DP 44056 sample rate error

[MIDI Life Crisis]

Don't bogart that Koolaid, my friend. Pass it over to me-e-e!

[EMRR]

The largest perceptible differences, IMO, are between 44.1K and 48K, and between either of those and 88.2K. I still don't know why it makes so much difference at those low levels, with such a small change. 88K I can understand.

Phase distortion differences caused by the aliasing filters are easily the perceptible portion in the audible range. It isn't frequency response.

Phase comparisons of MOTU 16A and MOTU 2408mkIII below. While 44.1 to 48 may look small, I've found it easy to perceive measured phase changes smaller than this in A/B tests. This is one of those things that conventional wisdom says is imperceivable on average, but can be detected under scrutiny.

128K could make for a smoother filter, and anything within reason that saves data headroom is a good thing. I live at 88.2 for headroom reasons, perceiving no functional difference there versus 96....with my converters...caveat, caveat, caveat.....

Full resolution here

[mikehalloran]

I thought about bringing up the anti-aliasing filters but when what's his name stormed out like a three year old, I figured this thread was dead.

I was a beta tester for early CD players around 1981 and hated the sound, especially on orchestra recordings -- a Chabrier recording was especially painful. I didn't know why till later when articles came out about the phase distortion and it all made sense. 48k pushes the affected frequencies higher where fewer can hear them.

The filters are much better than they were 30 years ago and my hearing is worse so I no longer hear the difference even if I can see it on the graph

As for his rant, perhaps 128k might have made sense when storage was more expensive. Now it's cheap.

One thing I like about 176.4/192k is that hi-hats and cymbals no longer sound all the same, a quality found on analog recordings. Weird that I can hear the difference. Going strictly by the numbers, I shouldn't... except that the numbers don't say everything. It is known that brain uses frequencies in the 25kHz range to locate objects in the stereo field even if our ears can't perceive them.

[Shooshie]

The largest perceptible differences, IMO, are between 44.1K and 48K, and between either of those and 88.2K. I still don't know why it makes so much difference at those low levels, with such a small change. 88K I can understand.

Phase distortion differences caused by the aliasing filters are easily the perceptible portion in the audible range. It isn't frequency response.

Well, duh! I guess the operative words there are "audible range." Makes sense when you put it that way, though I knew it wasn't really a frequency thing, since all of the differences are happening way above my hearing level. But phase between those aliasing frequencies would be something that, in theory, could be perceived in the range.

Phase comparisons of MOTU 16A and MOTU 2408mkIII below. While 44.1 to 48 may look small, I've found it easy to perceive measured phase changes smaller than this in A/B tests. This is one of those things that conventional wisdom says is imperceivable on average, but can be detected under scrutiny.

128K could make for a smoother filter, and anything within reason that saves data headroom is a good thing. I live at 88.2 for headroom reasons, perceiving no functional difference there versus 96....with my converters...caveat, caveat, caveat.....

Very interesting. Thanks for the chart. I think I'll drag that to a folder for future evidence somewhere down the line. I also like 88.2K for the same reasons. If 192K were not such a drag on the resources of both CPU and MOTU box, I might use it all the time, but I honestly can't tell the difference between that and 88.2K. At least, I haven't learned where to listen for it. Would you be more likely to hear such differences around the edges of cymbals? Spatial effects of a hall reverb? Instrument placement? The differences I hear between 44.1 and 48K sound... (I hate to use words like this in this context) luxurious. Connected. Smooth. Totally meaningless words, used subjectively to describe something whose exact context I can't place, empirically.

88.2K is where I like to be, though if I'm recording gigabytes of audio, I'll do it in 48K to save space. The gain between 48K and 88.2 is perceptible, but not enough to worry about if space is an issue. Of course, these days with 3TB drives... Oh well, old habits are hard to break.

Shoosh

[EMRR]

I think of my average LP project done at 88.2 as a 50GB bucket. Divide that out among even a 1TB drive....I don't want that many project archives reliant on a single drive (ignoring the obvious backup drives). If the CPU will do it, burn the space. They're only making more.

A bunch of old timers poo-pooed my phase chart, stating that all anti-aliasing converters should be phase linear by now. Test gear reports otherwise.

As to what you can hear, I was looking at the difference in a freq/phase flattening filter after noting the sonic difference, and found only very small movement in either graph above 20K, no change below that. The effect of the filter on program material was that it would make the cymbals sound more forward, or more set back in the mix, to such a degree that one was the clear winner for the mix at hand.

[Tritonemusic]

I feel compelled to mention that I love reading EMRR's posts. Always bringing up interesting facts and observations. Thank you, Doug, for your contribution to Motunation.

[Shooshie]

I feel compelled to mention that I love reading EMRR's posts. Always bringing up interesting facts and observations. Thank you, Doug, for your contribution to Motunation.

Same here! Always a pleasure reading your posts, Doug!

Shoosh

[mikehalloran]

A bunch of old timers poo-pooed my phase chart, stating that all anti-aliasing converters should be phase linear by now. Test gear reports otherwise.

Ahhh, I love it when people claim that the laws of physics don't apply.

True, they have improved over the analog anti-aliasing filters of 30+ years ago but phase issues are still built into the process. It can't be helped.

CD players went on the market around 1983 as I recall. I refused to buy one till the mid 1990s when they got good enough so my ears no longer hurt when I listened to them.