Cambridge Audio EXN100 vs Eversolo A8 comparison blind test.

[Mnyb]

A valiant effort more people should try for themselves , these things are really really hard to do properly, I’ve done these informal things for myself sometimes, but would not put them online for their lack of information that translates beyond my own skull .

However if you try again you have to randomise the selection of A & B even to the helper ( wife ) so no one knows what’s playing during the test .

And be extra careful with the volume differences <0,1dB it is can’t be done with an spl meter in your sofa you need a multimeter and a test tone .

The only ABX attempt I’ve done have been between file formats then I used the random functions in my player and made the tags in the files so that title appears the same regardless of if it’s A or B made 10 copies of each and shuffled them up and skipped forward in the list taking notes..

Also regarding these products, make sure they’re no shenanigans in the background choose fast linear filters for thier DAC’s if there is a choice .if there is no choice in one product then try to set a similar filter in the product that’s has a selectable filter .

The backlash you got comes from that you tested a product category where differences are very very unlikely and you meet with disbelief .
Namely that you could differentiate these two are an indicator of method errors rather than actual differences?

 

[voodooless]

The backlash you got comes from that you tested a product category where differences are very very unlikely and you meet with disbelief

No, it’s because the good advice and clear explanations were simply dismissed, ignored or waved away with bad car analogies.

 

[antcollinet]

waved away with bad car analogies.

Is there such a thing as a good car analogy? I've never seen one - or at least not applied to audio gear. Often (as above) it boils down to :

"Dac's can't sound the same because all cars are different. An analogical equivalent of comparing apples to intercontinental ballistic missiles.

 

[francolargo]

I respect the intentions and methods. I know enough about hearing and listening to not assume the tester's results will be statistically significant without a ridiculous sample size. So it is what it is. My only *suggestion* is that an "out-of-box" evaluation does not account for the significant factor of oscillator break-in, or "settling". I would want to know how the two units compare at the 500 hour mark. "Shrill" is one area that can easily improve, and additionally, the A8 can easily swap clocks as an upgrade.

 

[ahofer]

Trust your ears. In the end that's all that matters regardless of the nay sayers and "experts

Absolutely. And the only way to make sure to trust your ears is to get your other senses out of the way - a level-matched double blind test.

 

[RandomEar]

[...] My only *suggestion* is that an "out-of-box" evaluation does not account for the significant factor of oscillator break-in, or "settling". I would want to know how the two units compare at the 500 hour mark. "Shrill" is one area that can easily improve, and additionally, the A8 can easily swap clocks as an upgrade.

Yeah, this is total nonsense. "Break-in" does not exist with semiconductor parts. Crystals/oscillators in particular have extremely tight tolerances of a couple dozen ppm (±50 ppm are common) out of the factory and stay within tolerance for decades. For comparison, 50 ppm are equal to 0.005% or a change of 0.05 Hz at 1 kHz. Humans can barely detect a 1 Hz shift at typical audio frequencies [1].

 

[francolargo]

Yeah, this is total nonsense. "Break-in" does not exist with semiconductor parts.

I modded my DAC to have a standby mode that keeps the clocks running and warm all the time. It worked really well and, yes, it took significantly less than 500 hours for the phase noise of new SC-Pure clocks to settle down. However, I tried one of Ian's FemtoMck clocks, and it was still settling after 200 hrs. Here is a published observation from Ian in Canada regarding his FemtoMck clock. "FemtoMck crystals require an extended period to achieve optimal phase noise performance. For new clocks, please refrain from serious listening until they have operated continuously for more than 24 hours. Further improvements can continue to develop over weeks or months of continuous operation."

Bottom line: You enjoy your theory, I'll enjoy MY ears, and Amir's data .

 

[ahofer]

I modded my DAC to have a standby mode that keeps the clocks running and warm all the time. It worked really well and, yes, it took significantly less than 500 hours for the phase noise of new SC-Pure clocks to settle down. However, I tried one of Ian's FemtoMck clocks, and it was still settling after 200 hrs. Here is a published observation from Ian in Canada regarding his FemtoMck clock. "FemtoMck crystals require an extended period to achieve optimal phase noise performance. For new clocks, please refrain from serious listening until they have operated continuously for more than 24 hours. Further improvements can continue to develop over weeks or months of continuous operation."

