recording at 24 bit 48k?

[TyphoidHippo]

I misused "upsample" and "downsample" at least a couple of times, when I should have said "upcast" or "downcast", and possibly vice-versa. I'll look at that post more tomorrow to see if it matters....g'night hr...

 

[Farview]

Are you sure that the extra 8 bits are used for "extra loud signals" only? I think this is an inaccurate statement for the same reason I think Boulder's was - just at the other end of the spectrum (and with the same caveat that I kinda doubt it even matters.... I just don't think you're right about what the extra bits are used for).

I'm pretty sure he was trying to say that it would only be useful in situations where you need a huge dynamic range. In other words, a situation where you might expect huge, unexpected volume jumps over the average level.

I'm not sure it exactly relates to how DAW's work, but this is exactly the reason I always upsample to floats or doubles (64 bit floats) when doing any DSP programming - so there's not a clamp to 16-bit values at every stage - you only have to do that once at the end of the chain. Of course... the DSP programming I've done always uses 16 bit source material, so I think there may be a difference when taking into account higher resolution sources....but your point stands, and I agree with it - except that I would say it *does* sound better, to not downsample to 16 bits, repeatedly, every step of the way. Audibly, literally, better - but I think most DAWs use 32 or 64 bit internal processing anyway, which is basically the same....in which case - I still don't know if the source depth *really* mattered beyond 16 bits. I suppose some math examples might help.... if you start with x[n] and do y,z,a,b to the n samples in x, do you end up with a more accurate sample if x was sampled at 16 bits or 24? Almost certainly 24...but I don't know if it *really* matters, to be honest, since it's upsampled to 32 or 64 from the beginning, regardless, and we can't really distinguish more than is expressible in 16 bits.

The reason it would make a difference to use 24 bit is when you do offline processing and rewrite the file. In that case, you are probably better off having the extra bits to collect any rounding errors and such.

 

[bouldersoundguy]

The reason more bits (greater word length) doesn't mean higher resolution is that each bit still represents a 3dB change. A signal recorded at -6dBFS looks exactly the same at any word length until you get down to the noise floor.

 

[TyphoidHippo]

I'm pretty sure he was trying to say that it would only be useful in situations where you need a huge dynamic range. In other words, a situation where you might expect huge, unexpected volume jumps over the average level.

Ok, I think that makes sense to me. If, for some reason, you had a much louder signal at some point than the average, having the extra resolution would still let you get an accurate recording of the average parts and the super-loud parts....whereas without it, and the loud part required all 16 bits, the much quieter part might be forced to be recorded using only a few of the least significant bits and would be subject to artifacts and such....right? Makes sense to me, yea.

The reason it would make a difference to use 24 bit is when you do offline processing and rewrite the file. In that case, you are probably better off having the extra bits to collect any rounding errors and such.

I forget about offline processing sometimes (I never do it). I think I understand your point about it, though - it would be kind of the same reason that it's ideal to not use 16 bits in real-time processing - truncating down to 16 repeatedly causes a real, tangible difference, but if your offline processes have a 24 bit file come out each step... well that would be better.

...interesting thread, so far

 

[TyphoidHippo]

The reason more bits (greater word length) doesn't mean higher resolution is that each bit still represents a 3dB change. A signal recorded at -6dBFS looks exactly the same at any word length until you get down to the noise floor.

Well then that just flips everything upside down...and seems like an illogical use of the extra data. Isn't the noise floor, as it relates to usable signal levels, still going to be dependent on the hardware? Why would you want to only describe noise in greater detail? Are you sure about this?

 

[bouldersoundguy]

Well then that just flips everything upside down...and seems like an illogical use of the extra data. Isn't the noise floor, as it relates to usable signal levels, still going to be dependent on the hardware? Why would you want to only describe noise in greater detail? Are you sure about this?

But the whole point to 24 bit is that you don't have to record as close to 0dB to stay above the noise floor. You can record with 12, 16 or even 18dB of digital headroom and not get near the noise floor. Your analog noise floor is always above the digital one and you have plenty of headroom left for processing. It's like driving in a lane that's extra wide instead of driving in one that's just two feet wider than your car. Sure, you can fit in that narrow lane but if there's some debris on the road or someone drifts into your lane it's nice to have some spare space to maneuver.

But no matter what level you record at 1 bit is always 3dB.

 

[TyphoidHippo]

Ok, it's becoming more clear now - describing quieter signals in more detail is good specifically because it allows you to record quieter signals and still retain as much information about the original source as if you had used all 16 bits and made sure that the loudest signal was exactly 0xFFFF (all 16 bits set to 1). So you record quieter, get more headroom for processing, still get as accurate of a recording as if you had recorded a "hot" 16 bit recording, and your quietest real signal is still far from the noise floor.

I suppose this would also nullify (or at least minimize) the rounding errors and artifacts of using the less significant bits to record audible signals (as has been said many times here - but now I really see why, I think).

...right?

