[charleski]

If I'm reading the posts correctly, in his first test he down-sampled the 24/96 original to 16/44 with foobar, but didn't then up-sample that back to 24/96. But then went back and used dbPoweramp to down-sample and then up-sample for the second. Is that correct?
 
There's another question lurking here which might be a bit more interesting than the old 'Can I hear a difference with hires' chestnut. And that's whether the DAC produces audible differences when fed with hires vs std signals. The Modi 2 uses an AK4396 DAC which, like most modern DACs, upsamples everything through an interpolation filter before feeding it to a ΔΣ D/A block. As seems to be the fashion, the AK4396 allows the low-pass filter used in this process to be switched between fast and slow roll-off. While Schiit seem to be a little secretive about the filter choices they make it seems possible they set this up for slow rolloff, which runs the risk of introducing IM from aliased images when fed with 44.1kHz signals, but will be operating far beyond the audible band on 96kHz ones. On a track with a lot of high-frequency material like the one used, this might be a factor. Maybe.

 

[PDVJAM]

  If I'm reading the posts correctly, in his first test he down-sampled the 24/96 original to 16/44 with foobar, but didn't then up-sample that back to 24/96. But then went back and used dbPoweramp to down-sample and then up-sample for the second. Is that correct?

Yes, you're right.

 

[PDVJAM]

  There's another question lurking here which might be a bit more interesting than the old 'Can I hear a difference with hires' chestnut. And that's whether the DAC produces audible differences when fed with hires vs std signals.

Right now I'm more interested in 'Whether different DAC's/DAP's sound different' You know, those old tales 'Sabre sounds sooooo different to Cirrus Logic'. I have iBasso DX90 and thinking, if there any reason to buy something more 'audiophile grade' like QLS 360, L&P LP5Pro, etc

 

[RRod]

 
I was able to hear difference when I used foobar to downgrade files. I wasn't converted them back to 24/96 after downgrade, so they were mp3 320 and 16/44 flac. And original was flac 24/96.

 
Thanks for clearing that up. Something has always made me uncomfortable about the foobar black box.

 

[PDVJAM]

   
Thanks for clearing that up. Something has always made me uncomfortable about the foobar black box.

I will try new attempt with foobar, but this time I will downgrade files and then upgrade them to 24/96.

 

[sterling1]

 All,
 
Back in the 1970's when digital audio was for the most part just an idea, it was theorized  that 16/44.1 was more than adequate to get all the fidelity to our ears that our ears could handle; and, Sony told us it was perfect sound for ever.  SACD did not come along until Sony and Phillips were at the end of CD patents, telling us then that CD was not really perfect but SACD resolved the issues.  And now, interestingly, with patents expiring on SACD, Sony again leads in "Hi-Res" propaganda with a new catch phrase called DSD, to satisfy the appetites of audiophiles, telling us SACD was not the end all. Well, maybe we indeed are not at the end of real audio improvements but, seems to me, CD's still reign and therefore the development of multi-channel and the streaming of it with out hassle would be a better use of audio engineering resources. How about low cost little battey powered speakers driven directly from our laptops for multi-channel surround sound with as many channels necessary to immerse us in the experience.

 

[arnyk]

  If I'm reading the posts correctly, in his first test he down-sampled the 24/96 original to 16/44 with foobar, but didn't then up-sample that back to 24/96. But then went back and used dbPoweramp to down-sample and then up-sample for the second. Is that correct?
 
There's another question lurking here which might be a bit more interesting than the old 'Can I hear a difference with hires' chestnut. And that's whether the DAC produces audible differences when fed with hires vs std signals. The Modi 2 uses an AK4396 DAC which, like most modern DACs, upsamples everything through an interpolation filter before feeding it to a ΔΣ D/A block. As seems to be the fashion, the AK4396 allows the low-pass filter used in this process to be switched between fast and slow roll-off. While Schiit seem to be a little secretive about the filter choices they make it seems possible they set this up for slow rolloff, which runs the risk of introducing IM from aliased images when fed with 44.1kHz signals, but will be operating far beyond the audible band on 96kHz ones. On a track with a lot of high-frequency material like the one used, this might be a factor. Maybe.

 
If one studies http://www.akm.com/akm/en/file/datasheet/AK4396VF.pdf  page 11 and 12 one finds that the AK4396 has standard features that are easy enough to implement for introducing high frequency losses that can reasonably be expected to be reliably audible.  It has other similarly easily implemented  features that can give it the performance characteristics of  the objectivist's "Good DAC".  It's up to the audio component designer to make some relevant choices or expose them to the consumer.

 

[charleski]

   
If one studies http://www.akm.com/akm/en/file/datasheet/AK4396VF.pdf  page 11 and 12 one finds that the AK4396 has standard features that are easy enough to implement for introducing high frequency losses that can reasonably be expected to be reliably audible.  It has other similarly easily implemented  features that can give it the performance characteristics of  the objectivist's "Good DAC".  It's up to the audio component designer to make some relevant choices or expose them to the consumer.

