8th generation ADC/DAC loop vs the original.....can you hear it?

[fas42]

I didn't do this test, as I had done the previous ones and posted the results from that. Out of curiosity I downloaded the Hyperion Knight piano samples here just now, and had a listen. The 8th generation copy is most easily distinguished, for me, by the quality of the transient attack of the notes - the original has an intensity, jump factor to the initial impact of the hammer on the strings, which is quite dulled by the 8th copy - as if a piece of cloth had been slipped between the hammer and string.

 

[Blumlein 88]

I didn't do this test, as I had done the previous ones and posted the results from that. Out of curiosity I downloaded the Hyperion Knight piano samples here just now, and had a listen. The 8th generation copy is most easily distinguished, for me, by the quality of the transient attack of the notes - the original has an intensity, jump factor to the initial impact of the hammer on the strings, which is quite dulled by the 8th copy - as if a piece of cloth had been slipped between the hammer and string.

Was this before or after you looked at the answer?

 

[amirm]

Amir, this kind of test I did was a triangle test. If you provide an original, and ask for people to choose between another original and the 8th generation that would be a duo-trio test. Either can be made to work and might be used for various reasons. Triangle is considered more discriminating when looking for non-specific differences. Duo-trio might be used with a specified aspect to listen for like "pick which file has the deepest bass".

I might try one with a tetrad format next. You are given two original files, and two 8th generation files and are asked to pair them up. The advantage here is if you can't pair things up you know you haven't heard it yet. I am thinking maybe try a 32nd gen vs originals. Assuming I don't go bonkers repeating this 32 times.

I understand the difference. Multi-way tests are fine when differences are large. But when they get this small, they lead to frustration and lack of outcome. I know that was the case for me. I wanted to see if there were any differences in ABX so I had to go through all the combinations on top of the boring job of running ABX.

 

[Blumlein 88]

I understand the difference. Multi-way tests are fine when differences are large. But when they get this small, they lead to frustration and lack of outcome. I know that was the case for me. I wanted to see if there were any differences in ABX so I had to go through all the combinations on top of the boring job of running ABX.

So if it is this hard 8 generations down the line would it be safe to conclude simply playing back some music is a transparent process? At least for you and the gear you used for playback.

 

[RayDunzl]

So if it is this hard 8 generations down the line would it be safe to conclude simply playing back some music is a transparent process?

Your demonstration is compelling up to the point it hits the speakers, then it's anybody's guess as to transparency.

Even so, observation of the waveform shows differences, so, no, it is not safe to conclude the process is transparent (as in an exact copy).

No argument, really, just stretching the point a bit.

 

[amirm]

So if it is this hard 8 generations down the line would it be safe to conclude simply playing back some music is a transparent process? At least for you and the gear you used for playback.

Yes. I am confident almost no one can pass this test without detecting the tells. It is good reference to have here next time someone claims to have better hearing than others .

I had passed Ethan's generational tests with much more ease than this one.

 

[Blumlein 88]

I wonder how much of the degradation (what little there is) is from the cable.

I don't think much. When I have done null tests the differences between cables is usually right near the noise floor. So I suspect it is really the noise being the main difference. The cables I used were 12 feet long so sort of like running thru 96 feet of cable.

I did something of a software simulation of my 8th gen test. I generated distortion with software like what my gear has when I measured it. I added pink noise at the level I measured with the gear. I repeated 8 times. The resulting noise level was slightly less than in the test. The distortion was almost spot on in terms of THD. It was pretty similar to the real test. Maybe the cables let a touch extra noise get picked up.

 

[fas42]

Was this before or after you looked at the answer?

After. I wasn't interested in actually doing this run, but was curious why people found it relatively difficult - the piano piece was a type which made evaluation more difficult, the music was too "simple" and gave less clues than more complex arrangements.

The quality of the transient attack of a strongly struck or blown note is a very good indicator of generational losses, I've used this signature when listening to all the examples put up in this forum - subjectively, the sense is that a multi copied version has a dullness, lack of sparkle, or mojo about it ... an analogy might be an extremely sharp knife, use it a number of times - yes, still cuts, does everything it should but that last "purity" of the sharpness is now lost - which would be invisible to a user, unless he had an unused knife to directly compare it with ...

 

[Blumlein 88]

After you knew the identities it's obvious. The lost edge. While on another forum it was described as obvious because of the sharp edges that builds up from distortion due to repeated playing with recording. The easy way to find the original was to listen for the analog naturalness to the sound or so I was told.

