Looking for advice: an upgrade from Topping D50 III (under €600)

[RandomEar]

Again, no please... The so called "solid measurements" are useful but i don't think they are enough to tell the entire "story". Yeah, sure, THD, SNR superb but, FR apart, correct me if i'm wrong (i'm sincere), we are measuring only at a specific frequency value and not anything else, right? I don't want to sound arrogant, i understand the meanings of most of the measurements but i don't have the knowledge of an electrical-acoustic engineer for sure, but what about the other elements?

It's fair to question statements. But you say yourself that you don't fully understand the measurements and have little knowledge of the science behind all that. Then why reject the knowledge presented to you here by people with a better understanding of the science behind it? Again: Asking questions: Good. Plain rejection instead of learning more: Bad.

Circuitry design, the other various components (caps, resistors, IC ecc. with their specific values, quality ecc.), both in the supply stage (this one with basic, cost-effective PSM or toroidal...), the core stage of the DAC and the Analog part?
Imho, I think that, with this mindset, we are kickin' many other things out of the 'equation'.

Why would it matter? If one device uses transistors for 0.1 € and another device uses some for 10 €, but the devices deliver the same measured performance, why would it matter which one uses which transistors? Why would a different supply stage matter if the measured, proven result of both is the same? If anything, the device using the cheaper 0.1 € parts seems to have a more capable designer, because he managed to achieve the same performance at a lower cost. Unless you have a deep understanding of electronic circuits and all their pitfalls, why worry about this at all?

If you're looking for reliabilty or longevity, it may make sense to go with a more expensive part from a reputable manufacturer. But for pure audio performance, which - again - can be measured and proven to be identical, why pick the more expensive solution? There is no logic in this.

It's no placebo, i can guarantee, i perceived micro to almost macro differences between the three DAC, just one example: the D90 has an easy audible better dynamic than the D50 III and SMSL Raw-dac 1 (listened carefully, with the same cans, same connection, same audio track segments ecc.).

Comparing audio devices is non-trivial and most people trying to do it set up severely flawed comparisons and draw the wrong conclusions from them. That's what we're trying to tell you.

• Did you switch between devices within les than 3 s? Because our echoic memory is extremely short and any longer switching time will make the comparison useless.
• Did you level match your devices to within 0.2 dB? If not, volume differences alone can easily explain any subjective difference - especially stuff like "dynamic range". Playing music louder will make it sound better.
• Did you play both devices through the exact same signal chain?
• Were you in the exact same listening position (if using speakers)?

These are some but by far not all aspects you need to consider to come close to a valid comparison for audio equipment. And bias is added on top of it all. We're all affected by bias, no matter if we're aware of it or not.

Do you think that i wanted to ended up spending €999 on a DAC? Hell no! But, after extensive searching by Article, forum (including this of course) and YT reviews (sadly, most of the audio related yt channels don't seems so reliable, in part or at all... sponsor money? i just saved a couple, one is "Reviews By WaveTheory", i guess he's reliable, for now), i couldn't find a decent and appealing substitute in the €600 (living in Italy i'm forced to choose products from Europe and from Amazon in particular, for their convenient return and refund policy, S.M.S.L doesn't offer much choice and i wanted to avoid Topping with ESS... so, gathering as much info as i could, narrowing the choices, i opted for the D90 III D.
(Edit: words corrections and others)

No, I don't think you want to spend more money. But it's a common logic amongst subjective HiFi enthusiasts that more expensive audio devices must sound better. After all, why else would they be more expensive? However, past a certain and surprisingly low amount of money, this isn't true anymore for most product categories - maybe with the exception of speakers.

 

[somebodyelse]

1. The system for comparing has a major flaw: we can't select a portion of the track, for both A,B and X (yes i know how it works, you must choose what is X), at our will (start and end point), as we can do in a DAW; this is a flaw because of the limitation in time of the AWM (Auditory working memory). Value may vary: there are many factors that affect the duration, but, according to different studies, the amount is within few sec. (4 sec, 2 sec. e.g.). So, without an optimal situation like that i described before, in a DAW, it's hard and annoying

That would be a flaw in a specific ABX implementation, not in the method in general. Some software implementations do let you select a portion of A, B and X.

2. In a compressed file, even at max bitrate, something is discarded: if i remember well, high frequencies above a certain threshold. This may affect, in a subtle way, harmonics and even dynamics (based on the psychoacoustics phenomena that bind frequency and perceived loudness, see equal-loudness contour). According to the database (this one) that gather infos on dynamic range of plenty of recordings (with media and codecs info), there are several of these with poor DR rating. So, what i'm trying to mean, are we sure that the (allegedly) lossles recordings or the original source went, at least in some portion, beyond that threshold and that the dynamic range wasn't bad? Because if not, it's very hard to distinguish, or a waste of time imho

I'm not sure whether you are confusing dynamic range compression with data compression - two very different processes that happen to share a common word. Dynamic range compression changes gain based on some measure of the signal level over time - loud bits have the gain turned down and quiet bits have it turned up, so there ends up being less difference in volume between the loud bits and quiet bits. Frequency response is not changed. Data compression involves either a more efficient way of encoding the data (lossless), or a reduction in the amount of data (lossy).

What is discarded (if anything) depends entirely on the data compression format. For lossless compression like FLAC nothing is discarded - the clue is in the name, and you can do a binary comparison before and after to verify. For lossy compression some information is discarded, generally that which is least likely to be audible based on some psychoacoustic model. That could include stereo information, masked frequencies, high frequencies and other things depending on the model used in the format. Dynamic range is not usually changed.

Different dynamic range for recordings on dr.loudness-war.info usually relate to different releases using a different mix or master - the studio has chosen to release an altered recording. Usually in regard to the loudness war this means later releases of old material having dynamic range compressed. Occasionally (like Metallica's Death Magnetic) we see the reverse - the original was heavily compressed and had clipping, while the later release was less compressed after fan complaints about the original.

 

[tccalvin]

Yes, i've known it for a while... Have you ever wondered why it's so difficult to perceive the differences? I've come to some conclusions

I think the simpler explanation is that our ears are just not as sensitive as we believe them to be.

This knowledge is what allows me to enjoy music with a MacBook DAC + Spotify + 150€ Headphones. I now know that spending more won't give me any real benefit.

That's not to say that my search to improve my listening experience is over. I'm still learning about optimal gain levels to minimize distortion/noise, components with design defects, room correction etc.

The important part (at least for my own peace of mind ) is verifying my experience with the scientific method.

 

[bric-out]

That would be a flaw in a specific ABX implementation, not in the method in general. Some software implementations do let you select a portion of A, B and X.

Maybe I didn't explain myself well, i didn't wanted to state that the ABX method is flawed in general, it's that particular implementation, as you pointed out.

I'm not sure whether you are confusing dynamic range compression with data compression - two very different processes that happen to share a common word.

No, i wasn't confusing at all, i know the difference.

What is discarded (if anything) depends entirely on the data compression format. For lossless compression like FLAC nothing is discarded - the clue is in the name, and you can do a binary comparison before and after to verify. For lossy compression some information is discarded, generally that which is least likely to be audible based on some psychoacoustic model. That could include stereo information, masked frequencies, high frequencies and other things depending on the model used in the format. Dynamic range is not usually changed.

i didn't specify well but i was talking about lossy compression, not lossless one like FLAC, ALAC or APE.

 

[wilhelm2015]

upgrade your amp not dac ...

 

[bric-out]

upgrade your amp not dac ...

I can't deny that this tought crossed my mind for a while. Now i'm quite happily enjoying listening with my current setup, in the future, maybe, when i'll collect further data and info, maybe...