Bottom line: You enjoy your theory, I'll enjoy MY ears, and Amir's data .

See above on “enjoying ears”

Post in thread 'Cambridge Audio EXN100 vs Eversolo A8 comparison blind test.'
https://www.audiosciencereview.com/...o-a8-comparison-blind-test.70034/post-2544857

 

[Beave]

Just when I was getting used to the idea that people out there do op-amp "rolling," and now along comes the idea of clock rolling?

I modded my DAC to have a standby mode that keeps the clocks running and warm all the time. It worked really well and, yes, it took significantly less than 500 hours for the phase noise of new SC-Pure clocks to settle down. However, I tried one of Ian's FemtoMck clocks, and it was still settling after 200 hrs. Here is a published observation from Ian in Canada regarding his FemtoMck clock. "FemtoMck crystals require an extended period to achieve optimal phase noise performance. For new clocks, please refrain from serious listening until they have operated continuously for more than 24 hours. Further improvements can continue to develop over weeks or months of continuous operation."

Bottom line: You enjoy your theory, I'll enjoy MY ears, and Amir's data .

Mods, please move to Humor Thread. Thanks!

 

[Beave]

To think that an oscillator needs hours to stabilize for optimal phase noise performance is quite the leap.

The think that such "stabilization" would lead to audible changes over a period of hours is a leap akin to me jumping up and touching the moon.

And, FWIW, I had to design and test the oscillators on a lot of audio DAC-related products. I never once came across one that needed hours to stabilize. That would be a completely broken design. We tested frequency and phase noise stability over time, and over a wide range of temperature and humidity in controlled environmental chambers.

 

[francolargo]

And, FWIW, I had to design and test the oscillators on a lot of audio DAC-related products. I never once came across one that needed hours to stabilize. That would be a completely broken design. We tested frequency and phase noise stability over time, and over a wide range of temperature and humidity in controlled environmental chambers.

I can't argue with your comment regarding "broken designs" because my stuff is DIY. But my class A Aleph-J amp takes an hour of warmup to really sound its best, and the DACs too are clearly better an hour after cold startup, and slightly better still after 6 hours. All determined by ear, of course, but I would imagine that the subtle but clearly audible improvements in sound stage and 'transparency' have to live "way down in the grass" of any instrumental analysis. As a neuroscientist, I know that ears are not microphones; they include tunable filters. Roughly 75% of the cochlear neurons are not directly involved in sound transmission into the brain. It's asking a lot of instrumentation to detect the detailed parameters that the tunable ear can discern. But yes, clock rolling is a thing!
That said, I'll check out of this casual discussion. All the best to you all.

 

[Beave]

I can't argue with your comment regarding "broken designs" because my stuff is DIY. But my class A Aleph-J amp takes an hour of warmup to really sound its best, and the DACs too are clearly better an hour after cold startup, and slightly better still after 6 hours. All determined by ear, of course,

Not determined by ear but by brain, using ear *and other senses*. Try it again using *only* your ears, ie, implement proper controls into your experiments and then determine by ear what you can really hear - and what you can not.

but I would imagine that the subtle but clearly audible improvements in sound stage and 'transparency' have to live "way down in the grass" of any instrumental analysis.

So you're claiming people can hear changes that electronic instrumentation can't measure? Do you have any understanding of the capabilities of electronic instrumentation? It FAR, FAR exceeds any person's hearing ability.

As a neuroscientist, I know that ears are not microphones; they include tunable filters. Roughly 75% of the cochlear neurons are not directly involved in sound transmission into the brain. It's asking a lot of instrumentation to detect the detailed parameters that the tunable ear can discern.

Again with the claim that hearing is more acute than instrumentation when it comes to electronics. Simply not correct. At all.

I'm truly surprised that a trained neuroscientist doesn't understand the limitations and problems with sighted listening comparisons. I'm not so surprised that you don't understand the capabilities of modern electronics measurement equipment.

