Brand of SD Card used makes a difference in music quality

Hi all, yep, it is unbelievable, but I tried different cards on different RPI’s.

Some SD cards sound incredible flat with Samsung, worst is scandisk. For me toshiba is the best for me atm.

I found out when checking out the new RPI4, sound was nasty, very flat with a new install of volumio on a scandisk 32GB. I plugged my favorite toshiba from my RPI3 in and the sound was much better. As a sidenote the RPI4 still did not live up to my RPI3B+.

Anybody can share this impression and perhaps even has an explanation.

Thanks

Sorry friends, for me this is sounding quite “esotheric”.

I’m testing with over 30 Raspberry PI’s and different soundcards (98% of them I2S). I do setups nearly every day on around 5 diifferent brands of SD cards and absolutely cannot confirm any SD card related difference in sound quality.

btw. i never heard of this “scandisk” brand…

Best Regards
Josef

1 Like

Sorry, SanDisk, was a typo.

Well, what can I say, I know I had the same approach when starting with hardware for sound transport. For me it was always 1 and 0, is this information not there, then it will re-read etc etc…

I started with an old netbook and Daphile, the netbook broke and I used an old laptop I had, again with Daphile. Man, this was a difference already. i was baffled.
I used that laptop and put Volumio on it, and yet again, could get a better sound.
Then i changed to RPI with Volumio, again, better sound quality.

In this process I came across the SD Card <-> sound issue. Playing around with different Volumo versions I used different SD-Cards for my convenience, and they sounded different in the same RPI.
Yesterday I was playing around with the sound difference between a RPI3 and a RPI4. The RPI4 had a new install on a Samsung EVO Plus class10 with the latest Volumio, the RPI3 the same Volumio with my old Toshiba SD Card.

The RPI4 sounded sharp compared to the RPI3, so I just put my Toshiba in the RPI4, suprised it was booting flawless the RPI4 and the RPI4 sounded so much better.

The RPI4 and RPI3 is feeding a very nice sound system via a high quality USB DAC, so everything in the sound chain is the same, it is just the SD-Card that has been changed…

That’s what I can report…

strange. :confused:
everything from volumio should be in RAM.
maybe an issue related to the USB port - not to I2S…
What mSD Card did you use? expecially the toshiba? Toshiba EXCERIA M302-EA 32GB?

2 Likes

I am using a audiophile external DAC with usb cable connection, not an I2S hat on the RPI.

The best sounding SD-Card I am using currently is the Toshiba Exceria 16GB micro SD HC I U1.

I used really fast SD-Cards for testing and it does not seem to be related to the R/W speed. It would be interesting whether the way Volumio has configured the eg swapping that makes the difference.

I remember back when I was using Daphile. Daphile had the option to load the whole track into the RAM and was playing it from there and this made a difference to the sound quality.

What I am saying with this is: The way the RPI or the hardware generally is used in sound streaming can make a difference to the sound quality. It does perhaps not necessarily be the SD-Card, but how Volumio is organizing the data during processing which causes the difference.
I will have a look into the underlying Linux version Volumio is using whether there is a benefit to increase the space it uses on the card. At the moment, even a 16GB card is only used by half, the rest is unused which in return could give LInux more space when swapping data. Then again, mostly swapping is disabled for running the OS from the SD-Card because of the wear and tear of those cards.

Another option would be to let the RPI boot from a hard drive and let the OS run from there. This is perhaps an interesting option…

Perhaps one of the Volumio programmer could share a though about this…

But again: Yes, as outlined in the first post above, the hardware makes a difference and I think the SD-Card is no exception obviously. You will probably need a good sound system and FLAC files to hear the difference. And you also probably need a good piece of music you are familiar for years.

Hi volumiouser9,

I just ordered a “Toshiba Exceria 16GB micro SD HC I U1” and will test it against a Sandisk card as soon as it arrives.

Are you especially speaking of listening to FLACs that are stored on the SD Card or do you feel all sources are affected ?

Best Regards
Josef

1 Like

My FLAC files are stored on an external spinning USB hard disk. The whole setup HHD->RPI->external DAC->Amplifier->loudspeaker is the same, I just exchange the cloned SD Cards, boot the Pi and the sound is different.

There is nothing about an SD card that could possibly affect sound playback quality in Volumio. :slight_smile:

I’d say this is a strong case of confirmation bias at work here… :laughing:

Concentrate on the things which do actually make a difference to sound quality and don’t waste time or money on things which don’t.

At most a better SD card will give you a snappier more responsive user interface (one reason I have a reasonably fast SD card in my Pi 4) and will also last longer before eventual failure, as all SD cards used to boot Linux based operating systems with writeable filesystems eventually fail due to write fatigue.

But it won’t make anything sound better.

1 Like

+1

1 Like

Hi all,

My Toshiba Exceria 16GB micro SD HC I U1 arrived yesterday.

Test scenario 1: PI3 (LT1963 Power supply, Terraberry DAC2 (dual LT3042 power supply), NAD Preamp, DENON POA4400a Mono Blocks, Dynaudio MSP400 Speakers, (cloned) SD cards: SANDISK Ultra 8GB ; very cheap 32GB card from discounter ; Toshiba Exceria 16GB micro SD HC I U1.

I listened to my best quality well known music in FLAC HQ format and some Hi-Res test files. —> NO DIFFERENCE

But i want to know for sure and will also test with other DACs (ES9018K2M, ES9038Q2M, Allo Boss with isolator and also a USB-DAC), other AMPs (NAD, and tube monos…) and other Speakers beginning this evening. (Have some quite good stuff here :wink: )

I’m excited to see if I can hear any difference. I really guess no - but i will test and see.

