Intro:
Like many others here, I have been extremely interested in HQPlayer as a tool to maximize the performance of my DAC. One such feature is the unique apodizing filters that recognize and correct errors in the mastering process. After exploring my music library, I discovered tracks with hundreds or even thousands of detected apodizing errors. Initially, when casually flipping between filters, it seemed that the apodizing filters were reducing high frequency artifacts, but I wasn’t sure so I decided to rigorously test my perception of these differences.
Chain: HQPlayer 4 Pro trial → Gustard A26 (ASIO usb driver &) → Ferrum Oor amp + Hypsos power supply → Hifiman Arya Organic Headphones
Source Files:
“Bubbles” by Yosi Horikawa 0:00s to 0:59s (~300 apodizing errors) 16bit 44.1kHz PCM .flac
“Inu K A 3l” by Hiroyuki Sawano 0:30s to 0:59s (~1200 apodizing errors) 16bit 44.1kHz PCM .flac
Upsampling:
File type: PCM .wav
Dither: LNS15
Bit Rate: 705.6k
Apodizing filter: poly-sinc-gauss-xla
Non-Apodizing filter: poly-sinc-gauss-xl
File type: DSD .dsf converted to DoP (DSD over PCM) .wav using XRecode 3
Modulator: ASDM7ECv2
Bit Rate: 44.1k x256 = 11.2M
Apodizing filter: poly-sinc-gauss-xla
Non-Apodizing filter: poly-sinc-gauss-xl
Playback software: foobar2000 Download foobar2000
ABX plugin: foobar2000: Components Repository - ABX Comparator
Delta waveform comparator software: DeltaWave https://deltaw.org/
Methods:
I used HQPlayer 4 Pro trial to upsample and save both PCM and DSD output files. To make the DSD files compatible with the foobar2000 ABX plugin, I converted the .dsf files to DoP files using XRecode, which results in lossless DSD playback when received by my DAC. I used foobar2000 ABX plugin to run blind comparison tests to assess audible differences between the Apodizing and non-Apodizing filters in both PCM and DSD256 formats.
I also used DetlaWave to compare the difference in PCM waveforms between the apodizing and non-apodizing files. This software provides both a visual representation and an audible playback of the delta waveform.
Results:
I was not able to pass the ABX blind comparison tests for either of my two test tracks in PCM or DSD formats. When comparing the PCM waveforms in DeltaWave, high frequency artifacts were present in the non-apodizing files compared to the apodizing files on both test tracks, however, they were inaudible during playback of the delta waveform. They became audible at around +20dB amplification.
Conclusion:
In the two test tracks used, each containing hundreds of apodizing errors, I perceived no audible difference between comparable apodizing and non-apodizing filters. The apodizing filter seems to be eliminating high-frequency artifacts given the delta waveform comparator results, however, these artifacts were inaudible at normal listening volumes. A more definitive test should be performed by recapturing the output from my amp with a analog-to-digital converter and comparing those waveforms in DeltaWave, however, I do not currently have the means to run that test. Also, I assume that not all apodizing errors are created equal, and it is possible that despite the high number of apodizing errors in my test tracks, those errors may be relatively inaudible compared to more severe apodizing errors in other tracks. I would gladly re-run these tests on tracks with a high degree of perceived difference if posted here.