components above the Nyquist frequency are generated which alias back into the audible range. The DAC will faithfully reproduce these aliasing components.
DonH may not have been talking about exactly the same thing.
Last edited:
The conclusion is yours - I would never make such conclusion from this debate. Anyone sees that all this is most probably just semantics.
BTW, although you consider us funny, your post was not even funny.
No kidding. If you clip inside the digital framework, as I just said, what would have been harmonics in an analog chain become aliases that fold back into the passband, which of course then go right through the DAC's antiimaging filter.
Ditto if you clip at the ADC.
So, yes, you can get aliasing inside a digital system, and of course, since aliasing products are in-band, they go right through the DAC's anti-imaging filter.
I would suggest that you consider the graph you saw. It shows in-band signals, which are very obviously not going to be removed by an anti-imaging filter.
I have no idea how you think anyone is saying that. Once you've created in-band aliases, you're out of luck. You're stuck with them.
The point is to never CREATE them. Ever. Except, I suppose, as a part of some strange artistic sound...
Well, if I remember correctly, the vinyl file was created from the digital file, right?
The same digital file that contains these folded back aliases, right?
These would simply pass through the analog system also?
Therefore these two systems are exactly the same in their treatment of such aliases.
Your use of this as some sort of explanation for the differences between the example digital & resulting vinyl file seems rather incorrect
No tangent. JJ introduced this & you agreed but now say there is no clipping. Care to bring some clarity?
Sorry I stepped into the middle of the thread, just happened to catch the comment and did not read back to see the full context. After the DAC output no sampling takes place and thus no aliasing. The DAC will generate images at multiples of Fs/2 and those should be filtered. Any signal that equals or exceeds 1/2 the sampling rate before the ADC, due to clipping or whatever, will cause aliasing. As for the digital chain between the ADC's output and the DAC's input, that depends upon the processing and can go either way. Being an analog guy a decade or two from my last grad classes in digital signal processing, I'll defer to the experts and stay out of this one.
Enjoy - Don (need to stay away and find WGE forum now...)
Enjoy - Don (need to stay away and find WGE forum now...)
So he was just making a general comment about digital Vs Vinyl handling of clipping? OK! It came directly after your detailed post on the differences you heard. Kinda confusing, particularly when followed up by the posts in defence of his statement!
All good. Btw, Don, any kind of clipping in the purely digital domain is going to do "bad things" if it's extensive enough to create any sort of noticeable harmonics or intermod.
Similar threads
