I can't agree or disagree with what you say, I don't know enough about the ins and outs of the subject to do that. All I can detail is how the files were arrived at (which were done in good faith I might add, no attempt to do anything misleading with them).
Would anything you see be related to Audacity ? and I'm assuming it wouldn't be PC specific, meaning the machine used to run Audacity to create the files.
Would anything you see be related to Audacity ? and I'm assuming it wouldn't be PC specific, meaning the machine used to run Audacity to create the files.
Can't tell since I don't know how the original files looked like. Could you tell us the source of the files?
These tests may be rather confusing than revealing, if an exact scientific method is not used, and if the testing hardware is not perfect. What kind of dither has been applied in Audacity?
I don't know how to confirm that. So these particular test tracks are not all they perhaps seem quality wise ?
Very interesting... I think something must have gone wrong somewhere in the processing, I guessed right at how you made the files comparable sizes, but really thought that X was the 'upcycled' mp3, and Z was just too much information for the processor and hence the 'interference'. I have downloaded a fair number of 24bit files but not found anything I know or already have on CD to really be able to compare. The 24bit music I have heard though is all obviously superb recording and production, we need to be able to listen to exactly the same recording and production in all the different formats to be able to judge, preferably with something familiar.
I don`t work with this software but I guess 44100 on the right is original file sample rate? Where did you found the original hi-res file? The master file should be recorded in hi-res in the studio, when the song is recorded, and you shoul know it for shure otherwise this kind of test doesn`t make sense .
I listened to these on my Mac with MusicFans, through the audio output jack into Cambridge Audio powered PC speakers.
It was enough of a difference to be clearly audible, even on this lousy setup.
The problem that always crops up in tests like this is getting files to look the same and be in the same format. If that is not done then everyone picks the 'correct' file without even listening. So how to get the different formats to look the same yet preserve the quality of the originals is something to be looked at.
The files were from the link I posted earlier with the one labelled 24bit/88200 taken at face value that that is what it is. Do you mean you think the originals were really only 44.1 and upsampled to provide the so called 88.2 master ?
Yes, and the results show a definite trend to be able at least to differentiate the lowest quality MP3 from the two masters. So I think there has been some validity in this even if its not all quite worked out as I thought it might.
There are a lot of complaints on the internet that many so called high resolution downloads are actually upsampled from the original CD master. No wonder so many people claim not to hear any worthwhile improvement from many of them, me included.
I did think X&Y were VERY close but i could hear the degradation on Z.
As you point out, the test does at least show that most of us can tell MP3 from FLAC at commonly used rates.
I did think X&Y were VERY close but i could hear the degradation on Z.
As you point out, the test does at least show that most of us can tell MP3 from FLAC at commonly used rates.
Last edited:
That's interesting to know. Its all a new subject for me, something I've not done much (well any) real research on.
I've been using computers (among other things) as a source for about 8 years now but i've got loads still to learn about this digital stuff 
Lol, I know what you mean. Pavel (PMA) has set me right on a few things before now but there is still loads I don't know. But its how we learn 🙂
The somewhat bizarre thing is that the spectrum of Y, the CD quality one, actually extends slightly higher than the supposedly HiRes version, at the 20kHz mark, by about 1kHz's worth ...
There is so much nonsense being talked about the capturing the "information" above 20kHz being so important - well, so far every time I've looked I've found this to be crap. Valid above 20kHz captures only show any data in moments of extreme percussive sound, crescendos, and comparing the energy levels in the region above 20kHz, in the very short periods when it occurs, with the intensity of signal below 20kHz shows that the ultrasonic component will be completely buried, irrelevant. Sorry, it's a no-go for me ...
As people have hinted many times, high resolution recordings sound good because so much care has gone into recording them - that's 99.9% of the answers, IMO ...
There is so much nonsense being talked about the capturing the "information" above 20kHz being so important - well, so far every time I've looked I've found this to be crap. Valid above 20kHz captures only show any data in moments of extreme percussive sound, crescendos, and comparing the energy levels in the region above 20kHz, in the very short periods when it occurs, with the intensity of signal below 20kHz shows that the ultrasonic component will be completely buried, irrelevant. Sorry, it's a no-go for me ...
As people have hinted many times, high resolution recordings sound good because so much care has gone into recording them - that's 99.9% of the answers, IMO ...
How do you know that the 1kHz above the 20 kHz on the spectrum analysis is really part of the music? It could be some kind of noise or distortion artifact.
When I listen to Y as compared to X, there is slightly more fuzzy stuff to the sound that should not be there. I hear this, even though my hearing drops off at 11.3 kHz.
Even though my older ears are not as good as they used to be, musical instruments in a live concert still sound very much the same to me as they used to when I could hear up to 18 kHz.
When I listen to Y as compared to X, there is slightly more fuzzy stuff to the sound that should not be there. I hear this, even though my hearing drops off at 11.3 kHz.
Even though my older ears are not as good as they used to be, musical instruments in a live concert still sound very much the same to me as they used to when I could hear up to 18 kHz.
I don't, 🙂 ... as you say, most likely an artifact of the original creation of the files, done elsewhere - not as a result of Mooly's setting up of these samples.
That "more fuzzy" quality should be a sign of some significant variation between X and Y - what I'll do now is a careful subtraction of those two from each other, and see if I can isolate anything that may be significant ...
Wish me luck ... 😀
That "more fuzzy" quality should be a sign of some significant variation between X and Y - what I'll do now is a careful subtraction of those two from each other, and see if I can isolate anything that may be significant ...
Wish me luck ... 😀
Frank - i actually agree with most of what you said in your previous post. I take it you don't believe in the notion that frequency's outside of our hearing range can impact on or effect in some way those which are audible?
Cornelis Spronk - I didn't notice any increase in noise when going from X to Y, just a very vague impression of more clarity in certain areas. If noise existed on the CD quality file i don't see how that noise would be removed by the upsampling process unless it was somehow raised in frequency. Would the upsampling process also act as a filter in some way? I don't know that much about the processes going one, it involves enough maths for my concentration to start wandering. I had that problem way back in school too 😉
Cornelis Spronk - I didn't notice any increase in noise when going from X to Y, just a very vague impression of more clarity in certain areas. If noise existed on the CD quality file i don't see how that noise would be removed by the upsampling process unless it was somehow raised in frequency. Would the upsampling process also act as a filter in some way? I don't know that much about the processes going one, it involves enough maths for my concentration to start wandering. I had that problem way back in school too 😉
Frequencies outside our range can have an audible impact, but not for the reasons typically thought of - what they highly likely will do, is subtly alter the behaviour of the DAC, or other circuitry, so that distortion "folds down" into the audible area ... or, they might actually improve the distortion characteristics, for some reason! - "there is a difference, therefore we must be hearing the higher frequencies!", 🙄.
Well, there are differences between X and Y, significant ones, to a sufficient level to be audible - and there shouldn't be!! - somewhere, something has been done, accidently, or intentionally, to make those hires and redbook files not match! I won't proceed further in examining them - a better test setup might be needed, perhaps ... 🙁.
Well, there are differences between X and Y, significant ones, to a sufficient level to be audible - and there shouldn't be!! - somewhere, something has been done, accidently, or intentionally, to make those hires and redbook files not match! I won't proceed further in examining them - a better test setup might be needed, perhaps ... 🙁.
- Status
- Not open for further replies.