[gregorio]

Does over-sampling reduce perception of resolution? Does this have to do with interpolation?

No and no. Changing bit depth from a lower bit depth to a higher bit depth should result in absolutely identical data/sound and over-sampling should be completely audibly transparent. I'm with bigshot on this one, something else must be going on when you choose that mode.

G

 

[1800yolk]

Noise!
I'm not sure I understand, why would the amp not be doing that quite as much with 16/44 stuff? 0dBFS is the loudest a digital file can be and 0dBFS is exactly the same with 16bit as it is with 24 or any other bit depth. What changes with increased bit depth is the quietest signals we can theoretically record.
G

Thank you for your reply! The 16/44 vs 24/44 would make no difference with the amp in regards to generating any harmonics that would influence the sound. I think there is a difference in 16/44 vs 24/88 though.

 

[bigshot]

If you're taking 16/44.1 and bouncing it up to 24/96 it shouldn't change the sound at all. Everything is identical within the range covered by 16/44.1. The rest is all zeros so to speak.

 

[1800yolk]

If you're taking 16/44.1 and bouncing it up to 24/96 it shouldn't change the sound at all. Everything is identical within the range covered by 16/44.1. The rest is all zeros so to speak.

I'm discussing digital releases that are 24/88 to begin with, obviously putting 16/44 in a 24/88 container would do nothing

 

[bigshot]

Well then it's recorded sound that you can't hear because it's frequencies outside the range of human hearing, or it's sound so quiet, you would have to turn the loud stuff up to deafening levels to hear it. Check out the link "CD Sound Is All You Need" in my sig.

 

[RRod]

Thank you for your reply! The 16/44 vs 24/44 would make no difference with the amp in regards to generating any harmonics that would influence the sound. I think there is a difference in 16/44 vs 24/88 though.

The harmonic distortion for any content > 11kHz will be inaudible, so you're asking in essence about audibility of IMD. This would require that the high-res frequencies contributing to the IMD be loud enough so that the IMD product is audible *over the other content potentially masking it in the audible range*, unless you have some kind of unbounded distortion occurring. If you look up the people trying their damndest to prove audibility of hi-res, you'll find examples such as playing back recordings of gamelan music really loudly, because only certain content has a snowball's chance in hell of any hi-res induced IMD being audible, and that's assuming the distortion heard is due only to the non-linearities of the ear and not the speakers.

 

[vatch]

This thread is quite old and any and all questions have been answered suitably. Before asking questions I suggest reading it through. There's really no purpose for all this extra data; it simply is very high frequency noise, something we almost always try to reduce in audio.

I highly doubt a tube amp capable of producing much greater than 20hz-20khz frequency range anyway which is the same for most speakers, ss amps, etc. There is EMI and RFI noise on the power lines we always try to reduce. Why add to the noise?

It's not nice to blow your dog/cat/guinea pig/parrot's ears out!

 

[gregorio]

The 16/44 vs 24/44 would make no difference with the amp in regards to generating any harmonics that would influence the sound. I think there is a difference in 16/44 vs 24/88 though.

That is possible, in terms of the sample rate rather than the bit depth though. Expanding on what @RRod stated, we'd effectively be talking about IMD. If we feed a signal to an amp beyond what the amp is designed to accept, say a 40kHz signal input into an amp designed for signals in the range of 20Hz to 20kHz, an amp can respond with distortion in the audible range. I read an article or paper from a reliable source some years ago, which tested a number of speakers and amps and concluded that this type of audible distortion occurs much more commonly than is generally assumed, even with high sample rate material which only contains relatively small amounts of ultrasonic material. Unfortunately, I can't remember the title or enough other details of the article to locate it with a cursory search. Does anyone out there know the paper/article in question?

G

 

[1800yolk]

That is possible, in terms of the sample rate rather than the bit depth though. Expanding on what @RRod stated, we'd effectively be talking about IMD. If we feed a signal to an amp beyond what the amp is designed to accept, say a 40kHz signal input into an amp designed for signals in the range of 20Hz to 20kHz, an amp can respond with distortion in the audible range. I read an article or paper from a reliable source some years ago, which tested a number of speakers and amps and concluded that this type of audible distortion occurs much more commonly than is generally assumed, even with high sample rate material which only contains relatively small amounts of ultrasonic material. Unfortunately, I can't remember the title or enough other details of the article to locate it with a cursory search. Does anyone out there know the paper/article in question?