So which makes the most difference, the picture one has in their head with known identities or the sound of the file?

 

[fas42]

After you knew the identities it's obvious. The lost edge. While on another forum it was described as obvious because of the sharp edges that builds up from distortion due to repeated playing with recording. The easy way to find the original was to listen for the analog naturalness to the sound or so I was told.

So which makes the most difference, the picture one has in their head with known identities or the sound of the file?

Why is it obvious? Depending upon the quality of the playback/record chain the signature of the distortion that builds up over multiple loops will vary: sometimes the sound will dull; other instances an unpleasant edge will increase; or perhaps a low level random noise will start to intrude more with each pass. Every situation will have its own characteristics, there is no set "answer".

What matters is whether one can hear a difference or not. If you can't, then everything's good; if you can, then you decide whether it's worth making a fuss about - record companies make a fuss about "going back to the master", as do the customers ... once audio passes through a bit of analogue circuitry there will always be losses, and some people will find it easier to pick this up.

 

[Blumlein 88]

Why is it obvious? Depending upon the quality of the playback/record chain the signature of the distortion that builds up over multiple loops will vary: sometimes the sound will dull; other instances an unpleasant edge will increase; or perhaps a low level random noise will start to intrude more with each pass. Every situation will have its own characteristics, there is no set "answer".

What matters is whether one can hear a difference or not. If you can't, then everything's good; if you can, then you decide whether it's worth making a fuss about - record companies make a fuss about "going back to the master", as do the customers ... once audio passes through a bit of analogue circuitry there will always be losses, and some people will find it easier to pick this up.

I would be more impressed if you had these opinions before knowing which was which and could point to the correct file.

The comments about the edge increasing btw were about the same files which you said basically the reverse about. Also after knowing the identity of them.

"What matters is whether one can hear a difference or not." About this we agree, in fact that was the point of the poll.

 

[fas42]

Interesting that some registered the copied version as having more edge - this may have something to do with the playback setup they were using ...

I was not interested in being right here, I was interested in whether I could discern the variation, since others had mentioned that it was difficult.

In the bigger picture, worrying about how many generations one is down is irrelevant - I have some "appalling" recordings here, equivalent to a couple of hundred generations loss - there is no way to get back to anything near what one can call a master. My energies are spent on getting the most value, listening satisfaction from these - not on grinding my teeth because I want to revert to an earlier, "truer" version ...

 

[TBone]

88, not certain how relevant this is considering your OP, however ...

~about a decade or so ago, (for some reason, which I now forget) I ripped a song (Pretenders, Private Life) from my cdp (Ikemi) analog outputs into the pre-amp tape-in/tape-out, then into the ADC inputs of a Sony rcd-w1 cd recorder (IIRC, Sony first commercial grade, cheapest CD recorder) burned into a CDR.

The following results are consistent at any freq.

20-300hz (green = CD, yellow = rip(CDR) ...

1,200-1,400hz ...

10,000-10,200hz ...

 

[Blumlein 88]

88, not certain how relevant this is considering your OP, however ...

~about a decade or so ago, (for some reason, which I now forget) I ripped a song (Pretenders, Private Life) from my cdp (Ikemi) analog outputs into the pre-amp tape-in/tape-out, then into the ADC inputs of a Sony rcd-w1 cd recorder (IIRC, Sony first commercial grade, cheapest CD recorder) burned into a CDR.

The following results are consistent at any freq.

20-300hz (green = CD, yellow = rip(CDR) ...
View attachment 1934

1,200-1,400hz ...
View attachment 1936

10,000-10,200hz ...
View attachment 1937

Certainly consistent with results I get. Looks as if the player and the ADC are flat enough in response not to alter it much.

Many pro recording people consider ADC's a solved problem. One they don't worry much about.

 

[Thomas savage]

In your honour I found a theme tune for this thread...

Thanks for another fantastic contribution that's helping define our forum in the 'right' way

 

[Blumlein 88]

Thomas, yes the perfect song for this topic. Good find.

 

[Blumlein 88]

I just made available again these files. Updated with no tell-tale in the Fiona Apple this time. Poll is closed, but you can listen and comment if you wish. Thought maybe newer members might have a little interest in listening.

This one was done with different gear than my most recent version of 8th generation copies. Another tidbit of info. The SINAD on the 8th gen of this is 87.5 db.