What is "warm" sound anyways? Serious question. The term comes up a lot and the only thing I can think of when hearing it is "limited bandwidth, bloated bass, low slew rate".
What is generally associated with that term?
Again, it is a serious question.
What is generally associated with that term?
Again, it is a serious question.
I dunno, but early generations of digital audio used a low sampling rate of 44 or 48 KHz compared to the high-frequency bias of 60 to 120 KHz of analogue tape machines. If fsampling would have been fixed to 100 KHz and lowpass filters properly designed (not just brute-force oversamping), we would have never had such a big problem. Take your time, take it easy.
Tape HF bias and digital sampling rate are not directly comparable, as they do quite different things. You might as well say that 455kHz sampling is needed as that was used for AM radio IF stages.
At 100kHz sampling it would not have been possible to fit Beethoven's Ninth on a CD - I believe that was the criterion imposed by Sony?
"brute-force oversamping" is a very clever way of achieving a good filter at reasonable cost and reliability. Without this we would have inferior digital audio.
It amuses me that people worry about 'inadequate' sampling rates and bit depths, yet mostly listen to highly processed pop music with very small dynamic range and no proper acoustic signal to compare it with anyway.
At 100kHz sampling it would not have been possible to fit Beethoven's Ninth on a CD - I believe that was the criterion imposed by Sony?
"brute-force oversamping" is a very clever way of achieving a good filter at reasonable cost and reliability. Without this we would have inferior digital audio.
It amuses me that people worry about 'inadequate' sampling rates and bit depths, yet mostly listen to highly processed pop music with very small dynamic range and no proper acoustic signal to compare it with anyway.
About comparability, feed a loud 30 to 60 KHz sine sweep into an analogue tape recorder and hear the aliasing. Oversampling has some use, but lowpasses for digital audio should produce more post- than pre-ringing (linear phase lowpasses produce equal parts of pre- and post-ringing). Man's senses vary with specimen, time and place, and he phantasises and forgets, but i hold my ground, that CDs are as much inferiour as superiour to vinyl.
Last edited:
Tape recorders don't do aliasing, because they don't do sampling. They do heterodyning, due to second order distortion. It just happens that heterodyning and aliasing may be at the same frequency, but different mechanisms so not comparable. Good tape recorders will include filters.
It is easy to frighten yourself by looking at the output of filters given unlikely inputs. That is why filterless NOS DACs are so popular; people complain about digital audio but then deliberately set out to fail to reproduce the input, and regard that as progress!
It is easy to frighten yourself by looking at the output of filters given unlikely inputs. That is why filterless NOS DACs are so popular; people complain about digital audio but then deliberately set out to fail to reproduce the input, and regard that as progress!
I must smile when I remember Mike Uwins article in Linear Audio when he had people playing vinyl when secretly he played a digital recording.
Almost all were happily pointing out the great warm, smooth sound of vinyl.
People are funny.
Jan
Almost all were happily pointing out the great warm, smooth sound of vinyl.
People are funny.
Jan
Overcoming magnetical hysteresis by a high-frequency bias is a sort of sampling in the time domain. Aliasing means, that new frequencies being minus fnyquist plus fsignal are produced.
Is it? I thought the HF bias dropped in intensity as the tape left the head gap - there is no trace of the bias left on the tape so it isn't sampling.Grasso789 said:
If aliasing artefacts are present on the recording, it must have been sampled.
Sampling means, that a sameness is established, for example by cutting a bamboo stick at water height. Space has been sampled but not rasterized, because the cut has been made as exactly as possible at water height. Time is rasterized, as the cut has been made at the certain time of morning. Next morning may be made another sample.
If one and the same bamboo stick is used every morning, with millimeter marks on it, and integer millimeter values read and written on a paper, then space is also rasterized.
Sampling means, that a sameness is established, for example by cutting a bamboo stick at water height. Space has been sampled but not rasterized, because the cut has been made as exactly as possible at water height. Time is rasterized, as the cut has been made at the certain time of morning. Next morning may be made another sample.
If one and the same bamboo stick is used every morning, with millimeter marks on it, and integer millimeter values read and written on a paper, then space is also rasterized.
If someone alleges that recorded artifacts are due to aliasing this does not prove that they are due to aliasing so it is not proof that sampling has taken place. There are other phenomena which can give rise to the same artifacts, such as heterodyning.Grasso789 said:
If we are talking about digital audio electronics then it helps if we use words with their normal engineering meanings. 'Sampling' means taking snapshots of a signal, so all we have is a set of values and times (which may or may not be uniformly spaced - but usually are uniform).
This is a troll, right?
I've said it before, I'll say it again: My system is more digital now than it's ever been - about as digital as it can be! - and it's never sounded more natural, real, and yes, "warm" than it does right now.
In conclusion, let me just say:
-- Jim
Same goes for me. And Ill also add detail to the list lots and lots of detail, and ease of use and storage, fewer components,...
Different strokes.
I vehemently reject all digital audio. Also all tape recordings and all vinyl. Rubbishy new-fangled garbage, all of it!
Everyone with true golden ears (mine are platinum alloyed with gold) knows that the only really good audio recordings were made on tin foil, wrapped around the cylinder in an 1870's Edison phonograph.
The digitally reproduced letters on my computer screen are are starting to make my neck itch, which is a well known side-effect of digital letters. So it's time for me to go back to my all-analogue quill pen and iron gall ink.
-Gnobuddy
Everyone with true golden ears (mine are platinum alloyed with gold) knows that the only really good audio recordings were made on tin foil, wrapped around the cylinder in an 1870's Edison phonograph.
The digitally reproduced letters on my computer screen are are starting to make my neck itch, which is a well known side-effect of digital letters. So it's time for me to go back to my all-analogue quill pen and iron gall ink.
-Gnobuddy
No. Aliasing is always a result of feeding wrong signals into a sampling machine. Class-D and radio are of course sampling machines, too, tho with very high sampling rates.
Times yes, values no. A value is a mental thing, one, which cannot be transferred to other minds without the use of numbers. A number for water height is either inexact or theoretical. But a sample is no value but a physical thing.
Hence digital audio does NOT calculate with samples. It rasterizes amplitude and so generates amplitude values. But a value is not a sample of an analogue, real thing.
Last edited:
You have your logic back to front. Aliasing can happen when sampling takes place. The existence of a signal component at a frequency consistent with aliasing is not proof that it is an alias, and so not proof that sampling has taken place.Grasso789 said:
Class D I don't know about. Radio as a system does not use sampling, although particular implementations of radio may use sampling - sometimes without realising it. For example, a traditional AM transmitter does not use sampling but a receiver envelope detector does use sampling. There are other reception methods (e.g. product detector) which do not use sampling.
You are still playing with your own private meanings of words. This does not aid communication.
- Status
- Not open for further replies.