G

This is great! I appreciate everyone's answers, b/c there are lots of points to consider. I found an interesting article about IMD that explains why the even order harmonics I was hoping to achieve aren't as natural and desirable as they're made out to be:
http://sound.whsites.net/valves/thd-imd.html
So in the end, pumping 24/88 music through a tube amp, in addition to other sources of distortion mentioned above, will create distortion that may or may not be better than settling for clean, predictable 16/44 music. It's hard to say what music benefits from this in any capacity, but my list would include albums that were poorly mastered. It could also make them worse! It's an interesting gamble.

 

[gregorio]

So in the end, pumping 24/88 music through a tube amp, in addition to other sources of distortion mentioned above, will create distortion that may or may not be better than settling for clean, predictable 16/44 music.

Pumping high frequency material through an amp MAY (not necessarily "will") create IMD. If our goal is high-fidelity reproduction, then in no sense could the introduction of IMD during reproduction be considered "better". Furthermore, as introduced IMD is typically unrelated to the content, there would be few occasions where it would be "better" even purely in terms of personal subjective preference. Often with tubes, the goal is the addition of even harmonics, which although by definition is deliberately "lower fidelity", can nevertheless be euphonic (pleasing). Even the addition of odd harmonics can be desirable under various conditions but tones which are not harmonics, not related to the fundamental frequencies at all, are much more rarely desirable. IMO, the use of IMD should be restricted to specific places and specific amounts, and that means restricting it's application to the mix/master itself and not randomly applying random amounts during playback!

G

 

[71 dB]

...and over-sampling should be completely audibly transparent.

G

Audibly transparent perhaps, but:

Oversampling 44100 Hz to 88200 Hz adding zeros between every sample point and low-pass filtering the result is very transparent.
Oversampling 44100 Hz to 96000 Hz with linear interpolation is theoretically not very transparent (interpolation distortion).
Oversampling 44100 Hz to 96000 Hz with zinc-interpolation is very transparent.

If we feed a signal to an amp beyond what the amp is designed to accept, say a 40kHz signal input into an amp designed for signals in the range of 20Hz to 20kHz, an amp can respond with distortion in the audible range. I read an article or paper from a reliable source some years ago, which tested a number of speakers and amps and concluded that this type of audible distortion occurs much more commonly than is generally assumed, even with high sample rate material which only contains relatively small amounts of ultrasonic material.

G

Was it a paper by Lavry?

The benefit of "low" sample rates 44.1 kHz and 48 kHz is that after proper reconstruction filtering hardly anything above 20 kHz is fed to amps.

 

[gregorio]

[1] Oversampling 44100 Hz to 96000 Hz with linear interpolation is theoretically not very transparent (interpolation distortion).
[2] Was it a paper by Lavry?

1. True but I don't believe that still occurs much (or at all) these days does it? The ADCs/DACs I was using more than 25 years ago were oversampling 128 times.

2. I don't think so. While Lavry has certainly mentioned IMD and IMD avoidance with the standard sample rates (44 and 48) in at least one of his papers, the paper/article I'm referring to actually tested some amps and speakers for the occurrence of IMD and concluded it's occurrence was far more prevalent than generally assumed. IE. It's generally assumed to be fairly rare, except when using signals designed to elicit IMD but it may in fact be quite/very common, even under normal usage consumer usage conditions, when reproducing typical commercial music recordings.

G

 

[skwoodwiva]

While not science at all!
I believe this fundamental is why fone, the Italian, label is the finest producer of CDs Then they skipped so called better pcm formats and went exclusively DSD.

 

[RRod]

While not science at all!
I believe this fundamental is why fone, the Italian, label is the finest producer of CDs Then they skipped so called better pcm formats and went exclusively DSD.

And their microphones are from 1947!

 

[skwoodwiva]

And their microphones are from 1947!

You know this to be true
Or joking I cannot tell.