Sparkos SS3602 Opamp Rolling In Fosi P4

[MAB]

Ok.
I got my Fosi X5 a few weeks ago after reading the review about it at ASR.
Product sound very good in my system and Fosi have more products available.

Yes, it looks like a very nice little phono preamp. I am not surprised it sounds good.

Many Youtubers are talking very positive about Fosi Products at a low price.
Several Youtubers talk very positive about changing opamps, but the price
is about same level as the product itself. Actually very expensive.

Yes, this is absurd. I think you have correctly identified the absurdity.

Is there any difference in sound? Yes....is stated by several Youtubers.
No is stated by measurements. I can hear differences in sound samples.
That is why I am searching some answers to make a decision.......

How about asking New Record Day to make new files with reduced gain?

The recordings don't meet basic experimental standards, high school science fairs demonstrate better control. And most of us want to avoid speculating on what went wrong, just that these are total failures of evidence.

You have the preamp. Get some opamps. Make some recordings without messing the entire experiment up. The steps to do that are spelled out in this thread. You just learned in this thread what Deltawave is (it's the difference). Download it (don't forget to donate!), learn to use it on the recordings. Use it to establish a control and verify you are not introducing experimental error. Then use it to establish if your recordings actually exhibit measurable differences. If the do, evaluate the magnitude of those differences compared to limits of human audibility. Lastly, use Foobar's ABX comparator or some other tool that provides blind comparison capability and see if you can actually tell the difference between the files.

It all sounds exhausting. For me, the concentration applied to the blind testing is by far the most tedious. I actually don't mind the measurement part since electronics is my hobby too, although secondary to music. Using my music listening time to do shootouts and comparisons, trying to engage my friends to join me, changing setups endlessly, all became incredibly tedious years ago for me.

There is also a practical side. I have a fairly high performance preamp design that I have used for a few projects. I just built one up with opamp sockets.

I am already jostling the capacitors as I swap opamps. See those yellow ceramic caps near the socket?

The closer electrically to the supply pins are to that little capacitor, the better the performance on certain opamps with truly exemplary performance. For instance, NE5532:

That low level junk I circled in red didn't show up in my other builds as far as I looked. It may be that I never looked hard enough. I am hypothesizing it is due to the larger drop from those little capacitors to the opamp introduced by the socket. The junk is 1 or 2 dB in a circuit with nearly -110dB THD+N. At this level of performance, the socketing resistance and IR drops become important on some opamps. NE5532 is one of those race-horses that needs proper attention if you want the absolute best out of it. I will test this hypothesis at some point.

I do want to emphasize the junk circled in the above graph is far below audibility in any application. I exclude adding downstream noise amplification, etc.

In fact, if you stare closely at the above graph and the pictures of the preamp, you will see I actually have different feedback resistor values on left and right channels of the initial gain stage. This was to specifically test how much the opamp vs. how much the application matters. I'm planning on writing up a test of this preamp, with some details on actual performance difference between opamps, and the differences when changing the surrounding circuit values.

 

[oleg87]

As far as I know, based on the GR-Research video, the recordings were done at New Record Day.
New Record Day has been a Youtuber for many years and is still going strong.
Can I trust this Youtuber and the information he is providing?

If he can't be bothered to get the absolute basics right, obviously not. Most audio youtubers are grifters and morons.

 

[lini]

A quarter, not 'four times lower', it's a division, not a multiplication.
Sorry, got my pedant hat on.

Sorry, but you're wrong.

I.e., "the distortion of B is four times lower than the distortion of A" would usually be supposed to mean, that the distortion of B equals one fourth (= one quarter or respectively 25 %) of the distortion of A. Whereas "the distortion of B is one quarter lower than the distortion of A" would rather mean, that the distortion of B equals the distortion of A minus one quarter of the distortion of A, hence equalling three quarters (or respectively 75 %) of the distortion of A.

