At times, companies try to promote their new products with unverifiable or unsupported claims. This time it’s Panasonic who issued a press release about the SC-PMX152 and SC-PMX82, their new CD players. They seem fine, but there’s one thing that definitely does not work here: Panasonic claims the players can “re-master” common CDs and make them hi-res. Which is pure snake oil.
The company writes that “the SC-PMX152’s CD High-Res Re-Master function […] lets you enjoy a wide range of sound from various sources – including your favourite CDs, MP3s and smartphones. By delivering high-frequency sounds of 20kHz or more, far beyond ordinary CD playback, the CD High-Res Re-Master produces sounds that rival high-resolution audio.” To pile it on, “The Bluetooth Re-Master also compensates for any audio signals that are lost due to data compression when streaming Bluetooth sounds. This feature gives you high-quality playback of sound sources from Bluetooth-compatible devices such as smartphones and tablets.”
That’s an impressive feat, isn’t it? Except it’s not really possible. Let me delve a little into detail. When you digitally record a sound, two measures come into play: bit depth and sampling frequency. To put it simple, greater bit depth allows to discern quiet sounds from loud sounds (it’s called dynamics), while a higher sampling frequency makes it possible to capture more details and nuances. To compare this to digital pictures, bit depth represents all the different colours you can use, while sampling frequency is the resolution. The better the bit depth, the more colours you can use and the more vivid the pictures become; a bigger resolution means you can better portray all the tiny details. This is an over-simplified explanation, of course, but it’s enough to understand what we’re talking about.
Now, the point is: can you recover lost details from a digital picture? In short, the answer is no. Many attempts have been made in the past, but to no real avail – once information is lost, there’s no way to recover it. You can do things like upscaling, but it will never be the same as the original – you will see apparent differences. Google is now experimenting with AI to recover details, but it’s guessing what should be there and its accuracy has still to be improved a lot to be useful. The same applies to audio – and that’s why you should not buy “high resolution remasters” of Dire Straits: those were digitally recorded in 16 bit, 44.1 kHz quality, so “high resolution” is nowhere to be seen. Panasonic can make use of algorithms that process data and try to guess what could fill the holes, but you won’t hear any difference – the process could actually introduce artefacts and distortions, much like those edgy curves in upscaled pictures.
The issue with Panasonic’s new CD players is that the company claims it can get high resolution audio from a low resolution source – not just from CDs, but from Bluetooth and even unspecified-bitrate MP3s. Anyone in their right mind this simply cannot be true. What if I told you that $1 earphones can be made to sound just as good as $3,995 Audeze LCD 4 just with equalization and a few black magic tricks? You’d call this quackery, and rightfully so.
If this “hi-res re-master” thing was real, we’d see plenty of products which could do the same and a service such as Tidal would make no sense to exist. Heck, all the audiophile world would burn to the ground if this was true. But it’s not, sorry.
The SC-PMX152 and SC-PMX82 seem like otherwise very good and capable products. Panasonic’s lack of explanations on the so-called “re-master” process and their very broad assertions, however, cast a shadow on what could otherwise be potentially very interesting devices.
I also was wondering what re-mastering should be good for. As you explain here, converting a 16 bit number into a 24 (or 32) bit number does not add any information. And Panasonic’s marketing sentences do not make any sense to me either.
So, I have been thinking about what Panasonic engineers might have done. As far as I understand, Panasonic is using a DSP to process data digitally. For example, bass and treble adjustments might be perfomed in the digital domain rather than in the analog domain. And now the digital signal processing in the DSP might be done on “re-mastered” 24 (or 32) bit numbers rather than on 16 bit numbers. A similar approach is done in photography by RAW converters. For example, RawTherapee converts 12 bit CCD data into 32 bit floating point numbers, applies all calculations to the 32 bit floating point numbers and eventually converts the processed data into 8 bit JPEG data. This approach ensures that the final histogram of the JPEG is gapless (of course only, if the original histogram is gapless too). So, converting the 16 bit data into a higher resolution data and applying all calculations to the higher resolution data rather than to the 16 bit data would make sense to me. From a technical point of view, such an approach would indeed improve the quality of the processed signal. Whether someone really can hear a difference is, however, another story.
Another possible explanation might be that increasing the sampling rate from 44 kHz to, for example, 96 kHz (or even much higher) would remove the (unwanted) periodic spectrum in the frequency range from 22 kHz (44 kHz / 2) to 48 kHz (96 kHz / 2). This would result in a higher suppression of the (unwanted) periodic spectrum above 48 kHz by the anti-aliasing filter, which should have a 3 dB cut-off frequency just above 20 kHz in this case.
From a technical point of view, also this approach would indeed improve the quality of the processed signal. Maybe there exist some ears that really can hear a difference.
However, from the marketing garbage it is difficult to conclude what Panasonic engineers have actually implemented.
What do you think about these possible implementations (processing is done on high-resolution data rather than on 16-bit data and/or shift of Nyquist frequency to better suppress unwanted periodic spectrum)?
Hi, Armin!
You have very good points here and your arguments make a lot of sense. The signal quality may indeed benefit from the processing and whichever DSP is in the device could actually improve the perceived sound quality or apply good quality filtering or equalization.
The point I was trying to make, though, is that Panasonic advertises the system as being able to restore lost details and frequencies – but that is impossible, as we know. So, even if the “remastering” could actually benefit the sound (but it remains to be seen whether anyone can hear a difference, as you pointed out), the way Panasonic worded it is completely wrong and is utter marketing garbage.
Processing on higher-resolution data and/or Nyquist frequency shift could indeed make the device sound better and bring additional value to it, but I think we have too little data to work with to reach a definitive conclusion on the matter since it’s all concealed behind Panasonic’s terrible marketing efforts.