How to Improve Sound Quality [Step-by-Step Guide]

Tutorials

In this article term "to improve sound quality" mean "to decrease sound playback distortions at the analog output of digital-analog converter". Read below how to try to decrease the playback distortions for your audio equipment or portable device (digital audio player, mobile phone, etc.).

Disclaimer: Result of sound quality improvement method, that described below, depends on used playback hardware and software and sound quality improvement isn't guaranteed.

DOWNLOAD FREE
[Windows]

DOWNLOAD FREE
[Mac]

DSD audio file conversion

Resampling and bit depth impact to sound quality

Digital-analog converter audio hardware may have different distortion levels for different modes. Here DAC mode is a combination of sample rate, bit depth, PCM/DSD.

Different distortions for different DAC modes

How to Improve Sound Quality

One or several modes may have minimal distortions. Below such mode is called as "the best sounding mode".

Audio file resolution may be converted to values (sample rate, bit depth, PCM/DSD) of "the best sounding mode" to improve playback sound quality (optimization of audio files).

Read below about:

dithering, audio resolution conversion software, and
step-by-step guide on how to check your audio system abilities. Probably it will help to improve the system's sound.

Example:

DAC can be able to provide lesser distortions for mode 192 kHz/24 bit/PCM than for 96 kHz/24 bit/PCM.
Therefore, the conversion of audio file resolution to 192 kHz/24 bit/PCM is recommended for playback at the DAC.

Theoretically, a higher sample rate and bit depth cause lesser noise/distortion level. However, it is an implementation matter rather.

"The best sounding mode" may be found via measurements of the test signals (see at the picture above). However, it demands precise measurement equipment and skills in the measurements.

The article describes step-by-step how to found "the best sounding mode" by ears. It is not a precise way, unlike measurements. But ears allow approximately estimate the DAC's modes at home without additional measurement equipment.

Read more:

Dithering impact to sound quality

In contrast, audio-file-resolution changing, that work together with DAC, dithering allow increasing the sound quality of an audio file as itself in some cases. Because non-linear distortions of quantization, initially correlated with an audio signal, are smoothed and transformed into uncorrelated noise after dithering.

Dither can improve the sound quality of an audio file via correlated distortion decreasing

If DAC has a noise level above the quantization noise of audio file, the dithering impact may be decreased or disappeared even. But modern DACs, as rule, have noise level below 16-bit error quantization and dithering noise. Therefore, for such DAC dithering may impact to perceived sound quality.

Read more:

About dithering...

Impact of audio conversion software to result

Below pro audio conversion software is suggested to convert audio files. It has low distortion noise. Theoretically, we can neglect quality losses of conversion (including non-multiple resampling) for pro class of audio converters.

However, different modes of the audio converter's resampling filters may are there, that can cause different ear perception due to different features:

steepness of transient band,
ringing,
flatness of frequency response,
phase response linearity,
etc.

Different converters also have different features of the resampling filters.

Currently, the author knows nothing about reliable results of researches of resampling-filter-features impact to sound perception.

Also, the filtration bandwidth may impact intermodulation distortions at an analog output of DAC.

Watch video about the experiment: we can hear ultrasound noise at some equipment.

Video about intermodulation distortions:
Experiment with removing ultrasound upper 20 kHz ("Optimized resampling filter") during audio file conversion

Watch and share: How to remove audible noise for high-resolution audio

Read more:

Recommendations for audio player

For a purer experiment, audio playback software (audio player and system audio driver) should transfer digital audio data from file to DAC without changes of binary content. As example, a preferable way for audio player software is using ASIO or WASAPI output and turned off processing.

For hardware (devices) adjusting of settings for minimal/without sound processing is recommended.

About digital acoustic-correction of room read below.

Read more:

Read step-by-step guide on how to check your audio system abilities. Probably it will help you to improve the sound.

How to improve sound quality via "the best sounding mode"

Described below testing algorithm may be used for DAC, digital audio player, music server, any audio device, that have DAC chip.