Where it gets more difficult: "The distortion of B is four times higher than the distortion of A." Because that's not quite the same as "the distortion of B is four times as high as the distortion of A". So while the latter would simply mean "distortion of B = 4 x distortion of A", the former might rather mean "distortion of B = distortion of A + 4 x distortion of A = 5 x distortion of A".

I'd agree with you, however, that there's a bit of a "logic break" there - 'cause while "is four times lower than" would equal "is four times as low as", "is four times higher than" might rather not equal "is four times as high as".

Greetings from Munich!

Manfred / lini

 

[MaxwellsEq]

Ok.
A lot of information about these files are posted by experienced people.
Differences has been discovered.
Why would anyone manipulate a file and include clipping as part
of the manipulation when so many people have tools to discover
every bit in a file?


Are differences due to
1) manipulated files
2) opamp differences

Clipping has been detected.
What can cause clipping?

The product in which the opamps were tested has a variable gain button.

If the variable gain button is lowered 3dB it will be no clipping
in the file with clipping which is presented in the video?

How about asking New Record Day to make new files with reduced gain?

I recommend trying some critical thinking.

Who benefits from the two samples sounding different? The answer is the reputation of the person making the claim is enhanced. They claim there is a difference. They claim to hear it. They publish files which genuinely sound different to anybody. People now trust them to give further guidance based on their knowledge and demonstrable hearing skills. This, may be why these files are different...

Science experiments, meanwhile, are often dull because a lot of care needs to be taken to guarantee repeatability by anyone else in a different lab. Doing this experiment properly (as shown by Amir, Pavel and others) shows there are very tiny differences, which are probably inaudible.

The files shared on YouTube have (relatively speaking) massive differences between them, which is not what the science predicts. How could this have happened? Insufficient effort at doing the dull stuff like properly setting levels etc. is a reasonable conclusion we can draw. But, critical thinking reminds us - it's in the YouTuber's benefit to have files which measure wildly different and sound different. Is this deliberate manipulation? We have no way of knowing their intent.

 

[Flageborg]

Many comments on YouTube highlighting the clearly flawed sound clips have been removed or hidden since the video was posted. It does raise the question -why?

Looks like the video was published Youtube a month ago.
Some days ago I saw the video and found the files for downloading.
Downloaded the files for listening.
Visited same Youtube video now and the files are still ready for downloading.
As far as I know the files was.......and are still........ available for downloading.

Some times the Internett is ok and files behave as expected and many times not so good.
Shit happens when it comes to computers........

 

[amirm]

Why would anyone manipulate a file and include clipping as part
of the manipulation when so many people have tools to discover
every bit in a file?

They are just naive, not knowing that people use tools to analyze these files immediately. Hence, you want to do that yourself before releasing such files. They didn't. This type of machine analysis makes many online listening tests hard or impossible as people can figure out what is in them objectively. For example, if I want to see if people are biased, and release two identical files but say they are different, the public figures that out instantly by using such tools.

The solution here would have been for them to match levels, then bring a few audiophiles in, and perform the testing blind on them. And repeat a few times to rule out chance. Danny once promised us that he would get his workers to do this test type of test. He never did. Likely they couldn't pass or he was afraid to even try.

Importantly, if measurements don't show an audible difference, you better check, double and triple check your listening tests to make sure you have not made mistakes. You don't quickly run out and shout that you have found audible differences that are so easily debunked.

Finally, there is a volume control on the amp they used. It is trivial to hit that and change volume as you mess with replacing the opamp. In my case, I used my analyzer to make sure the volume was matched. They are so sure that they are right that they didn't bother to do this and hence, have their hat handed to them.

 

[Flageborg]

WHY would one alter the gain (or volume control) when comparing op-amps.
Opamps don't alter the volume... they can't. The gain is determined by components around the op-amps.

With manual gain level turned to the max the signal can get to strong and cause clipping as in wav 3-file ?
The clipping may be a result of combining 2 opamps(Sparkos + Sparkos) raising the total signal to high ?

 

[Flageborg]

How about you ask him/them and do it properly...

