Audible Performance Differences in HiFiBerry Digi+ and Digi+ Pro/Digi2

[changer]

You have nerve Michael, stepping into my thread an giving me a roll eyes, on my weekend when I spend the rest of the day laboring. Funny guy

I didn’t ask you if you wanted me to be your service supplier, file me a contract and I will decide whether I spend my time working for an anonymous dude on the internet.

I give a damn, I was trying to solve my problems here an I did, get lost or be more gentle.

 

[changer]

Other than you know to actually understand why there is a difference so you can make sure that you always get the benefit in the future.

@Tangband posted some BS earlier claiming differences between USB bridges but wasn't even using the same source material and hadn't done the most basic controls like level matching.

If you think there is a difference fine, but telling people that you prefer one over the other without any actual investigation just adds noise and is completely worthless.

Michael

Spare me!

 

[thorvat]

@Tangband posted some BS earlier claiming differences between USB bridges but wasn't even using the same source material and hadn't done the most basic controls like level matching.

Taking into account how USB bridges actually work the only audible difference between 2 devices can only come from enourmous ammount of jitter and that wouldn't sound anything similar to what folks who allegedly hear the difference are describing.

 

[mdsimon2]

Taking into account how USB bridges actually work the only audible difference between 2 devices can only come from enourmous ammount of jitter and that wouldn't sound anything similar to what folks who allegedly hear the difference are describing.

I've definitely seen level differences with digital output products, not huge but +/- 0.5 dB is not uncommon.

Obviously this assumes there is actually a difference and not just sighted bias. I like to give people the benefit of the doubt when they report audible differences but I have yet to see a case where an audible difference cannot be explained by a measurement.

Michael

 

[BDWoody]

Taking into account how USB bridges actually work the only audible difference between 2 devices can only come from enourmous ammount of jitter and that wouldn't sound anything similar to what folks who allegedly hear the difference are describing.

The amount of jitter needed before it becomes an audible issue is much higher than most believe. Clocks are the new place to hide the woo apparently.

what does jitter ACTUALLY sound like?

In a previous post we had an encounter with jitter in audio. The talk about jitter is hard to miss these days. There are plenty of products on the mar...

www.sereneaudio.com

 

[thorvat]

The amount of jitter needed before it becomes an audible issue is much higher than most believe. Clocks are the new place to hide the woo apparently.

what does jitter ACTUALLY sound like?

In a previous post we had an encounter with jitter in audio. The talk about jitter is hard to miss these days. There are plenty of products on the mar...

www.sereneaudio.com

Exactly.

Excellent link to jitter examples!

 

[Tangband]

Other than you know to actually understand why there is a difference so you can make sure that you always get the benefit in the future.

@Tangband posted some BS earlier claiming differences between USB bridges but wasn't even using the same source material and hadn't done the most basic controls like level matching.

If you think there is a difference fine, but telling people that you prefer one over the other without any actual investigation just adds noise and is completely worthless.

Michael

I DID use the same source material - Apple lossless. I found that my Yamaha wxc50 did SRC to 48 kHz in a rather bad way.
All listening was done with my Genelec 8340 with digital signal.
There was also audible differences between douk u2 , skysong xmos208 and AirPlay with the same source - a MAC M1 computer with Apple Music lossless.
So no BS , its you thats talking nonsense.

Please remember we are talking spdif here. Read stereophile:s review of Karik/ Numerik :

Linn Karik/Numerik CD player Page 5

Without hearing the effect of the separate sync cable, one would not suspect something was missing. But plugging it back in returned the magic and threw into sharp relief the Linn's best attributes. These impressions were repeated with a variety of music. It's tragic that a separate sync cable...

www.stereophile.com

Linn Karik/Numerik CD player In Sync!

