CD Sound Quality/Accuracy - A B X testing

[kfbkfb]

http://drewdaniels.com/audible.pdf^^^
...our tests failed to substantiate the claimed advantages of high-resolution encoding for two-channel audio...


(I've been looking for this test report)


Kirk Bayne

 

[Mr. Afternoon]

Yep, the difference is imperceptible. There was a study a while back that found ultrasonics do emit a subconscious response, but in actual listening, there's virtually no difference.
But let's be honest, we have hard drives to fill, and it's nice having something in 24-bit 96-kHz for the hell of it.

 

[Doug G.]

I have always thought all the high resolution stuff was BS. People wanted to hear it so they thought they did. They also couldn't stand that the unwashed masses could have great sound without spending thousands of dollars.

Doug

 

[kfbkfb]

Kinda blending 2 threads - my Atmos DSotM just arrived, just now playing it - the LPCM 2 ch is 192kHz sampling rate quantized with 24 bits (overkill), but if it doesn't cost a lot more to digitize content with these rates and bits, why not (will have to check and see what if anything my Sony player is doing resamplingwise)?


Kirk Bayne

 

[par4ken]

There are many reasons to extend frequency response well above the hearing range.

Back in the analogue days frequency response of 20Hz-20kHz was often quoted. That often meant 3dB down at those extremes. Connecting a number of components with similar specs together would result in much greater rolled off bass and treble. Only by improving those specs as much as possible would ensure that those roll-offs don't directly add. Phase changes at the frequency extremes is another factor that could affect sound quality.

Early CD players had a very artificial sounding high end. As DAC's improved and over sampling was introduced the sound of the players improved greatly.

When I do my vinyl rips at 192kHz the audio editing program has much more to work with when processing the clicks and pops. Processing in the ultrasonic range helps to ensure near perfect repairs are done, artifacts would be inaudible. I could save at a lower sample and bit rate but choose not to, disc space is cheap!

SACD's rendered in DSD rather than being converted to PCM sound much better via my Oppo. I get good results converting to PCM using Foobar with the "short filter" installed (details in another thread). Less ultrasonic filtering is applied that way, resulting in better sound.

Audiophiles always insist on specs that are much better than what can actually be heard. That is the only way to ensure perfection! It is a goal that should always be strived for rather than this particular spec is good enough!

 

[ar surround]

SACD's rendered in DSD rather than being converted to PCM sound much better via my Oppo.

"Much better," as opposed to "somewhat better?" Don't you find the click of the relay annoying when the Oppo switches in and out of DSD?

 

[par4ken]

"Much better," as opposed to "somewhat better?" Don't you find the click of the relay annoying when the Oppo switches in and out of DSD?

Much better, and I never noticed the relay clicking.

The difference is that the sound seems more detailed. At times perhaps a bit harsh. When converted (by the Oppo to PCM) the recording sounds more like a tape. Yes I know it is from a tape but what I originally thought to be tape anomalies limiting overall sound quality disappeared with DSD playback.

 

[ar surround]

Much better, and I never noticed the relay clicking.

Oppo seems to believe that straight DSD and DSD to PCM sound different because they repeatedly state in the manual to choose one or the other depending on what one likes better.

When my Oppo 205 is set for SACD playback via DSD, a relay clicks when switching from a PCM to a DSD recording...and then back again. I understand that the Oppo 105 does not have this relay.

 

[barfle]

Mark Waldrep of AIX Records actually did a reasonably scientific study (he has a PhD in the field of audio) that concluded that it’s rare that people can detect the difference between CD and higher bit rates.

Note that his label was adamant about 24/96 recordings, which he still feels is the best recording spec, even though, for most people, it’s not perceptible. And anything beyond that might be good for editing and production purposes, but certainly not for distribution.

 

[JediJoker]

Note that his label was adamant about 24/96 recordings, which he still feels is the best recording spec, even though, for most people, it’s not perceptible. And anything beyond that might be good for editing and production purposes, but certainly not for distribution.