Yes...will send a request to New Record Day/GR-Research on this matter

 

[wwenze]

That low level junk I circled in red didn't show up in my other builds as far as I looked. It may be that I never looked hard enough. I am hypothesizing it is due to the larger drop from those little capacitors to the opamp introduced by the socket. The junk is 1 or 2 dB in a circuit with nearly -110dB THD+N. At this level of performance, the socketing resistance and IR drops become important on some opamps. NE5532 is one of those race-horses that needs proper attention if you want the absolute best out of it. I will test this hypothesis at some point.

I do want to emphasize the junk circled in the above graph is far below audibility in any application. I exclude adding downstream noise amplification, etc.

In fact, if you stare closely at the above graph and the pictures of the preamp, you will see I actually have different feedback resistor values on left and right channels of the initial gain stage. This was to specifically test how much the opamp vs. how much the application matters. I'm planning on writing up a test of this preamp, with some details on actual performance difference between opamps, and the differences when changing the surrounding circuit values.

Your junk (wait that doesn't sound right) looks like multiples of 50Hz and 60Hz. High chance at least 1 of them is due to either insufficient power supply filtering or "coupling" issues.

 

[amirm]

Yes...will send a request to New Record Day/GR-Research on this matter

I will give you their answer: "volume doesn't make a difference. you have to be deaf to not hear the fidelity difference."

 

[Flageborg]

I recommend trying some critical thinking.

Who benefits from the two samples sounding different? The answer is the reputation of the person making the claim is enhanced. They claim there is a difference. They claim to hear it. They publish files which genuinely sound different to anybody. People now trust them to give further guidance based on their knowledge and demonstrable hearing skills.But, critical thinking reminds us - it's in the YouTuber's benefit to have files which measure wildly different and sound different. Is this deliberate manipulation? We have no way of knowing their intent.

Yes......and that is why all the questions.
New Record Day has promoted this video with sound examples, but can you hear the differences?
The music files are ready for download too

 

[Ropeburn]

Sorry, but you're wrong.

I.e., "the distortion of B is four times lower than the distortion of A" would usually be supposed to mean, that the distortion of B equals one fourth (= one quarter or respectively 25 %) of the distortion of A. Whereas "the distortion of B is one quarter lower than the distortion of A" would rather mean, that the distortion of B equals the distortion of A minus one quarter of the distortion of A, hence equalling three quarters (or respectively 75 %) of the distortion of A.

Where it gets more difficult: "The distortion of B is four times higher than the distortion of A." Because that's not quite the same as "the distortion of B is four times as high as the distortion of A". So while the latter would simply mean "distortion of B = 4 x distortion of A", the former might rather mean "distortion of B = distortion of A + 4 x distortion of A = 5 x distortion of A".

I'd agree with you, however, that there's a bit of a "logic break" there - 'cause while "is four times lower than" would equal "is four times as low as", "is four times higher than" might rather not equal "is four times as high as".

Greetings from Munich!

Manfred / lini

More pedantry (leave that to us Germans lol):

The phrase "four times higher" is already highly colloquial and not accurate. It is commonly used as a substitute for "four times as high", but if you're being precise with language, it's misunderstandable already. If you start with a value of 1 (that's one time), and then it's "four times higher", what is it? Right: it's 5, not 4 logically. From a pedant's (= accurate, scientific) point of view you should've used "four times as high".

Into the negative, this linguistic construction falls apart completely. "Four times lower"? Did you start with 5, and now it's one? No, what you mean is you started with 4. Again, it's much better to use the (perfectly established) phrases "one fourth" or "a quarter" for numerical comparisons. Just so much better, and without any possible confusion.

Greetings from a Saupreiß!

 

[Flageborg]

They are just naive, not knowing that people use tools to analyze these files immediately. Hence, you want to do that yourself before releasing such files. They didn't. This type of machine analysis makes many online listening tests hard or impossible as people can figure out what is in them objectively. For example, if I want to see if people are biased, and release two identical files but say they are different, the public figures that out instantly by using such tools.