 

[RandomEar]

I can't argue with your comment regarding "broken designs" because my stuff is DIY. But my class A Aleph-J amp takes an hour of warmup to really sound its best, and the DACs too are clearly better an hour after cold startup, and slightly better still after 6 hours. All determined by ear, of course, but I would imagine that the subtle but clearly audible improvements in sound stage and 'transparency' have to live "way down in the grass" of any instrumental analysis. As a neuroscientist, I know that ears are not microphones; they include tunable filters. Roughly 75% of the cochlear neurons are not directly involved in sound transmission into the brain. It's asking a lot of instrumentation to detect the detailed parameters that the tunable ear can discern. But yes, clock rolling is a thing!
That said, I'll check out of this casual discussion. All the best to you all.

You cannot compare details of sound across hours of time by using your ears, because echoic memory is only a couple of seconds. Sighted listening is also heavily affected by bias and essentially useless when it comes to generating reliable comparisons. As a "trained neuroscientist", one would imagine that you would be aware of such basic facts about our brain.

This is apart from the fact that all your claims about oscillators and ears being superior to measurement instruments are total bogus and provably incorrect. You also fail to provide any evidence to the contrary - on a science-based forum. As a "scientist".

 

[asteroth]

To keep the topic on the useful side, can you try to record them both with parts of a song that you consider to sound harsh on one and not on the other?

LE: Yeah, I know that details are lost when recording and when the recording is played on different devices, but still, when something sounds so different, you could still pick it up even from a recording.

 

[ahofer]

As a neuroscientist, I know that ears are not microphones; they include tunable filters. Roughly 75% of the cochlear neurons are not directly involved in sound transmission into the brain. It's asking a lot of instrumentation to detect the detailed parameters that the tunable ear can discern. But yes, clock rolling is a thing!
That said, I'll check out of this casual discussion.

Nice try Mr. Bluster.

 

[JustJones]

Of course Clock rolling is a thing.

 

[Drakon]

I hadn't planned to add more to this so I haven't been reading responses however a day after my last post our piano tuner called to stop by for the semi annual tuning. I thought well here's an opportunity for professional input. I've also spoken with an audio engineer.

The piano Tuner is classically trained and well experienced on all manner of keyboards but played grand. He also teaches. With the machines db balanced as best I could (1kHz test tone) I played his requested piece Messiah. Then followed by The 1812 Ovature. His choice was the Cambridge for the Messiah. The 1812 was a draw.

His reason for the Cambridge jived with mine. The notes are clearer and sharper. Greater depth.

I swapped the Cambridge and Eversolo back and forth during the music and also let each play the entire piece without interruption. He never knew which machine was playing.

Yes it's subjective. Don't waste time replying. I'm not asking. But its now 2/0 in favor of England.

The Audio engineer agrees ears can be fooled but stated your ears make the final choice. Not a machine or data. Data only tells how the machine is built, it's (electrical/electronic) performance and may again may hint at how it might sound. I asked about using a multi meter to balance the outputs. He didn't feel that was required. Use a db meter and test tones.

Choosing something that doesn't please the ear just leads one down the long, potentially expensive and time consuming road of sound correction. But there are those that love tweaking i.e. tube rolling, turntable cartridge swapping etc. Eh, whatever makes one happy. Again it's their money. Their ears.

Acting on the engineers suggestions I've scheduled recording studio time at one of the better studios on the east coast. I've chatted with their chief engineer, let him know my desires and ideas. We've alloted two hours for evaluation. Two very expensive hours.

One of the reasons for the studio is the audio engineer, not theirs stated a proper evaluation has to be done on speakers with a flat responsive like studio monitors. They also have the proper equipment and expertise to evaluate and advise. I'll ask again about balancing the machines outputs.

I'll post my/their findings here and likely in a stand alone posting.

D

 

[RandomEar]

I hadn't planned to add more to this so I haven't been reading responses however a day after my last post our piano tuner called to stop by for the semi annual tuning. I thought well here's an opportunity for professional input. I've also spoken with an audio engineer.

