A great explanation @Stan_Halen. Thank you!!
1 Like
The ultra sonic and subsonic frequencies affect what we do hear. These harmonics are in the ranges that canāt be heard by the human ear but, they react in a way that gives an instrument it characteristic sound such as piano and stringed instruments in particular. When these instruments are recorded at low sample and bit rates these frequencies are lost and many instruments donāt sound right and are fatiguing to the listener just as mp3 are. This is because the brain is trying to piece together the bits and pieces that are lost in the conversion and or dithering. At higher sample and bit rates more of this information is recorded to the point where at 192k/32 bit and above, the recordings are approaching the capabilities of tape that capture most all frequencies and sound the most natural. This is also why there is a resurgence in vinyl and even cassette tape sales. CDāS could sound much better than they do. We have the capability to, why not?
2 Likes
That may or may not be true, but you arenāt reproducing those frequencies simply by increasing the sample rate. Microphones, amplifiers, filters and speakers are all designed and built to operate within the human range of hearing, i.e. 20Hz to 20kHz. Youād have to redesign and manufacture the entire chain to have any chance of hearing the effects.
Simply not true. It is a proven, accepted, scientific fact that a soundwave only needs to be sampled a maximum twice in order to reproduce it perfectly. It is an impossibility to record āmore of this informationā when you already have all of it. Itās like buying an apple, and then buying more apples so that your first apple somehow becomes more āappleyā. Itās totally flawed logic. Further, the ācapabilities of tapeā, even if you refer to the finest machinery ever made, never extended past a bit depth of 13.
Bit depth (I assume you mean bit depth and not bit rate) is a statement of headroom/noise floor capability. That is all. Even 16 bits gives a colossal headroom of 96db, only the most specialised of applications could ever justify needing more headroom than that.
Higher bit depths are useful inside a DAW because of the multiple tracks; the sum bus can get overloaded at 16 bits. However, this is nothng to do with āsounding betterā, it is simply easier to operate, with less chance of clipping the sum bus.
Further, at 192kHz, the sample rate is so fast that errors are inevitable. So not only is it a rate that is totally unnecessary - literally - but it degrades the sound quality due to errors of speed, thus completely defeating its own (alleged) purpose.
1 Like
How would you do that? ā¦Suppose someone wanted to.
1 Like
That is my point. You couldnāt, therefore itās pointless using 96kHz for the purpose of an extended high frequency range.
Even if you could somehow manage it, there is still the problem of the general listening publicās equipment.
1 Like
I misunderstood what you were saying. So that makes sense.
Some monitors have a shelving configuration up to 40k. Mine sound different in full range mode. I donāt care if most people canāt hear it, I find it easier to make mix decision with everything in full range, then reduce to standard mode, then reduce to 2 way, then reduce to mono, when checking for translation at the very end.
ā¦I wouldnāt claim this has anything to do with 96k though.
1 Like
It makes sense that anything with electronics or āmanufacturingā such as speakers has the limited range, but as DaveP specifically mentioned piano and strings as cases with extended harmonics isnāt it possible that some fine acoustic instruments like that could (theoretically) benefit from being recorded at higher sample rates? In other words, if they demonstrate harmonics above and beyond half the Nyquist (22, 050 Hz) then they might not be fully reproduced by a 44.1k sample rate?
So youāre also citing Nyquist here? As per above, I would ask that if the soundwave did happen to be beyond human hearing (or 22,050 Hz) with harmonics/overtones then itās possible that a higher sample rate could conceivably capture a tiny amount of subtle harmonics that would impact the more fundamental (lower) frequencies? If so, I think weāre looking at very subtle differences and nuances that might only be heard by the most discriminating audiophile. Iām just trying to be clear on the discussion.
As to tape capturing all frequencies and sounding more natural, I would tend to think that since tape is mechanical/electronic/manufactured that it also has limitations on capturing a wide range of sound. Tape technology was notorious for high-end roll-off; losing higher frequencies. And the electronics contained filters, whether intentionally or by the action of capacitors etc. Rather than saying it sounds more natural, I would say it sounds more āpleasingā to some people (due to the inherent recording limitations). It could be said that analog does sound more natural or pleasing because to some degree it emulates how our ear/brain hears things. A long standing complaint about digital (whether deserved or not) is that it sounds too āclinicalā and āsterileā. In other words, itās not ādirtied upā.
But do you still subscribe to the info you referenced in Post #34 that recording at 96k might have some practical benefits in terms of electronics (aliasing, filters) and processing?