Best regards
Josef

1 Like

… also no difference with Khadas tone board

1 Like

for the next step i will prepare two completely identical setups with a very good (but unsupported out of the box) ES9018K2M based DAC to be able to do a quick AB comparison while playing the same title.

@Jens: Maybe you also want to have a check on this at your next blind listening test in Hamburg. I don’t want to hold on stating this is esotheric before having tested conscientious…

@voumiouser9: Would you like to describe your setup more in detail ? What super USB DAC do you use - what Amp - what Speakers ? Maybe i’m not able to test because my listening equipment is below 50000$

Best Regards
Josef

Of course it is in RAM. And even if it wouldn’t be there would still be absolutely no difference as data on both SD cards is identical and buffering when reading SD cards would prevent any difference in jitter (which would again be inaudible even if it would happen).

The explanation is simple: as long as you know what you are listening your mind is playing tricks on you and this cannot be avoided. For that reason the only test for audibilty one can trust should be performed blindly. If you want to compare if there is an audible difference between 2 sources (any 2 different sources) listener(s) shouldn’t know what source are they listening when making comparison.

More info here: en.wikipedia.org/wiki/ABX_test

1 Like

Hi Krunok,

You are right - i solved the blind listening issue by having a toggle switch that switches between 2 sources. At the starting i close my eyes and press it as often that i really do not know on which input it has latched. When i open my eyes again after comparison i have a look where i am.

All loaded to RAM - again you are right - as an electronics engineer i know that there may (maybe) be an influence by the internal memory controller of the SD card itself - that maybe causes interference and may do some kind of bad influence on the sound quality. This is the only reason why i am going on testing this in deep…

But maybe time is wasted and everything shows up to be esotheric. Let’s see…

Please everybody - hold on buying this “magic” Toshiba card before the test staff published the final report to it.
The card costs are twice of a normal Sandisk Ultra - may be of a better reliability, but only 1 user confirmed a remarkable better sound quality.

Have a look at my first post to this topic - i will test critically - but fair.

Best Regards
Josef

1 Like

Thank you for sharing volumiouser9 and also josh for your work to verify.
i should belive you both :sunglasses:

Therefore, I think now to use the same SD card for all devices in my soundcheck.
I see no disadvantage to use only the same SD cards.

…oh no Jens

I hoped you will be the one to help finding the truth around this woowoohuwidoo.
I‘m still in heavy doubt about this „Finding“ and would still not do a recommendation nor buy another Toshiba „woowoo super SD card“

Josef

HaHaHa Josh,
you mean, i should invite for listening different SD cards?
you, volumiuouser9 and other would come to my soundcheck…
let me say, i will have enough beer here. :smiley:

By the way, did you check the USB detoxer?
https://forum.volumio.org/terra-berry-dac2-review-t7871-60.html

Hi Josef,

as an eletronic engineer myself I assure you DAC chip and analog section that follows signal on it’s analog output (I/V converter, LP filter and output buffer) are immune to such low noise as the one emitted from SD memory card controller. DAC chip itself and USB controllers (like XMOS chip) are much more complex than SD card controller and, as they are in the DAC device box), are much closer to analog circuit and yet their effects on analog signal are extremely low and certainly impossible to hear. Any component in the digital path of the signal, such as various chips in your device that runs Volumio and hosts your SD card, can only interfere with jitter but as modern DAC chips are doing re-clocking on the receiving end those effects cannot possibly be audible.

If SD card controller would have any influence on the sound quality then just try to imagine what kind of influence would microprocessor have - it is much more “noisy” than simple SD card controller!
Frankly, I can hardly imagine that effects coming from various SD cards to analog singal are even possible to measure with most sophisticated measurement instruments as it is totally unclear how they could influence the analog signal at first place.

If you are interested in the topic of measuring DAC performance I strongly suggest this site (audiosciencereview.com/forum/index.php). There you will find many myths busted as measurement revealed that various linear power supplies, USB “purifiers/cleaners/detoxers”, SPDIF coax cables etc. didn’t affected analog signal at all.

Speaking of audible limits one should be aware that even the best loudspeakers/heaphones have THD over 0.1% so any THD lower than that will anyhow be masked by loudspeaker/headphone distorion. Regarding the noise, you would be hardly challenged to hear anything below 60dB in the room or 80dB with your headphones and DAC SNR is much much better than that.

Modern electronics are doing the job so well that you can’t possibly hear the difference between them - the only way to improve sound quality is to concentrate at loudspeakers and room EQ as that is where the difference is.

Br
Krunoslav

1 Like

Many thanks for explanation Krunoslav.
I can follow you in terms of your content with the SD card, etc.
However, I see a clear difference in the sound quality of the various dac boards on the market.
Also, but only sometimes i see that a PSU matters the soundquality.
As well we were able to successfully confirm the function of recklocker / purifier in the soundcheck.
And indeed, speakers and room EQ are probably one of the most important chains in SQ.
Br.

Have the differences you are mentioning been observed in a propery conducted blind test with output levels of each DAC been properly matched to be within +/- 0.2dB difference? If the output levels have not been properly matched DAC with even slightly higher output level will always be perceived as sounding better even in a blind test.

Personally I have never seen a measurement that confirms any positive effect of any reclocker/purifier. There are plenty of measurements on ASR forum and Archimago blog so you can see for yourself. Some DACs may benefit from a high quality linear PSU but most modern DACs do that properly by themselves so no enhancement is necessary, even in the scenario when they get the power from USB port of the computer. This has also been thoroughly tested on ASR and Archimago blog so no secrets there.