The solution here would have been for them to match levels, then bring a few audiophiles in, and perform the testing blind on them. And repeat a few times to rule out chance. Danny once promised us that he would get his workers to do this test type of test. He never did. Likely they couldn't pass or he was afraid to even try.

Importantly, if measurements don't show an audible difference, you better check, double and triple check your listening tests to make sure you have not made mistakes. You don't quickly run out and shout that you have found audible differences that are so easily debunked.

Finally, there is a volume control on the amp they used. It is trivial to hit that and change volume as you mess with replacing the opamp. In my case, I used my analyzer to make sure the volume was matched. They are so sure that they are right that they didn't bother to do this and hence, have their hat handed to them.

Ok.
I do make files for comparison between different PU's mounted on my turntables
to observe the differences and match up with different RIAA's
It is a part of my hobby playing records.
Blind testing is very interesting, but most people raving about differences do not want
others to participate in their own system to check new cables or other new products
and validate differences. I am ready to participate and get tested, but no response in my area.
The volume knob is very touchy indeed.........

Picture of some of my turntable comparison equipment
Leonard Cohen - who would know different pressings has different sound?

 

[Flageborg]

I will give you their answer: "volume doesn't make a difference. you have to be deaf to not hear the fidelity difference."

A good one........but worth a try. Worst case scenario I have to pay them a visit and find out all about Soundstage & Imaging

 

[SIY]

They are just naive, not knowing that people use tools to analyze these files immediately.

Or they understand their target market so don't care that people who aren't the target can do that. They can (and apparently do) suppress any negative comments to their promo videos, and again know that their target audience doesn't care about anything published on a science-based site. Same as faith healers, spirit mediums, and homeopaths.

"Naïve" is the last word I would use to describe them. Cynical, amoral, and skilled at deception are, IMO, more accurate.

 

[MAB]

Your junk (wait that doesn't sound right) looks like multiples of 50Hz and 60Hz. High chance at least 1 of them is due to either insufficient power supply filtering or "coupling" issues.

Ha! Good one.

I think it's exactly. It shows up predominantly on the left channel with an NE5532. I didn't see this effect with any other opamp. May be I just didn't get it socketed correctly for that measurement. Which is another reason to not roll opamps, and to make measurements of tweaks like this.

 

[Old_School_Brad]

Yes......and that is why all the questions.
New Record Day has promoted this video with sound examples, but can you hear the differences?
The music files are ready for download too

View attachment 445843

This has nothing in common with comparing opamps. Physically moving your speakers will naturally change the sound at the listening position -that’s expected. Swapping opamps, on the other hand, doesn’t alter the sound unless the surrounding circuitry “rejects the transplant,” so to speak.

 

[solderdude]

The clipping may be a result of combining 2 opamps(Sparkos + Sparkos) raising the total signal to high ?

No... gain is NOT determined by op-amps but the resistors around the op-amp which arguably do not change.
Changing an op-amp will never change the gain. For this reason the files you listen to are 'doctored'.

 

[Old_School_Brad]

I will give you their answer: "volume doesn't make a difference. you have to be deaf to not hear the fidelity difference."

That was essentially the standard response to every critical comment they received in the days following the upload.

I also noticed some variations of it, such as:
"It doesn’t matter if the volume is different -the audible difference is obvious regardless," and
"Volume doesn’t affect timbre etc.."

Both insulting and unintentionally amusing at the same time..

 

[KSTR]

I will give you their answer: "volume doesn't make a difference. you have to be deaf to not hear the fidelity difference."

Funny thing is this is a clever answer because it is not totally wrong, as usual. If you compare stuff with very significant differences, like a lot of distortion or changed stereo width (mono in the extreme case), then a few dB of volume difference don't mask the differences.

Obviously, the smaller the differences, the larger the need to have matched levels for a meaningful initial comparison followed by successful ABX or AB/HR etc. However, IME once one has nailed the difference in knowing what exactly to listen for in a successful DBT then volume levels again have some wiggle room, you still identify the artifacts/differences (of course with different DBT protocol than simple ABX).