You dont have to be gone. I am not offended by any means. Quite the contrary I feel that this is quite an informative topic. The way I see it, this is always going to be a gray area. You do what works best for you.
I had an opportunity get some samples of my favorite orchestral instruments. Wish I had recorded the event, it would have made a great youtube video…but whatever. I had asked a friend who was a professional audio tech at an orchestra to record my flute and some other instruments like bass drum, bassoon and some brass instruments. He recorded them simultaneously in 16 bit and 24 bit. Then he asked which I prefer so we did some blind tests about a 100 times and I was preferring the 24 bit version 90+ times as I could hear the richer tones of my gold flute better in 24 bit version. The timbre of the instrument, the early reflections and the ambiance of the hall acoustics were more prominent in the 24 bit version. As if it had an envelope on the sound. Same with the other instruments. 16 bit recordings of these high end instruments felt quite a bit “mundane”. I had a feeling that the head was too close to the roof of the car lol. He ended up deleting the 16 bit recordings and gave me the 24 bit versions. I wish I saved the 16 bit versions.
@FluteCafe we all hear differently I guess and my hearing is not good in general but is this difference something that some or all of us here on this forum could hear, occasionally at least? Would it require that you can hear above 10kHz for instance?
the tone and timbre differences are subtle and have audible overtones (harmonics) in the 10k-18k range depending on the note played. In an orchestral setting, our ears are trained to hear the overtones of the A (8th octave) note when looking for timbre nuances. These frequencies modulate audibly between 8k and 20k so any kind of hearing loss near the 10k will hamper the ability to hear this.
the spacial nuances can be heard by anyone though if you know what you are listening for. The first reflections of the sound in a hall, or the binaural resonances that provide the acoustic envelope are generally low frequencies, so if you know what you are listening for, you can pick up on these. A good mic will pick these up regardless of bit depth but the deeper the bit depth, the bigger the headroom and more ambiance information is recorded clearly to file. That is what is perceivable by some people, though a small number of the population.
Personally, I have not been able to tell the differences between different sample rates in a blind test. Only the bit depths. I will be listening for transients in different sample rates next time I have an opportunity to do this.
Hey, your contributions here are meaningful, and I know you are very good at what you do.
Whenever discussions like this occur, the subjective side of things is going to be involved, as well as differences in hearing ability, playback equipment, and in some cases the state of mind of the listener.
I doubt I could conclusively hear higher bit depth in a blind test, but I don’t doubt I could be convinced that I could be persuaded with some tips about what to listen for.
The point to me is always going to come back to what the listener is using to hear what we slave over to record. 99.9% of the time the playback system is severely lacking in a normal situation, so higher bit depth will always be a very low rung on the ladder for popular music.
That’s not to say we shouldn’t continue to strive to raise the bar, it only means the audience is not likely to get larger.
Thank you to Michelle and Bob for your encouragement. I have returned to reply to points made.
In other words, two separate signal paths, two separate sets of hardware and processing.
I don’t doubt it, but it was not because the recording was 24 bit. I surmise that the 24-bit signal path utilised higher quality hardware. Two entirely different recordings (albeit simultaneous) of the same instrument is no test of anything - unless you are specifically comparing the hardware involved.
But you have the 24-bit versions; you don’t need the 16-bit versions. Just convert the 24-bit to 16 bit and see if you can tell the difference.
Hint: You won’t be able to. If you import both versions into a DAW and reverse the polarity on one, they will null - i.e. literally no difference (to the human ear).
I agree, however, whatever the equipment, if a listener tells you that they can hear the difference between 24 bit and 16 bit, you can be confident in knowing that they are either mistaken or lying. The difference between 24 bit and 16 bit is in an inaudible noise floor. Other than that they are identical.
Is that OK for you Brian? Or do you think I need another reprimand for having the audacity to prevent indisputable facts?
Downsampling or down-dithering is not the same as recording with original bit depth. Just like upsampling wont add the original data that was never recorded to file, downsampling wont be able to effectively erase the extra data that was captured. Just like you cant see the photographer who took the photo by enhancing the photo, you cant erase the photobomber by compressing a digital photo to a lower resolution.
Digitally converted sample rates or bit depths usually aren’t decipherable by ear, but a recording in higher bit depth can possibly be detected by human ear, and there are some who can detect that. That is what I believe is being discussed and debated here.
These arent indisputable facts though, they are all debatable welcome back! and I hope you wont receive a slap on on the wrist from Holster
Precisely my point. You are detecting differences in two separate recordings, and you are erroneously ascribing those differences to the bit depth.
To be clear, we’re discussing bit depth, not sample rate. But to address your point, as far as I know, no-one has said anything about erasing data. We are discussing whether you can tell the difference between bit depths (whatever that difference may be).
Again, that is precisely my point. That is exactly what we are discussing.
But you would need to listen to that same recording in a lower bit depth in order to prove that point, and you have already said it’s not detectable.
The point you have made is that you can tell the difference between two separate recordings. I doubt anyone would argue that point. The discrepancy is that you are ascribing the difference to the the bit depth. I’m not sure why you would do that when you know by your own admission that you are listening a whole host of differences (simultaneous recordings), not just bit depth.
