Apodizing Filters opinion?
I wonder if anybody compared Apodizing and Non-Apodizing Filters on DACs that allows selection. I found this written by Vice President (and Technical Director) of Benchmark Media Systems, Inc :
To "Apodize" or not to "Apodize"
Submitted by JohnSiau on Tue, 2009-08-04 17:16.
"Apodizing" filters offer significant advantages over some conventional systems, but only offer some of the advantages provided by Benchmark's UltraLock(TM) upsampling system. Benchmark is not planning to introduce an Apodizing filter because we feel this would be a step backward.
Apodizing filters make some improvements over conventional systems, but also have some shortcomings. They reduce pre-ring at near Nyquist frequencies, but they have non-linear phase response, and are prone to high levels of image fold back. There is no good evidence that the pre-ring is audible in a 44.1 kHz system, but there is evidence that non-linear phase response and image fold back may be audible.
Our tests indicate that image fold back may be the most audible defect in conventional 44.1 kHz PCM systems. Benchmark's upsampling system is virtually free from image fold back problems due to the implementation of a 46 GHz conceptual interpolator and zero-order hold. Fold back in conventional D/A converters can easily be seen on an FFT analysis and its effects can easily be viewed on an oscilloscope. Image fold back causes amplitude modulation of high-frequency sounds, but worse yet, it folds inaudible high-frequency sounds down into the audible band where they can be heard easily (but at the wrong frequency).
We believe the overwhelming success of Benchmark's DAC1 converter family is due to its freedom from image fold back as well as its immunity to jitter-induced sidebands. The Apodizing filters may sound different than Benchmark's system, but for the above reasons we do not believe that systems with Apodizing filters are as transparent as the Benchmark UltraLock(TM) upsampling system.
John Siau
Vice President
Benchmark Media Systems, Inc.
To "Apodize" or not to "Apodize"
Submitted by JohnSiau on Tue, 2009-08-04 17:16.
"Apodizing" filters offer significant advantages over some conventional systems, but only offer some of the advantages provided by Benchmark's UltraLock(TM) upsampling system. Benchmark is not planning to introduce an Apodizing filter because we feel this would be a step backward.
Apodizing filters make some improvements over conventional systems, but also have some shortcomings. They reduce pre-ring at near Nyquist frequencies, but they have non-linear phase response, and are prone to high levels of image fold back. There is no good evidence that the pre-ring is audible in a 44.1 kHz system, but there is evidence that non-linear phase response and image fold back may be audible.
Our tests indicate that image fold back may be the most audible defect in conventional 44.1 kHz PCM systems. Benchmark's upsampling system is virtually free from image fold back problems due to the implementation of a 46 GHz conceptual interpolator and zero-order hold. Fold back in conventional D/A converters can easily be seen on an FFT analysis and its effects can easily be viewed on an oscilloscope. Image fold back causes amplitude modulation of high-frequency sounds, but worse yet, it folds inaudible high-frequency sounds down into the audible band where they can be heard easily (but at the wrong frequency).
We believe the overwhelming success of Benchmark's DAC1 converter family is due to its freedom from image fold back as well as its immunity to jitter-induced sidebands. The Apodizing filters may sound different than Benchmark's system, but for the above reasons we do not believe that systems with Apodizing filters are as transparent as the Benchmark UltraLock(TM) upsampling system.
John Siau
Vice President
Benchmark Media Systems, Inc.
0 responses Add your response