I donāt have the faintest clue how this stuff actually works. My question would be the sameā¦if 23K audio can effect what we can hear based on the fact that energy waves from energy clearly interact with each other. That raise the question is we can hear the result of the interaction. I donāt knowā¦
I do buy into the theory that subtle matters. Dave Hill at CraneSong scientifically demonstrated that jitter differences between DA converters is audible, and that that jitter from an Antelope and Big Ben sound different than jitter from and Avid i/o.
Isnāt that supposed to be the reason izotope plugs upsample to the 384k?
1 Like
I think AJ mentioned earlier that some plugin processing can be more efficient at 96k (or whatever higher sample rate). Iām guessing that is a kind of āoversamplingā which might do something similar to the time-stretching effect you mentioned, where using a higher sample rate just minimizes distortions and errors in extreme circumstances.
Keep in mind that just because a manufacturer touts an impressive spec doesnāt necessarily mean that it improves anything or makes a difference. Such things can be used as marketing tools simply to get people to drop their jaws at NAMM or look better than the competition. Iām not saying it doesnāt do anything, just that Iām skeptical of marketing claims.
If I recall correctly, Boz said the code in that izotope stuff is real deep. Thereās a lot to be said for whoās using it as far as taking advantage of those cutting edge features. Iām not one to say if the up sampling algorithms are truly beneficial, because I donāt have anything to compare it with. Iāve only ever used Ozone for makeshift mastering stuff (that we all know is not really mastering). lol.
As much as it gets me by for what I need it to do, its not a plugin that Iād swear by, because I donāt anywhere close to enough experience in mastering to know.
1 Like
Yes, of course, I didnāt deny that.
I said that itās flawed logic to assert that simply increasing the sample rate to 96kHz is going to produce the desired effect. Even assuming that reproduction of the higher frequencies somehow became possible, you would then have to deal with assertion that it actually makes any kind of discernible difference.
Certainly we are talking about subtle differences. Whether anyone at all can tell the difference is debateable. As far as I am aware all experimantation thus far has shown that nobody can tell any difference at bit depths higher than 16, and sample rates higher than 44.1kHz. I mean, Iām not close-minded about this but it doesnāt look good for the high sample rate camp currently. However, that state of affairs does not seem to deter manufacturers and people who have bought their products from claming otherwise.
Yes of course, I did my first 96kHz recording a couple of days ago. I would have reported back to this thread but the only thing I have to report is that I havenāt anything to report. I did notice the computer slowing down a little bit but I couldnāt say for sure that it was due to the 96kHz rate.
2 Likes
Ha, well I guess thatās not unexpected. Kudos for trying though.
I assume double the samples means more work for the computer. I seem to remember that (theoretically) there is a heavier load for a higher sample rate than there is for a lower sample rate, given the same latency and other conditions. It would seem to make sense. Thatās what I learned IIRC, but havenāt ever tested that out.
Now you guys have me curious. I think Iāll sit down next week and see if I can do a quick acoustic thing (maybe voice and guitar) at 16/44.1, 24/48 and then again at 24 or 32/96.
Run it through a couple different spectrum analyzers, warp the crap out of it. Just generally beat it about the head and shoulders and see how each one holds up.
Super scientific testing methods.
2 Likes
Here is an interesting article in SOS that discusses the pros and cons:
https://www.soundonsound.com/sound-advice/q-should-i-use-high-sample-rates
1 Like
Good article! I didnāt even think about jitter being introduced as you increase the rate. Good point!!
1 Like
I thought about doing this as a test too, may still do it. My idea was to use 2 of the same mic (I think mine are even a āmatched pairā), use one mic going to one computer recording at 44.1k, and the other mic right beside it going to another computer recording at 96k. That might try to get things as similar as possible (same performance) for comparison results. While some may not have that kind of setup, I happen to use two different computers (Windows and Mac) for recording, depending on what and where Iām recording it, and have an interface for each.
1 Like
Iād love to hear the results! Although, itās not quite as identical as youād probably prefer given the PC and Mac involved, I think itāll still work for this purpose!
1 Like
Yes, having exact same computers and interfaces - everything the same - would probably be more objective. I donāt know that PC and Mac would record the sound differently, but they might. Mac is usually recommended for music recording (in some circles) because of ease of ease of audio configuration and workflow, and perhaps other criteria. It doesnāt record sound better than Windows is my understanding.
Identical interfaces might be the more crucial issue, preamps and converters, same settings, etc.
1 Like