If you play two instances of the same file together, one at its original bit depth of 24 bits, and the other reduced to 16 bits, and reverse the polarity on one, they will null. That is an indisputable fact. Ergo it follows that no-one can hear the difference between 24 bits and 16 bits - because there is no difference.
Last night I went to go see the latest Avatar movie in the ATMOS equipped theater. What a great work of art, visually and sonicallly!
To somewhat inject Adrian’s points to the movie, we all know that Atmos audio is above the 44.1 standard but not at the highest rates possible.
It sounded awesome even adding to the realism of the scenes we were watching.
From what I have read, the frame rate while shooting this movie was also higher than the typical 24 frames in order to eliminate the blur of action scenes. However, as great as this might be, when there are human shots in the movie, they reverted back to 24 frames in order to not make it look like you’re watching a TV soap opera. Fascinating.
IMO, if the audio would have been at let’s say, a DSD rate, it might have taken your brain out of the movie scene. I wonder what the dialog was recorded at? Or were there different degrees of audio recording rates/ playback rates?
If we could do that with our audio, specifically music, our tracks would have a perception difference. Much like how a lo fi 8 bit song has a certain grind at sets the mood. Thanks.
there is no way to prove or disprove that. You are erroneously assuming that I have erroneously ascribed my findings to it lol (light pun back at ya )
Reality is that there is a difference in recordings because higher bit depth allows for bigger headroom “during” the recording process. There is a difference in recordings according to both of us. You are saying that it isnt because of bit depth and possibly an error or difference during the process. I am saying it is possible because of bit depth and not just a difference in recording. The pro who recorded the different bit depth, used a similar or another of the same interface so there is a chance that the process captured the sound properly. If I get a chance again I will repeat the test. I will be surprised if I cant tell the difference.
Hence the debates on highrez sound persist online because some people can hear and feel the difference.
I have a friend who is synesthetic, she can see colors around people and colors in music. They exist. There is a strong possibility that there are also people who can hear or “feel” the difference, while I might not be one of these as I cant tell differences in sample rates, but I can certainly sense the bit depths. I would like to prove myself wrong though! so I will try another similar test sometime this year and post the two versions here.
All you need to do to prove that to yourself is reduce your 24 bit recordings to 16 bit and compare them.
But it is not a test of bit depth. I don’t understand why you can’t see that. If you approached a university/lab with this as a proposal for discerning bit depth you would get laughed out of the room.
Great! Reduce your 24-bit recordings to 16 bit and compare.
Computer science and physics alumni here from an Ivy league, perfect scores all around, no I wont be laughed at if i approached a uni.
there is no other way it is not a test of bit depth during the recording process and not simply a test of sample rate conversions and dithering.
I have no issues with a good debate and yes you are not wrong for assuming that the differences could be in the process and not bit depth but you can not ascertain that for a fact.
that is what I am trying to say here, that it is not the same test. We aren’t testing algorithmic dithering or sample rate conversions, we both agree that converting sample rates to different format aren’t going to be decipherable by human ear in most cases (extreme cases aside). I never argued that.
We are testing the real thing. Original recording with the different bit depths or sample rates set in place to begin with. They are not the same as conversions later down the chain.
As I gave the photo example earlier. You cant add something that a photo does not have to begin with and you cant remove something significantly by lowering the resolution slightly.
You state that you can tell the difference between 24 bit and 16 bit. The test I described is the exact test you would be asked to undergo if you wanted to prove your claim in a lab. The only difference between 24 bit and 16 bit is 8 bits. As you’ve already stated, no-one can tell that difference, so it’s not correct to claim that you can.
Which produces two separate recordings!
Which means it’s not a true test, because it’s not falsifiable. It’s two separate recordings, which means you can make whatever claim you want about them without fear of contradiction. I could listen to the same recordings, pick out the 24-bit versions 90% of the time just like you, but then claim it was due to spiders running up and down one of the cables (the 24-bit one). Or more realistically I could claim it was due to a higher quality interface. Either way, my claim is just as valid as yours - because it’s not falsifiable.
I’m not sure what you’re trying to prove with your photograph analogy but you’re comparing apples to oranges. Image resolution is not comparable to audio resolution. With images, the greater the resolution, the higher the quality the end result. With audio, the resolution is the bit depth, which is the amount of headroom (measured in db) or basically, the size of the noise floor. The “quality” does not increase as bit depth increases.
correct, but on actual recordings and not digital sample rate conversions or re-dithering.
As stated before, they arent the same thing. You are saying they are, and that is fine. There are others who think that digital dithering is not the same as actual interface recording. It is ok if you disagree with it. Agree to disagree here
That is a matter of perspective. Higher bit depth and head room are some of the parameters established by some industry pros as quality in sound…and I agree with it. Another agree to disagree moment I guess. Thanks for an insightful debate nevertheless.
I get it. My point is that any discussion about resolution needs to be tempered by the fact that the very vast majority has no interest in it altogether, and will happily spend $1300 on a new phone while they listen to music through a tin can and a string.
welcome back! and I hope you wont receive a slap on on the wrist from Holster 