Yeah, the slow rolloff filter introduces a -3dB drop at 18kHz, which might be marginally audible to younger ears. A slow rolloff will leave a reasonable amount of aliased image, though, which might generate IM products at around 6-10kHz (i.e. certainly audible) downstream in the amp, but this would depend on the amount of high-frequency energy present in the signal.

 

[arnyk]

  Yeah, the slow rolloff filter introduces a -3dB drop at 18kHz, which might be marginally audible to younger ears. A slow rolloff will leave a reasonable amount of aliased image, though, which might generate IM products at around 6-10kHz (i.e. certainly audible) downstream in the amp, but this would depend on the amount of high-frequency energy present in the signal.

 
I've done a number ABX tests with younger ears and filters that were -3 dB @ 20 KHz and -1 dB @ 10 KHz with positive (statistically significiant @ 99% confidence)  results, no cherry picking of program material or test system required.

 

[charleski]

   
I've done a number ABX tests with younger ears and filters that were -3 dB @ 20 KHz and -1 dB @ 10 KHz with positive (statistically significiant @ 99% confidence)  results, no cherry picking of program material or test system required.

 
I don't trust my decrepit old cochleas to resolve much of anything over 15kHz

 
I suspect the fashion for allowing a slow rolloff is related to the whole high res craze and the related hysteria about pre-ringing. It would be slightly ironic if this has become a self-fulfilling prophecy with DAC manufacturers designing devices that are audibly (and measurably) inferior at standard sample rates.

 

[Music Alchemist]

  Well, when I've converted files with dBpoweramp I could not hear any difference. Even between 24/96 and 320 CBR mp3

 

 
Yet another reason for me to recommend dBpoweramp!

 

[arnyk]

   
I don't trust my decrepit old cochleas to resolve much of anything over 15kHz

 
I suspect the fashion for allowing a slow rolloff is related to the whole high res craze and the related hysteria about pre-ringing. It would be slightly ironic if this has become a self-fulfilling prophecy with DAC manufacturers designing devices that are audibly (and measurably) inferior at standard sample rates.

 
No prophecy here, its already documented fact. First case of this I saw in a published test related to the Pono, but I have seen it happen since on my test bench with another commercial device. Designers are using those configuration options that I have documented!

 

[Joe Bloggs]

Screenshot of original 96kHz and various resampling algorithms. From top: original 96kHz, converted to 44.1kHz with: dBPoweramp (via new dBPoweramp engine option in foobar2000 1.3.9), Audacity 2.0.6 Best quality, foobar2000 1.3.9 PPHS Ultra, foobar2000 1.3.9 PPHS.

Note that some differences between the original and the converted are inherent to the difference in sample rate and how the spectrogram is computed different at the different sample rates. For example, the same 32768 sample Blackman window is used for the original 96kHz as well as the 44.1kHz conversions, which resulted in a wider effective window width view in the conversions, causing the start / end transient artifacts to appear wider. There are differing shapes of rolloffs, which aren't really visible in this view. The dBpowerAMP algorithm has crazy bandwidth, practically not rolling off until 22000Hz...

I will have to update to foobar2000 1.3.9 for the dBPoweramp engine option it sounds like, although the harmonics were probably inaudible using PPHS Ultra to begin with (will have to go home and listen some more)

 

[arnyk]

  Right now I'm more interested in 'Whether different DAC's/DAP's sound different' You know, those old tales 'Sabre sounds sooooo different to Cirrus Logic'. I have iBasso DX90 and thinking, if there any reason to buy something more 'audiophile grade' like QLS 360, L&P LP5Pro, etc

 If you study the scientific data about the sensitivity of the human ear to noise and distortion, as well as the relatively  large amounts of noise and distortion in commercial recordings, you will probably reach the conclusion that the excellent performance of modern electronics is generally overkill. That means that audible differences among good audio products is vanishing.
 
The basic specs of the iBasso DX90 are given as: Frequency Response: 17Hz~20KHz +/-0.1dB , Output Impedance: <0.1ohm,
THD+N: 0.0015% (32ohm load)  Output Level: 1.3Vrms(Low gain), 2.0Vrms(Mid gain), 2.8Vrms(High Gain)
S/N: -118dB +/-1dB(Low gain), -116dB +/-1dB(Mid gain), -115dB +/-1dB(High Gain) (32ohm Load)
Crosstalk: 75dB (1KHz 8 ohm Load)  which is symptomatic of this overkill.

 

[charleski]

Just out of interest, what settings are you using for the spectrograms in Audacity there? (the numbers in Edit->Preferences->Spectrograms) I'd like to repolicate them and compare to Audition. There's certainly a lot of alising and associated harmonics in the PPHS versions.