Sidebar 1: In Sync! The Linn CD player completely sidesteps a problem that plagues all other CD transport/processor combinations: jitter in the recovered clock. Linn's method is so simple I'm surprised no one else has done it before. In fact, I'm surprised that the S/PDIF (Sony/Philips Digital...

www.stereophile.com

 

[mdsimon2]

I DID use the same source material - Apple lossless. I found that my Yamaha wxc50 did SRC to 48 kHz in a rather bad way.
All listening was done with my Genelec 8340 with digital signal.
There was also audible differences between douk u2 , skysong xmos208 and AirPlay with the same source - a MAC M1 computer with Apple Music lossless.
So no BS , its you thats talking nonsense.

Please remember we are talking spdif here. Read stereophile:s review of Karik/ Numerik :

Linn Karik/Numerik CD player Page 5

Without hearing the effect of the separate sync cable, one would not suspect something was missing. But plugging it back in returned the magic and threw into sharp relief the Linn's best attributes. These impressions were repeated with a variety of music. It's tragic that a separate sync cable...

www.stereophile.com

Linn Karik/Numerik CD player In Sync!

Sidebar 1: In Sync! The Linn CD player completely sidesteps a problem that plagues all other CD transport/processor combinations: jitter in the recovered clock. Linn's method is so simple I'm surprised no one else has done it before. In fact, I'm surprised that the S/PDIF (Sony/Philips Digital...

www.stereophile.com

If you are using AirPlay on some and not others it’s not the same due to AirPlay resampling.

Again, I very much believe there may be audible differences but think they can be explained by measurements.

Michael

 

[phofman]

Maybe until HiFiBerry fixes the problem with the Pi4, ...

If the problem is the missing reset feature, they will hardly fix it remotely. Have you checked the board, where the pin 6 of WM8804 is connected?

 

[changer]

Can someone point me to a high quality test signal, like a sweep or the correct sort of noise, as an audio file?

I was thinking of doing a gated measurement as usual, but from the RasPi's playback. I do not know if this is possible, probably REW does not like it when I do it all manually, but I thought it could be the best way to satisfy the demand for data.

While I am at it again: After I had enabled ALSA 10 band equalizer from within piCorePlayer the first time, I noticed something like clipping when the source (music track) was high level. I first thought I had to find a place where I could, after I had slightly boosted some bands, lower the total level. But it went away without further actions on my behalf and I happily listened at 100% digital input level.

After the last listening orgy for this thread, starting with listening to the aforementioned recording "Good Morning Midnight", I was again noting some sound degradation again when playing at 100% digital input and interatively reduced it to 90%. At this level it is gone, according to my perceptual apparatus. I cannot tell for sure if the absence of this "clipping" or "overdrive" was there in the first place, or if I had simply not noted it at the beginning, but when it set in, it was very "there". At least this is the order of the story how my brain put things together for now.

Is this reasonable, that this occurs? I first thought I could only notice it with the Pi3 with Digi+ Pro, but now I found out it is also the case with Pi4 with Digi+ Transformer.

 

[mdsimon2]

I use Audacity to generate test files (white noise, sine waves, etc). A RPi4 will also run REW, you can do this even if you do not have a GUI installed on your RPi by using X11 forwarding (add -Y when you ssh in to your Pi) and is a great way to run a variety of test signals.

An acoustic measurement will be of limited value as it is going to have too much variability. You are better off using a DAC other than your Hypex plate amp and making an electrical measurement of the DAC output (with each Digi HAT as input) with your 2i2 (or other higher performance ADC).

Honestly the absolute best thing to do is use an audio interface (or a computer with a SPDIF input) to directly measure the digital signal from the HATs. Old mac minis (2012, 2014) have this capability as do countless audio interfaces.

Michael

 

[changer]

I understand this, but as I had previously stated, I have no interfaces that could pick up the signal and I don`t mind getting one, as it is of secondary concern to me.

Thanks for the tips on generating test signals.

 

[phofman]

E.g. https://gist.github.com/ideoforms/d64143e2bad16b18de6e97b91de494fd#file-sox-cheat-sheet-sh-L53

 

[changer]

Where to put the command lines?