24/96 is the highest PCM resolution I will buy on download. It's also my preferred resolution for recording, editing, and mixing.

 

[jimfisheye]

About 15 years ago and back the average AD and DA converters available had poor low pass filters. (Low pass filtering to avoid the sampling frequency bleeding into the audio band that was right up next to it.) That led to HD sample rates to avoid the issue altogether. ie. The audio band was now down the middle of the road with a very wide margin between it and the sampling frequency. HD sample rates were never intended to capture ultrasonics above the audio band. Yes, some sound design people started pitching ultrasonics they captured at HD sample rates down to create new sounds and that kind of thing. That was exploiting an artifact of the system that was never intended to be preserved. The bs about ultrasonics was truly telephone game. (Note that neither your amps or speakers reproduce those frequencies.)

Nowadays the filtering in SD running converters is better and it's mostly a moot point.

There was never any limitation with SD sample rate itself. And the workaround to upsample to HD was and still is a thing. (Most AV receivers still do this by default BTW.) This didn't come out of a vacuum and you could hear it. There was a brief period where the stock sample rate clocks in the converter units were not made to a tight enough spec and that led to external clocks for an upgrade. Fully a moot point nowadays and another thing of the past.

 

[JediJoker]

Note that neither your amps or speakers reproduce those frequencies.

There are systems that will reproduce ultrasonics, notably some powered studio monitor speakers from the likes of ADAM Audio et al.

 

[gene_stl]

Even if your amp and tweeter can reproduce ultrasonics the tweeter beams turn into laser beams. If you are not sitting in the beam you will hear nothing. I proved this to myself in experiments with a test oscillator. It was a large speaker (not mine) with single tweeters (it was a stereo pair)
Above about 13kHz you had to be right in the beam to hear anything. The response went away at about 18kHz. I don't know if that is where the tweeter , or my hearing quit. Probably a combination of both. (it was like 30 years ago)

In order for there to be any dispersion of a high frequency beam the radiating surface or horn structure needs to be smaller than the wavelength so as to radiate into "half space" This means the tweeter diaphragm needs to be very tiny or alternatively maybe a ring radiator. The great majority of spikkers aren't even close.
At 15 kHz the wavelength is about 0.9 inch or 23 mm . So for any dispersion to take place the tweeter diaphragm needs to be half of that or better less.

 

[JediJoker]

Even if your amp and tweeter can reproduce ultrasonics the tweeter beams turn into laser beams. If you are not sitting in the beam you will hear nothing.

Using nearfields in a studio, an engineer is likely to be in the beam. Does that mean ultrasonics are important? Only in that acoustic intermodulation is a thing. I was simply pointing out that there are systems capable of reproducing ultrasonics. It's even more common in headphones, in which case beaming is a non-issue, as the transducers are pointed directly at the listener's ears.

 

[Mr. Afternoon]

I personally don't believe ultrasonics make a difference, but there is a study that was was used as the basis for the Akira 2009 Hypersonic 5.1 Mix (24-bit 192-kHz) which found that ultrasonics are confirmed to emit a subconscious emotional response, albeit, we don't know how the body is perceiving ultrasonics (it's certainly not through the ear) and why it seems to be much more effective on loudspeakers.

I think for us plebians with plebian setups, ultrasonics don't matter as much, especially with modern DACs. I'm not gonna be measuring my brainwaves while listening to music.

 

[JediJoker]

I think for us plebians with plebian setups, ultrasonics don't matter as much, especially with modern DACs.

Especially if you're using any of the most common room correction DSPs, which all resample to 48kHz for processing. Bye bye, ultrasonics.

 

[gene_stl]

I used to do all kinds of experiments with my system , for my own amusement. One of them was turning off the 5khz Upwards channel. You definitely could hear the difference. But what was left did not sound bad at all. If you weren't listening critically you could have not noticed/missed it.