gpapag - "The live sound of harpsichord is heavenly. It is miles away from the recorded"
Yeah, I went to see my kid's high school band play "Sleigh Ride" in the gym, noted for its basketball game acoustics. Whatever instrument the one student was using to make the "whip" sound - a nice acoustic transient BTW - I was thinking "c'mon - no stereo is going to come close to how that simple little part sounds live". It was definitely the loudest instrument in the orchestra.
Yeah, I went to see my kid's high school band play "Sleigh Ride" in the gym, noted for its basketball game acoustics. Whatever instrument the one student was using to make the "whip" sound - a nice acoustic transient BTW - I was thinking "c'mon - no stereo is going to come close to how that simple little part sounds live". It was definitely the loudest instrument in the orchestra.
Did you read the rest?
Its saying Hi res audio sounds better because they are masterd differently. If you convert those hi res recordings to Cd theres no audible difference.
Just to satisfy my own curiosity, I made a symmetrical signal composed of a short 1msec exponential ramp-up, followed by a 0.5msec stable level, and at last a 1 msec exponential ramp-down.
Modulated frequency is 18Khz and Brick wall filter is at 20Khz.
Used time window was 10msec, magnified in the image below to 4.5msec.
Note that the difference between original and Brick wall filtered version has been amplified by 20dB to get a more accurate view on what was altered.
Again no signs of strange effects, everything looks well after filtering IMO.
Hans
Modulated frequency is 18Khz and Brick wall filter is at 20Khz.
Used time window was 10msec, magnified in the image below to 4.5msec.
Note that the difference between original and Brick wall filtered version has been amplified by 20dB to get a more accurate view on what was altered.
Again no signs of strange effects, everything looks well after filtering IMO.
Hans
Digital recording and reproduction strategies especially WRT media, dither algorithm and filter type was changing every month it seemed well into the 1990s. It was during these times we experimented with quite a few early AD and DAs, Apogee filters, etc...anyone remember the Pygmy?
The initial budgetary requirement I gave management was for a 4-replication line CD plant w/clean room for glass mastering, electroplating, QC, screen printing, Sonic Solutions DAW pre-mastering and all other needed services (2MW of new 480V power busses, 20 tons of chiller, 20MΩ H2O plant, 100HP of compressed air, waste treatment, exhaust...) was north of $20M in 1991 USD and commandeered 15k ft^2 of factory floor space from the previous cassette plant. AFAIK it was repaid in 24 months, by which time we were already into our second expansion.
Then there was the learning curve and training ops and techs, bugs in new technology to be worked on with mfgrs...and don't get me started on how difficult it was to train management and sales!
Busy times as the sole engineer...I slept in my office more than once and likely neglected my family more than I should have which I now regret.
All which is far more than you asked, thanks a lot for triggering the memories!
It is nice to be more part-time employed these days...
Cheers!
Howie
I had that many times when friends demonstrated me how much better their SACD sounded compared to the CD version on the other side of the disc.
However when playing this SACD and a down sampled 44.1/16 version on Jriver, all differences between the two were gone.
This must be one of the reasons how people were made to belief that Hi res sounds so much better.
In fact, quite some CD's are of mediocre quality, having nothing to do with the format.
Hans
The authors found the sound quality of the modern "hi res" records superior in comparison to "most CDs" .
It is a bit strange IMO to offer such an judgement - based apparently just on sighted listening - within a study that is primarly done because "sighted listening" comes with a higher risk of impact on listeners judgement.
The main problem - beside the statistical problems - was that nobody took measurements of the equipment before the experiment started and obviously , although they were running a long time experiment that lasts ~a year, nobody checked the equipment regularly.
After a couple of weeks they noticed that one of the mainly used units had a problem with lowlevel linearity but nobody had tracked how many trials were done up to this time.
Further the authors did not check if the records used actually contained any ultra highfrequency content and it was not tracked which records were used during the trials.
The authors offered additional information on a website (space restriction in the JAES) but the fundamental problems couldn't be resolved. Further some discussions on other forums brought some anecdotal additional informations about the records presumably mainly used.
The majority of trials was done at the main system, but some additional trials were done at other locations on apparently good quality systems, but I'm always surprised if an experimenter asked the subject in a sighted listening test if the added ABX comparator box introduced an audible effect and after the listeners denied such an effect, concludes that usage of the box is justified.
After that he does with the same listener "blinded" controlled listenening test because one can't trust "sighted" listening results.... 😉
So, as stated before, it might be that the hypothesis of the authors (the better sound quality of the so-called "hi res" records is not based on the "hi-res" property but on the better mastering/production) is correct, but one simply can't base such a conclusions on the data of the experiment due to the methodological problems.
Last edited:
Perfect simulation, thank you.
Your verbage of course, is beneath you.
Examine your in and out with a careful eye.
The frequency, as expected is absolutely positively rock solid, as expected.
The amplitude however, is dependent on the slope of the envelope. At the lower slope regions, the amplitude is the same. At the higher slopes, just before and just after the flattop, the filtered output has a lower amplitude than the original.
In addition, the difference, again exactly as expected, tells us exactly where in the waveform the upper sideband extends past the Fs, in this case...20 Khz.
So, your simulation is an absolute confirmation that:
1. The exponential modulation does indeed force a sideband
2. That sideband extension is dependent on the instantaneous modulation slope.
3. The sideband is seen by the filter and removed
4. The gibbs shows us exactly where the sideband exceeds Fs.
So, you have absolutely and positively confirmed that the choice of Fs too close to human hearing can indeed modulate the envelope of the content.
As an independent note...at no time does this result prove in any way that the CD brickwall choice is audible. It does however, prove that modulation of the audible signal can indeed cause a sideband above Fs that can alter the signal.
See, that wasn't so difficult now, was it?
jn
<Cough!> One may recall Mr. Wurcer has pointed out on multiple occasions that AES will publish anything.
Unfortunately the sword cuts both ways on that Mark. I have read through a number of high res vs CD papers, many of which claim a positive difference. In every case I could see (ab)use of statistics in ways that are considered "p-hacking" or the like, namely squeezing the data to tease out a predetermined result. A definite no-no.
But, yeah, methodologically the Moran paper leaves a LOT to be desired.
But, yeah, methodologically the Moran paper leaves a LOT to be desired.
I know..😉
It's the language he understands. From the jump, he has constantly told me I do not know what I am talking about, it's apparently his go-to reflex. That and casual dismissal.
And not appreciated.
jn
Well John, I'm not sure where all the confusion started and don't really care anymore, but it has been ubiquitous and I don't think Hans is anymore guilty than anyone else.
We are doing Satie for our grand daughter. We're on the third copy of wheels on the bus (book) she still likes to shred pop-up books.