Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Upsample, upscale low-fi audio to improve quality -what tool to use? (Read 16534 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Hello!

I have some low-fi audio and I want to upconvert bit-depth and sampling rate achieve a better sound. Are there any recommended software?

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #1
You cannot create information if it's not there. Simply increasing bit depth and sample rate will do nothing at all. You may be able to improve the perceived quality by using Digital Sound Processing (DSP) tools, but what to use really depends on what kind of sample you have and what you want to achieve.

FWIW, if you want to confirm my first statement yourself, you can try tools like Audacity or SoX.
It's only audiophile if it's inconvenient.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #2
Sorry, but I find it confusing that you have been registered here since 2007 and yet still not know that creating quality/information from nothingness is impossible!

Whether any DSP can improve the sound by altering the existing information depends completely on your perceptions and preferences.

To forestall a possible objection: Such DSPs might involve upsampling or increasing the number of bits from their current values to match the rates used by the DSP or to create headroom for guessed information (perhaps the prediction of high frequencies from lower ones), but it must be stated categorically that those parameters themselves will not be responsible for any subjective improvement you might attain. They are simply preconditions for the processing performed by the DSP, so again: changing only the rates, without any other mangling of the signal, cannot possible have any magical effect.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #3
You cannot create information if it's not there. Simply increasing bit depth and sample rate will do nothing at all. You may be able to improve the perceived quality by using Digital Sound Processing (DSP) tools, but what to use really depends on what kind of sample you have and what you want to achieve.

FWIW, if you want to confirm my first statement yourself, you can try tools like Audacity or SoX.

I know I can't create information. But upsampling can greatly improve the result, by decreasing artifacts and getting closer to the original signal. Not much fimilar with audio processing, but with images the difference between lanczos and nearest neighbor is drastic.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #4
Sorry, but I find it confusing that you have been registered here since 2007 and yet still not know that creating quality/information from nothingness is impossible!

But better use of the information is possible.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #5
Not much fimilar with audio processing, but with images the difference between lanczos and nearest neighbor is drastic.
This analogy is useless because image-processing is, by definition, not equivalent to audio-processing. In particular, it implies that the hardware does not automatically interpolate between samples properly already and that you should do its job for it. When you play a low-rate file, what comes out is most certainly not the series of ugly stair-steps drawn by programs that, or people who, should know better. The hardware, assuming it is not hopelessly awful, will already include a reconstruction filter to smooth the signal out before it is output in analogue form. For any one digital signal, there is precisely one mathematical function that represents its properly filtered and interpolated result, and this is what reconstruction filters aim to achieve. What exactly makes you think you know better?

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #6
When you play a low-rate file, what comes out is most certainly not the series of ugly stair-steps drawn by programs that, or people who, should know better.
I'm sure I used a version of windows that did pretty much that, if you selected low quality resampling in control panel. Probably '95 or '98 or something. Don't know. Haven't looked for a long time! I'm sure someone somewhere on the net has documented this in painful detail for each version of windows.

Cheers,
David.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #7
And about analogy with image processing. Lanczos does not improve image quality over Bilinear, it just does not destroy it. None of spatial resize filters produce new information, those only try to preserve existing one. There are superresolution resizers that try to interpolate new information in frame using surronding frames but that is not really posible within audio.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #8
I know I can't create information. But upsampling can greatly improve the result, by decreasing artifacts and getting closer to the original signal.


Unfortunately, this is not how it works.

Not much fimilar with audio processing, but with images the difference between lanczos and nearest neighbor is drastic.


Lanczos only looks good if you're comparing it to something that works badly, like in your example of nearest neighbor.  It doesn't actually improve the quality of the original though.  The upsampled version is larger but not actually better.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #9
The difference is that a digital image is always displayed as pixels, and you can display more pixels by interpolating between existing pixels.

In audio, on the other hand, the digital data are converted to analog in the DAC and reconstruction filter, and in this process there is essentially infinite interpolation, i.e. it becomes a continuous signal. It makes no difference if you add more data points in the digital domain because the result will be the same once you convert to the analog domain.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #10
The difference is that a digital image is always displayed as pixels, and you can display more pixels by interpolating between existing pixels.


Technically speaking, this is no different than audio.  Pixels are just 2D samples. 

In audio, on the other hand, the digital data are converted to analog in the DAC and reconstruction filter, and in this process there is essentially infinite interpolation, i.e. it becomes a continuous signal.


Diffraction is a low pass filtering operation that interpolates the discrete pixels into a continuously varying optical waveform.


Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #11
None of spatial resize filters produce new information, those only try to preserve existing one. There are superresolution resizers that try to interpolate new information in frame using surronding frames but that is not really posible within audio.


Well, you could have said that there's no more room for improvement.


Quote
Whether any DSP can improve the sound by altering the existing information depends completely on your perceptions and preferences.


You have any examples of these DSP-s.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #12
I know I can't create information. But upsampling can greatly improve the result, by decreasing artifacts and getting closer to the original signal.


Unfortunately, this is not how it works.

Lanczos only looks good if you're comparing it to something that works badly, like in your example of nearest neighbor.  It doesn't actually improve the quality of the original though.  The upsampled version is larger but not actually better.


Well, it's most certainly better perceptually.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #13
I suppose that audio upsampling is beneficial for non-oversampling zero-order hold DACs. But I don't think that it's possible to find such device these days.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #14
And about images: A Pixel Is Not A Little Square so a good image viewer should use lanczos or some other interpolation. In this case there will be no difference between original and upsampled versions of the same picture.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #15
Well, it's most certainly better perceptually.


Only if you define better to mean larger.  All Lanczos does is make the image look larger while slightly decreasing its actual resolution.  You might perceive larger images to look better in some circumstances, but thats a funny thing to claim in general.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #16
The difference is that a digital image is always displayed as pixels, and you can display more pixels by interpolating between existing pixels.


Technically speaking, this is no different than audio.  Pixels are just 2D samples. 

My point was that the display is displaying pixels, so it is conceivable that software interpolation could give a better-looking picture than just displaying raw pixels. No modern audio DAC outputs the equivalent, which would be a stair-step waveform.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #17
Simply put, a half decent sound card is already doing exactly what a mathematically correct upconversion will do.

In both cases, all frequencies above half the original sample rate are missing.

If you want to fake them, you need spectral band replication, and that doesn't work very well. For really low quality sources, it's totally unconvincing.

Cheers,
David.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #18
But upsampling can greatly improve the result, by decreasing artifacts and getting closer to the original signal.

This is a bold claim.

Do you have any examples or are you simply repeating yourself despite what you've been told?  If the latter then telling us so does not make it true.  If the former then you would have already answered the question asked in the discussion title.

Please provide some evidence to support your assertion.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #19
Hello!

I have some low-fi audio and I want to upconvert bit-depth and sampling rate achieve a better sound. Are there any recommended software?


You are badly mislead if you think that upsampling will provide any kind of practical advantage.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #20
In the space below lies the answer to clean energy that can be used to lower the average temperature of the planet:

In


the


so


for


is


sangria.

Can someone suggest a computer program that will fill in the missing information?

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #21
Most modern electronic devices (mobile phones, game consoles, media players, soundcards etc) use 16-bit 44.1/48k as the lowest natively supported resolution. If your audio files are lower than that resolution they will be automatically unconverted in real time when being played, there is no need to convert them manually. In fact they must be converted otherwise the files will be played at wrong speed and pitch.

You should ask which software converters can produce less artifacts or loss (than the default converters in the devices mentioned above) rather than "improving" quality.

PS: If 16-bit 44.1/48k means "low-fi" to your ears you may be challenged by TOS8 in this forum.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #22
Your search for a computer program is useless. The next glaciation cycle will swamp all results out of any perceptible range.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #23
Quote
I have some low-fi audio and I want to upconvert bit-depth and sampling rate achieve a better sound. Are there any recommended software?
Any audio editor can re-sample and/or save-as a different format.      An audio editor will also allow you to enhance/process the audio with noise reduction, EQ, etc.  Audacity is FREE!!!    With audacity, you can change the sample rate on the main editing screen, and you can change the bit-depth when you export (save). 

TAudioConverter (also FREE) can convert between many different formats.  It can convert almost anything to WAV at up to 32 bits and 48kHz.    (You change the sample rate with a "filter" option.)

BTW - I agree with what everybody is saying...  proper upsampling won't change the sound at all!!!  But, if you are going from 8-bits to 16-bits you may benefit from some noise reduction.  And, there may be some other enhancements/improvements you can make (with an audio editor) before re-saving.

Upsample, upscale low-fi audio to improve quality -what tool to use?

Reply #24
Quote
Whether any DSP can improve the sound by altering the existing information depends completely on your perceptions and preferences.

You have any examples of these DSP-s.
I have some badly recorded audio files (live concerts with missing high frequencies).

The quickiest way for me to "fix" this is the use of foosion's Noise Sharpening plugin for foobar2000.
This is HA. Not the Jerry Springer Show.