Thanks Fazal and Speed Racer. I will check the links. I have a Sony BDP S7200 I bought around 2014 or so. Maybe it will work.
not mine, just a cat spinning records
Oh. Ta. Very clever 
Why would I want all the hassle of ripping my SACDs when Roon and HQP will upsample the non SACD files of same music. What am I not understanding?
I donāt know anything about SACD, but arenāt they a higher quality recording than something that has been upsampled. Or is that what SACD is?
Like a lot of audio things, it depends. If the master used to make an SACD recording was recorded directly to DSD, then there will be more musical information in the SACD record. If, however, a PCM master is converted to DSD then the recording will have the information from the original master plus a lot of white space; effectively it will be upsampled.
SACD are DSD64. ROON AND HQP upsample to that and higher. I listen to most of our music at DSD64 as we prefer the SQ despite not being able to hear the difference as there cannot be oneā¦
I just read andybob. He knows better than I so gave a mucbetter answer. h
Iāve never used upsampling because I cannot hear any difference and no improvement in SQ. I prefer the purple light.
Cannot be one? Why canāt there be a positive difference between a DSD64 track played at DSD64 and upsampled to DSD128?
Because upsampling does not add any new information to the sound. Itās like when you view HD movie on a 4K display. If anything, I would consider it more like a DSP effects.
Knitman is being sardonic.
No one has ever suggested that upsampling adds information. What it does is enable better digital filters to be used that add less distortion and can shift noise away from the audible band.
Edit: Most modern DAC chip designs use internal upsampling. Jussi Laako of HQPlayer described such a DAC and compared it to HQP in this post:
Not at all like DSP effects. As @andybob said, with upsampling, you can use digital filters, noise shapers, and modulators that allow for the noise to be moved far away from the audio range and even remove it entirely. Upsampling without this would be pointless.
Most likely it is the DAC that makes the difference. Do you use the build in DAC in your CD player to output the sound or do you use a TOS link or SP/DIF out? Most likely that is where the difference (not better or worse, just different) lies. Your ripped files go through a different route hence the difference in sound.
are you referring to me?
I was. I think.
I was referring to the fact that my husband and I prefer the sound of all our ripped tracks upsampled to DSD64(I was using 128 but canāt tell the difference).
When I mentioned this a while ago I was told I was imagining the better SQ as upsampling does nothing. When I asked why then do Roon and HQP offer this, I didnāt get an answer.
I will say that my husband knows nothing about all this. He is a trained ;pianist and opera singer, tho he became a world renowned historian in his field instead.
I played him various tracks and all I did was ask him which sound he preferred. I think he thought I was messing with treble and bass.
Each time preferred DSD and he didnāt know that until I told him.
He also prefers MQA as it is āclean and openā.
1 Like
Although statements about upsampling similar to āyou canāt create missing informationā are made and generally blindly agreed to by everyone as fact, this is actually not true. Although itās way down into the technical weeds, it is actually a proven mathematical fact that if you are sampling audio within a fixed bandwidth, the Whittaker-Shannon sampling theory proves that you can recreate missing bits to perfectly recreate the original waveform with correct transients with oversampling using an infinite tap length FIR sinc filter.
It is because of this mathematical fact, that the Chord DAC (Hugo, Dave) and Scaler technology (either Blu or mScaler) were developed. Although itās not possible to implement an infinite tap length filter in real tech, Chord ā using a custom FPGA implementation that enables over 1M taps to upsample up to 768kHz and deploy their proprietary WTA (Watts Transient Alignment) algorithm ā have created tech which gets very close to the mathematical ideal and does indeed create missing information.
And itās also a known fact that your brain uses audio transients to create the sound stage of where it perceives sounds to come from. And while you donāt need more than redbook cd performance (44.1Khz @ 16 bits) to create audio within the frequency spectrum of 20-20K hz that you can hear your brain can process transients over 10x faster than the 44.1khz sampling rate of redbook. This is one of the key aspects of trying to create the missing audio informationā¦
Skeptics may argue that Chordās implementation doesnāt create differences that can be heard, that it doesnāt sound better or that it doesnāt improve the sound stage ā those are valid subjective arguments ā but cannot argue that it is impossible to create missing information in sampled audio that is critical to the listening experience. Mathematics has proven that is not true.
2 Likes
While I know little to nothing about this topic, I would think upsampling can create missing information by interpolation between two data points. I guess the created data point may or may not be equal to the missing data point, but should be close.
According to this article, it will work:
with the caveat that you have to use the āARMv7 AutoScript versionā, whatever that is.
1 Like
No, the Nyquist-Shannon theorem only proves you can reconstruct a signal that differs from the original by at most an error signal that is outside the band limit. It is not guaranteed to be identical, e.g. if you take a square wave and reconstruct it, you will get a Gibbs phenomenon, just above the 22kHz or whatever half the sampling frequency is. That will be true whether you upsample it or not.
The point is, upsampling will not magically recreate the portion of the signal that was lost when we applied the low-pass filter at half the original sampling frequency. Some interpolation algorithms look/sound better than others, however, and may outperform whatever sample rate conversion the DAC used to convert to its internal sample rate.
1 Like