-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
This question may be naive...
- Thread starter mike7877
- Start date
DVDdoug
Major Contributor
- Joined
- May 27, 2021
- Messages
- 3,035
- Likes
- 4,000
You can't "magically" add information.
You're just interpolating between samples, and the DAC (digital-to-analog converter) already fills-in the space between samples with a smooth, filtered, continuous analog wave.
Plus, CD audio (16-bit, 44.1kHz) is generally better than human hearing. It's very unlikely that you'll hear the difference between a high resolution original and a copy downsampled to 16/44.1 in a proper, scientific, blind, ABX Listening Test.
Up-sampling video is different because, assuming a digital display, there is no analog conversion and it IS useful to interpolate between pixels. You aren't truly increasing the resolution but you are making it appear "smoother" (essentially blurring between the real pixels).
You don't normally hear phase shifts either, unless it's relative to something. Like if there is a phase difference between the left & right channels.
At 1kHz, the acoustic wavelength (360 degrees) is about 1-foot so with a few feet between the speaker and your ear you've got a few-thousand degrees of phase shift. Higher frequencies are proportionally smaller and lower frequencies are proportionally lower so you get more phase shift at higher frequencies.
If you flip the phase 180 degrees (i.e. invert the polarity) of one channel you'll get bass cancelation and a weird "spacey-phasey" sound that sometimes creates a "stereo widening" effect. (You can do that by reversing the connections to one speaker, or you can do it digitally in Audacity.) If you reverse both speakers everything will sound normal again.
It's actually the out-of-phase soundwaves canceling in the air... With headphones the left & right soundwaves don't combine so you don't get the effect. (You can't electrically reverse one side of regular headphones because they share a common ground.)
If the left & right channels are entirely different, say a piano on the left and a guitar on the right the phase relationship is totally random and inverting one side won't make any difference.
If you have a recording, let's say you record yourself saying "Hello". You can make a copy, invert it, and if you mix it (digitally or electrically) with the original, you are effectively subtracting (adding a negative) and you'll get pure silence. (If you mix two identical in-phase copies, you'll simply double the volume). However, if you make two separate recordings of yourself saying "Hello", the phase relationships are totally random and if you mix them normally or invert one before mixing, either way it will come-out as a normal mix like you and your twin saying "Hello" together.
...You DO need an anti-aliasing filter in front of the ADC when recording because the aliasing frequencies appear down-lower in the audible range and they can't be filtered-out later.
You're just interpolating between samples, and the DAC (digital-to-analog converter) already fills-in the space between samples with a smooth, filtered, continuous analog wave.
Plus, CD audio (16-bit, 44.1kHz) is generally better than human hearing. It's very unlikely that you'll hear the difference between a high resolution original and a copy downsampled to 16/44.1 in a proper, scientific, blind, ABX Listening Test.
Up-sampling video is different because, assuming a digital display, there is no analog conversion and it IS useful to interpolate between pixels. You aren't truly increasing the resolution but you are making it appear "smoother" (essentially blurring between the real pixels).
You don't normally hear phase shifts either, unless it's relative to something. Like if there is a phase difference between the left & right channels.
At 1kHz, the acoustic wavelength (360 degrees) is about 1-foot so with a few feet between the speaker and your ear you've got a few-thousand degrees of phase shift. Higher frequencies are proportionally smaller and lower frequencies are proportionally lower so you get more phase shift at higher frequencies.
If you flip the phase 180 degrees (i.e. invert the polarity) of one channel you'll get bass cancelation and a weird "spacey-phasey" sound that sometimes creates a "stereo widening" effect. (You can do that by reversing the connections to one speaker, or you can do it digitally in Audacity.) If you reverse both speakers everything will sound normal again.
It's actually the out-of-phase soundwaves canceling in the air... With headphones the left & right soundwaves don't combine so you don't get the effect. (You can't electrically reverse one side of regular headphones because they share a common ground.)
If the left & right channels are entirely different, say a piano on the left and a guitar on the right the phase relationship is totally random and inverting one side won't make any difference.
If you have a recording, let's say you record yourself saying "Hello". You can make a copy, invert it, and if you mix it (digitally or electrically) with the original, you are effectively subtracting (adding a negative) and you'll get pure silence. (If you mix two identical in-phase copies, you'll simply double the volume). However, if you make two separate recordings of yourself saying "Hello", the phase relationships are totally random and if you mix them normally or invert one before mixing, either way it will come-out as a normal mix like you and your twin saying "Hello" together.
Something I found "interesting"... Once I connected an oscilloscope to the soundcard of my computer at work. I don't remember what the experiment was about, but I was SHOCKED to see a totally unfiltered stair-stepped output! I didn't use that computer for any "high fidelity" listening but I did sometimes listen to music after hours. I had never heard anything wrong! Then I thought about it... The harmonics and clock frequency are above the audible range, plus the amplifier (built-into the "computer speakers") might provide some filtering, and the speakers will certainly provide some mechanical filtering because they can't go to 44.1kHz. (Those speakers probably didn't go to 20khz, and neither do my ears.)
...You DO need an anti-aliasing filter in front of the ADC when recording because the aliasing frequencies appear down-lower in the audible range and they can't be filtered-out later.
Last edited:
- Thread Starter
- #3
I'm thinking about the filter specifically - when you don't attenuate above 1/2 the sample rate enough (using minimum phase filter) you can get reflections of the supersonic distortion/harmonics in the audible band. Would doing what I suggested above at least rectify this situation?
What about just resampling 44.1 to 88.2 or 176.4 in foobar - the filter would be far above any recorded frequencies - would this eliminate the potential above mentioned reflections?
I'm not asking about potential audibility here, just what would/should be observed by a scope on the output
What about just resampling 44.1 to 88.2 or 176.4 in foobar - the filter would be far above any recorded frequencies - would this eliminate the potential above mentioned reflections?
I'm not asking about potential audibility here, just what would/should be observed by a scope on the output
Because most DACs will resample at 64 to 256 times the base rate already
And the act of resampling, means applying a low pass filter. A slow filter will not do you any good. It will just leak and lead to a bad reconstruction.
Why do you think that there is a phase problem at the high-end? Linear phase filters are used, and as the name says: they are linear phase
oleg87
Senior Member
Every modern audio DAC (that isn't some audiophile oddity, which will surely advertise this "feature") oversamples. Assuming your DAC lets you select the filter, as most recent audiophile DACs do, if you get phase distortion that's a self-inflicted problem (but a likely inaudible one).
- Thread Starter
- #6
THD+n measurements I've seen for most DACs don't show itBecause most DACs will resample at 64 to 256 times the base rate already
This is the answer to my question, Thank you
I don't think there's a phase problem, just those reflections. And I never said problem "I'm not asking about potential audibility here, just what would/should be observed by a scope on the output" post 3Why do you think that there is a phase problem at the high-end? Linear phase filters are used, and as the name says: they are linear phase
Why do you think they don’t show that?
Similar threads
