Testing audiophile claims and myths

[Vamp898]

As an bonus, why not, i used the MP3 as a reference (which is closer to the original) and made a delta with the AAC Version. Here is the difference in linear and log as waveform and spectrogram






So from my personal Test with the 96/24 Song "Let it die" from ReoNa i came to the following conclusions

• I can hear the difference in an ABX with the MP3
• I can hear the difference in an ABX with the AAC even better/faster/easier
• Listening back to the Delta shows an impressive difference for the MP3
• Listening back to the Delta shows an even bigger difference for the AAC. The Delta contained enough information that shazam detected the song (this was not possible with the MP3)
• MP3 stops at 20kHz
• AAC wastes tons of resources to preserve information up until 23kHz. Also it wastes resources in preserving information below 20 Hz
• I will now buy used CDs of everything i own in AAC. Luckily not that much

Both MP3 and AAC were using 320kbps fixed bit-rate

 

[castleofargh]

The super generic answer with tunnel vision: A DAC will increase the sampling by filling in samples of the last value, or more likely with zeroes, and count on a filter to remove the extra frequencies born out of that.
A computer allows you to calculate a specific value for each new sample. It reduces the consequences of having a poor filter I guess. And of course, it has very real impact with a NOS DAC, as not oversampling a typical PCM file is a mistake in the first place.

More realistic answer, it depends on the DAC, how it works, what it can accept and if it will keep it that way. I was told long ago that the DAC upsampling/oversampling was part of the reclocking and anti jitter operation, but now a few very vocal people are all about forcing a sample rate that, they hope, will stop that process of their DAC.

The only measurements I've seen are from the guy selling HQplayer, which is better than vague empty marketing statements, but also really not ideal. I'm not suggesting he's feeding us false measurements, I'm saying I don't put my trust in a unique source of data who happens to sell the product.

 

[castleofargh]

@Vamp898
Showing that there is a difference in data in a format that exists specifically to remove data and is called lossy, is one of those ”thank you captain obvious” moments.
The entire thing is based on our ability to not notice certain sounds in a certain context. I’m not saying nobody can hear that a file is lossy. I’m saying your logic for making those measurements is wrong.

 

[Vamp898]

@Vamp898
Showing that there is a difference in data in a format that exists specifically to remove data and is called lossy, is one of those ”thank you captain obvious” moments.
The entire thing is based on our ability to not notice certain sounds in a certain context. I’m not saying nobody can hear that a file is lossy. I’m saying your logic for making those measurements is wrong.

I was not showing that there is a difference but how big it is. That is an small but essential difference.

You might not here if there is an difference at all, if its small enough. So if you are able to perceive it solely depends on how big it is. Hence i made measurements showing exactly this.

 

[Vamp898]

I was notified that there are quite big differences from encoder to encoder and to verify that, i got the song from a different source (and yes, i did pay for it, a second time) and there is actually a big difference.

Exact same song, both are 320kbps fixed-rate AAC, but using different encoders




As you have no influence over what decoder the platform uses where you get your AAC from, i'd just stay away from lossy at all

 

[bigshot]

A balloon full of air is bigger than a deflated balloon, but it's still a balloon.

If songs come from different sources, the likelihood that the mastering is different is high. That has nothing to do with the format.

 

[Vamp898]

A balloon full of air is bigger than a deflated balloon, but it's still a balloon.

If songs come from different sources, the likelihood that the mastering is different is high. That has nothing to do with the format.

They came all from the same source.

I bought the FLAC at レコチョク, i then converted it to MP3 and AAC using LAME/ffmpeg

As someone mentioned, there are differences in the encoding depending on the encoder (especially ffmpeg) as these are proprietary formats. So i bought the AAC Version from レコチョク (the same place where i got the FLAC) and compared it.

Converting the FLAC from レコチョク into an AAC vs. downloading the AAC Version from レコチョク gives different results. So they all came from レコチョク, they just used different encoders (although, of course, i have no knowledge if レコチョク changed the AAC in any way except for just converting it. It could be a different master).

But as converting the FLAC myself using common tools gives the worst results as it seems and we can not ensure the downloaded Version from the Portal is clean either, at least with the tools i have, there seems to be no way to get an transparent AAC out of this song.

And yes, i also tested just opening the FLAC and exporting it as AAC with highest Settings in Audacity, the results are pretty much the same. So it doesn't matter if i use ffmpeg myself or a Tool in between.

 

[bigshot]

I think the problem is the encoder you’re using. AAC should be transparent. I use a Mac. If you want to hand me your FLAC, I’ll convert it to AAC for you and it might be better. Just PM me a download link.

 

[Brahmsian]

I’m here to talk about ways to improve the sound of home audio systems.

To that end, two audiophile claims that might cause people to spend extra money.

1) Switching from a MacBook to a Mac Mini will improve the sound because there is less noise, particularly noise associated with having a screen.

2) In general, you get better sound from planar magnetic headphones if you give them more current, so they will sound better if you run them on the full output of a speaker amp. Lower impedance=more current.