The piano Tuner is classically trained and well experienced on all manner of keyboards but played grand. He also teaches. With the machines db balanced as best I could (1kHz test tone) I played his requested piece Messiah. Then followed by The 1812 Ovature. His choice was the Cambridge for the Messiah. The 1812 was a draw.

His reason for the Cambridge jived with mine. The notes are clearer and sharper. Greater depth.

I swapped the Cambridge and Eversolo back and forth during the music and also let each play the entire piece without interruption. He never knew which machine was playing.

Yes it's subjective. Don't waste time replying. I'm not asking. But its now 2/0 in favor of England.

After your inital post, members explained to you in detail the errors in your testing methodology and why your results could not be trusted (improper level matching, long switching time, didn't make sure to use the same settings on both streamers, only semi-blind). And now you've repeated the test and repeated the exact same mistakes? Like... why? This just adds noise to the void.

The Audio engineer agrees ears can be fooled but stated your ears make the final choice. Not a machine or data. Data only tells how the machine is built, it's (electrical/electronic) performance and may again may hint at how it might sound. I asked about using a multi meter to balance the outputs. He didn't feel that was required. Use a db meter and test tones.

Well... he's wrong. Just use test tones and a multimeter, get them to within 0.2 dB. Why use the provably worse method? And devices which measure identical also sound identical, measurements are not just a "hint". They are objective, reproducible truths. Your ears will tell you one thing today and another thing tomorrow.

Choosing something that doesn't please the ear just leads one down the long, potentially expensive and time consuming road of sound correction. But there are those that love tweaking i.e. tube rolling, turntable cartridge swapping etc. Eh, whatever makes one happy. Again it's their money. Their ears.

It's actually pretty much the opposite: With devices like speakers, where there's a measurable difference, you need to like the sound. Agreed. With devices which are audibly transparent, like the vast majority of DACs (including those in streamers) and many amps today, you will waste thousands of dollars chasing "better and better sound" while you actually change absolutely nothing about the performance of your system. Choosing something which is provably transparent by trusting your ears is a fools errand and will lead you down a long, expensive and time consuming road of despair.

 

[Drakon]

Testing redone in a professional recording studio. For me the Cambridge gets the nod for Rock and Jazz. The Eversolo for New Age. All other music was a draw. However when I asked the audio engineer to listen as I switched music the choice was 100% Cambridge song after song by him and another listener in the studio.

Going by all this and it's superior in home performance the Cambridge gets to stay and the A8 goes packing. I'm not spending time and money tweaking the room to make the A8 "work". Seems the room is already EXN100 friendly and ready.

Bottom line machines can be set up identically and still sound differently. So speakers, amps, turntables, streamers and DACS all can and do impart their own influences on the music. Call it whatever one wants: warmth, spaciousness, harshness. It's fact and it happens.

By it's design the A8 is not pleasing to the ear in my home. By testing it and the EXN100 I've experienced how the environment and equipment affects them.

Even if I owned a recording studio I'd still choose the EXN100. Vastly superior customer support - I can actually reach a person as well as unlimited tech access by Crutchfield, better controls etc.

Again the A8 definitely gets the input/output and flexibility check. Multiple output filter characteristics choices. Nice!

Words from the audio engineer, an actual qualified authority "the ears matter". You choose with your ears. In your home. He found it amusing that there would be disagreement. He knows it's the Internet where anyone (everyone?) can be an expert. We all know those guys - a little bit of knowledge and a lot of (unqualified) advice.

Thanks again for the good feedback. I think I've gone well above and beyond. Let the "yeah but you didn't . . ." or "to do a real A/B test you have to . . ." be flung. The testing was valid.

I'll may do a separate post.

 

[antcollinet]

Words from the audio engineer, an actual qualified authority

And just as subject to perceptive bias, lack of volume matching, picking up cues from whoever is swapping the wires and so an ad infinitum - as every other human.

You are of course free to keep on testing in any way that pleases you. But your "information" to anyone else is no more meaningful than a coin toss.