 

[phofman]

e.g. https://stackoverflow.com/questions/17667491/how-to-use-sox-in-windows

 

[Tangband]

I understand this, but as I had previously stated, I have no interfaces that could pick up the signal and I don`t mind getting one, as it is of secondary concern to me.

Thanks for the tips on generating test signals.

Static test signals dont measure timing errors ( jitter ) and are very different from real music.

 

[mdsimon2]

Static test signals dont measure timing errors ( jitter ) and are very different from real music.

It is important to get the simpler stuff right before worrying about more difficult issues. We’ve already seen issues with clipping that would have been easily identifiable with test tones.

If you take effort to get the basics right (matched output level, matched frequency response, no gross distortion or clipping) and you are still able to tell a difference then we have something interesting, until then worrying about jitter is just FUD.

Michael

 

[changer]

After my latest clipping discovery, I had to work a lot with the pCP equalizer, because it seems to never save the exact values that I set it too. I got suspicious if it was set to the same values when I was comparing and disabled it in both streamer-servers, because I could not open the alsamixer interface via ssh via alsamixer -D equal to check the exact settings. The equalizer settings might have been slightly off, as the sound is now very close to one another, if not the same. I figured that the sound difference before must have been measureable with a measuring microphone. It can as well be the case that both devices where actual set to the same equalizer settings.

Here is why:
I tried to verify my previous impression, and I have not gained certainty. But it still seems to me that the sound of the Digi+ Pro is more accentuated, i. e. the bass notes have more Gestalt, form, their presence is a bit more defined but this makes the overall playback better sounding anyway. When I was running a cheap Behringer A800 amplifier to my previous speakers, it had an incredible authority over the bass, compared to an 75w/8ohm AB amp that I also had standing around. What is the sound of the Digi+ Pro to me is like this surplus in authority, but in refined.

I will keep doing my sighted and biased comparisons until I get tired of it and you can happily comment on this deficient method, I’d be glad.

 

[changer]

My last step was testing the Digi+ Transformer S/PDIF HAT on the Pi3. You can call me a Golden Ear now, if you like.

The difference is subtle but it is there, and I am willing to risk not getting refunded for my late returns, and will live without a case for the streamer-server for some time. The difference in sound quality is: 1) openess. Listening to Charlie Haden & Pat Metheny's "First Song" from Missouri Sky, it is apparent in the first twenty seconds, the harmonics of the guitar strings reverberated in the studio give a different impression of space. 2) bass controll, well, this was a very bass heavy album and the Digi+ Transformer left a good impression. When listening to Steely Dan's solo in "Josie" the other night, it was apparent the bass had more "attack" and was at the same time more compact.

I cannot prove any of this as I am an amateur, have no tools to do it and will be fine anyways. But FYI, I have build a set of speakers following all objectivist rules and so far, measurements where a very good help.

Do the clocks make any difference? Do harmonics get shifted, if source sample rate and clock sample rate are not the same?

 

[Tangband]

My last step was testing the Digi+ Transformer S/PDIF HAT on the Pi3. You can call me a Golden Ear now, if you like.

The difference is subtle but it is there, and I am willing to risk not getting refunded for my late returns, and will live without a case for the streamer-server for some time. The difference in sound quality is: 1) openess. Listening to Charlie Haden & Pat Metheny's "First Song" from Missouri Sky, it is apparent in the first twenty seconds, the harmonics of the guitar strings reverberated in the studio give a different impression of space. 2) bass controll, well, this was a very bass heavy album and the Digi+ Transformer left a good impression. When listening to Steely Dan's solo in "Josie" the other night, it was apparent the bass had more "attack" and was at the same time more compact.

I cannot prove any of this as I am an amateur, have no tools to do it and will be fine anyways. But FYI, I have build a set of speakers following all objectivist rules and so far, measurements where a very good help.

Do the clocks make any difference? Do harmonics get shifted, if source sample rate and clock sample rate are not the same?

Yes, the clocks can make a differense when listening to music and using spdif. Trust your ears.