Found several samples of audio files with different bit depth (16, 24 bit, DSF, DFF, ISO, etc.) for testing of your audio equipment.
Launch AuI ConverteR 48x44.
For sound checking may be used AuI ConverteR 48x44 Free. Despite free version limitations, it allows to estimate the abilities of audio file optimization for your equipment or device.
Open source audio file(s) from step #1.
Go to AuI ConverteR's Settings > General:
Uncheck "Don't make DSP for unchanged sample rates" switch.
In Filter Mode list select "Optimized resampling filter" (read about filter modes).
Uncheck "Use minimum phase filter (without pre-ringing)" switch.
Minimum phase filter causes post-ringing only (no pre-ringing). But, approximatelly, energy of the post-ringing is doubled comparing post-ringing of linear (not minimum) phase filter, that causes both pre- and post-ringing.
Click OK button.
Select one of output format: WAV, FLAC, AIFF. If DAC supports DSD mode only, go to step #19.
Select output sample rate, bit depth.
Recommended to begin for maximal sample rate and bit depth, that is native for your DAC.
However, a higher sample rate and/or bit depth don't guarantee better sound quality.
In AuI ConverteR's main window set Dither On.
For both input and output bit depth upper 16-bit "Dither" button will be ignored.
Select new target directory
Advice: use special name directory for recognizing of used conversion parameters during further testing.
Example: "44100-16bit-DthOn", "DSD64-DthOn", etc.
Push Start button and wait for end of conversion.
In the main window set Dither Off.
Select new target directory.
Push Start button and wait for the conversion end.
Repeat steps #10 ... 16 for all available sample rate and bit depth combinations, that available for your DAC or device.
Repeat steps #10 ... 17 for all filter modes (Settings > General > Filter Mode).
For "Optimized resampling filter" test both checked and unchecked "Use minimum phase filter (without pre-ringing)" switch.
For other filter modes, this switch is ignored.
If your DAC support DSD mode, repeat steps #10 ... 18 for output format DSF.
Listen and compare similar files for all conversion modes: compare files with similar audio content, that placed in the different target directories (see steps #12 and #15).
Conversion parameters of converted audio files, that have the most preferable sound, you may use for conversion (optimization) of your library to playback on tested DAC.

In general, for the DAC mode comparison, the author recommends turn off digital room acoustic correction. But you can compare results with the room correction and without.

Filter mode description

AuI ConverteR has 3 resampling filter modes (Settings > General > Filter Mode):

Optimized;
Non-optimized;
Non-optimized wide;
Optimized G5 (since version 11.x).

Optimized filter

Optimized filter cut ultrasound (all above 20 kHz).

Non-optimized filter

Non-optimized filter:

for PCM signals, cut half of minimum sampling rate between input and output;
for DSD signals, cut all above 20...27 kHz depending on input sampling rate (always lesser half of minimum sampling rate between input and output ones).

Non-optimized wide filter

Non-optimized wide filter:

for PCM signals, cut half of minimum sampling rate between input and output;
for DSD signals, cut all above 20...100 kHz depending on input sampling rate (always lesser half of minimum sampling rate between input and output ones).

Optimized G5 filter

Optimized G5 filter cuts ultrasound (all above 20 kHz) and causes shorter ringing length for given threshold than Optimized one.

Our recommendations for the mode using:

Optimized mode is common recommended for better sound quality of playback.
However, you can use any mode, that you like more. Refer to your personal listening experience.
If non-optimized mode causes noise at specific software+hardware, try optimized mode.
If you distribute converted recordings, non-optimized mode for PCM may be recommended. For DSD files, either non-optimized or non-optimized wide mode may be recommended.
If you use optimized mode, the distributed recording may looks like conversion from CD. Despite it is not the same, your customers may have doubts in source of the distributed hi-res recordings, and additional explanation is need.

Tips

If Settings > General > "Don't make DSP..." switch is checked and no audio processing (resampling, gain, DSD to/from PCM, etc.) is applied, resampling filter is ignored.
Filters are calibrated for maximum level of seep sines, but big spectrum energy of musical signal can cause distortions. So, reducing of output level at 0.4 ... 3 dB may be recommended when distortions are happen.

Resume

If optimization of audio files gives audible sound advantages in your case, you can convert your audio library to "the best sounding mode"'s parameters: sample rate, bit depth, DSD/PCM.

If in your case no audible sound advantages by optimization, there is not necessary to optimize your library.

It is strongly recommended, original audio files are backup. Because you, probably, will buy new devices or/and new resolution conversion algorithms will be released.

Yuri Korzunov,
Audiophile Inventory's developer

Read the articles about sound quality

January 12, 2022 updated | since May 09, 2017

Русский