 

[bigshot]

1) I’ve had iMacs, Mac Towers, MacBooks and Mac Minis, dating all the way back to the 8500AV. I’ve tested every one and they’re all audibly transparent.

2) That’s right in general, but the Oppo PM-1s are the exception to the rule. They were designed to not need amping. I don’t know about the full output of a speaker amp though. It probably depends on the model you’re using.

My advice to efficiently improve sound quality…

When it comes to DAPs, amps and DACs, there is no reason to spend a lot of money. And focus on the transducers first, then match the amp to that. Pay attention to usability. A system that is simple to use and compact will melt into the background and make it more pleasant to listen to music. The best way to improve the sound of your system is to listen to better music.

 

[Vamp898]

To that end, two audiophile claims that might cause people to spend extra money.

1) Switching from a MacBook to a Mac Mini will improve the sound because there is less noise, particularly noise associated with having a screen.

2) In general, you get better sound from planar magnetic headphones if you give them more current, so they will sound better if you run them on the full output of a speaker amp. Lower impedance=more current.

If you give them more current, they will sound louder (that is how electricity works, ohms law).

Louder doesn't automatically means better.

 

[BS5711]

If you give them more current, they will sound louder (that is how electricity works, ohms law).

Louder doesn't automatically means better.

At some point the lack of current (for very low impedance planars) can result in distortion which will certainly adversely affect the sound quality albeit at high (but otherwise usable) volume. Some low powered amps seem to just run out of puff but don't actually result in distortion they just don't get loud.

I have a set of DCA Aeon Noire planars that are 13 ohms and 92db/mw sensitivity. They can be driven to adequate listening volume and sound good off pretty much anything including an Apple Dongle. However, when pushed to high volume, the kind you might want to use for a few moments on that favourite part of that favourite song, the current requirements created by the low impedance and somewhat low sensitivity can create problems.

If pushed hard they will drive many amps into current limit protection with momentary or complete shutdown and cause distortion with other amps which obviously makes them sound horribly harsh.

Driven off a fairly powerful amplifier that puts out maximum power into low impedance loads they sound great at any volume beyond what one can manage for but a moment. Best amplifier I used them with was a Schiit Asgard 3 which has 5 watts into 16 ohms and they simply don't distort even at volumes that are unlistenable for more than a couple of moments. The amplifier and headphone electrical properties are nicely compatible and there is no way to demonstrate a technical deficiency at any volume.

They will shut down my iFi Diablo because it can't output the current needed without the assistance of a robust power supply connected which bypasses the battery. They sound good but the technical limitation is very easy to demonstrate if pushing the volume high, they will get loud and then the unit will shut off. Conversely the Diablo works incredibly well with my HD600 because the 300 ohm load needs voltage swing which the Diablo can handle easily at any volume and at any gain setting because in that pairing the electrical properties are nicely compatible. The small battery powered Diablo performs better with the HD600 than the mains powered Asgard 3 because of the design of the amplifier.

I doubt any headphone needs a speaker amplifier beyond the handful of extreme outlier cases that are extremely low sensitivity and even then it is hardly likely to be truly essential.

TLDR - in some cases and at some level of still usable volume low impedance planars do indeed sound better with an amplifier that can output a lot of current, better if only only due to lack of distortion.

 

[gregorio]

So if you are able to perceive it solely depends on how big it is.

That assertion is incorrect and is the “fatal flaw” that invalidates your posts on the subject of lossy codecs. MP3 and AAC are lossy codecs which achieve their reduction in data by employing “perceptual coding”, as castleofargh basically tried to explain. Relatively high amplitude signals (within a mix) can be totally inaudible due to human “auditory masking”, so those specific signals are eliminated by the codecs. So, you end up with a “difference file” containing a wide range of frequencies that is trivially easy to hear on their own but as part of the original mix are inaudible. I suggest you start by reading the Wikipedia page on “Auditory Masking”.

G

 

[Vamp898]

That assertion is incorrect and is the “fatal flaw” that invalidates your posts on the subject of lossy codecs. MP3 and AAC are lossy codecs which achieve their reduction in data by employing “perceptual coding”, as castleofargh basically tried to explain. Relatively high amplitude signals (within a mix) can be totally inaudible due to human “auditory masking”, so those specific signals are eliminated by the codecs. So, you end up with a “difference file” containing a wide range of frequencies that is trivially easy to hear on their own but as part of the original mix are inaudible. I suggest you start by reading the Wikipedia page on “Auditory Masking”.

G

I am not sure if this was unintentional, but the article you sent literally says, that i am right and my observations are correct.

Maybe you should have read it first yourself...

 

[gregorio]

I am not sure if this was unintentional, but the article you sent literally says, that i am right and my observations are correct.

It literally says that you are wrong and your observations are invalid!

Maybe you should have read it first yourself...

Maybe you didn’t read all of it or maybe you did but didn’t fully understand it?

G