 

[Ethan Winer]

Are you sure that the extra 8 bits are used for "extra loud signals" only?

You misread what I said. The extra bits of course come into play at the bottom of a sample word, not the top. But using 24 bits instead of 16 lets you record at a lower level overall, to account for unexpected volume from an enthusiastic player.

It's a myth that 24 bits has more resolution than 16 bits, other than a theoretically lower noise floor. The notion that 24 bits has smaller steps is incorrect, and misses how a D/A converter's reconstruction filter works. Further, the noise floor of any recording - especially one made in a home studio - is usually limited by the ambient noise of the room and heating / AC system. Further, no 24-bit converters actually deliver 24 bits of digital audio. They can't even do that in theory due to the thermal noise floor (Johnson noise) which is around -131 dB for a 20 KHz bandwidth. This is probably more technical than anyone here cares about:

Johnson Noise at Wikipedia

Further still, even 16 bits is 20+ dB quieter than analog tape, which is almost clean enough for pro results.

--Ethan

 

[TyphoidHippo]

You misread what I said.

I misunderstood it, for sure. I think the post I made while you posted this is closer to a correct understanding. I'm seeing and understanding about the resolution myth now. This is good stuff.

The notion that 24 bits has small steps is incorrect

This was exactly where I was going wrong in my head.

Further still, even 16 bits is 20+ dB quieter than analog tape, which is almost clean enough for pro results.

--Ethan

lol

 

[Farview]

The reason more bits (greater word length) doesn't mean higher resolution is that each bit still represents a 3dB change. A signal recorded at -6dBFS looks exactly the same at any word length until you get down to the noise floor.

The thing is, you get down to the noise floor on every cycle. When you swing from negative to positive and back, you go through zero. The shorter the word length, the more dynamic range between the noise floor and zero, or the less low level resolution.

I believe that 1 bit represents 6db of dynamic range.

 

[TyphoidHippo]

The thing is, you get down to the noise floor on every cycle. When you swing from negative to positive and back, you go through zero. The shorter the word length, the more dynamic range between the noise floor and zero, or the less low level resolution.

That's what I gathered to be a logical inference from what he said (meaning I think you both said the same thing, just from different angles... or did you not?). This is all making more sense (at least to me) now. I still feel like I'm missing something, though, but I'm not sure how to word it... Something about the relationship between bit depth and convertor accuracy or something - I'll think more about what I'm actually wondering about and ask later, I suppose.

I believe that 1 bit represents 6db of dynamic range.

The wikipedia article on digital audio bit depth says that, too, or at least seems to imply it.

 

[bouldersoundguy]

I believe that 1 bit represents 6db of dynamic range.

Something like that. One bit represents a doubling of amplitude, which in voltage is 6dB and in power is 3dB. I think it's actually 6.02dB and 3.01dB, but those details don't help me decide how to eq a snare or compress a vocal so sometimes I forget.

 

[craveonmatter]

what is the difference between 16bit and 24bit cps are only 16 i think

 

[TyphoidHippo]

...what?

 

[Farview]

what is the difference between 16bit and 24bit cps are only 16 i think

I'm not even sure this would make sense with proper punctuation.

 

[Massive Master]

I *think* he meant: "What is the difference between 16-bit and 24-bit? CDs are only 16 I think..."

 

[RockinRobby]

Think of it like a digital camera? Because that's what it is, it's capturing "a picture" of your sound digitally? So cams w/more megapixles produce crisper images.

 

[bouldersoundguy]

The thing is, you get down to the noise floor on every cycle. When you swing from negative to positive and back, you go through zero.

The signal passes through zero frequently but very seldom does a sample actually land on zero.

 

[TyphoidHippo]

Think of it like a digital camera? Because that's what it is, it's capturing "a picture" of your sound digitally? So cams w/more megapixles produce crisper images.

That's a weird analogy. They're both converting tangible information into digital data, and they both have a resolution and frequency - but the similarity ends there. All input devices share those traits, including your computer's mouse. I wouldn't compare an audio a/d converter to a mouse or a camera...the source information, processing and digital representation are all completely unrelated.

 

[Tim Gillett]

The thing is, you get down to the noise floor on every cycle. When you swing from negative to positive and back, you go through zero. The shorter the word length, the more dynamic range between the noise floor and zero, or the less low level resolution.

I believe that 1 bit represents 6db of dynamic range.

I know it's a long used convention to speak in this way but I suggest that speaking of that line half way between positive and negative swings as "zero" or "the zero crossing point" only adds to the confusion here, especially when we already speak of the digital clip point, the maximum useable level, as a zero. If we do speak of the clip point as zero, then there is no meaningful "zero" down there at all, just points a certain number of db below clip point.

Ethan spoke of a top converter having Johnson noise at -130dbFS. Think of that as the outer borders, both positive and negative, in a band which extends equally above and below the middle line. Anything quiet enough to be within that band is effectively overwhelmed by that noise, so we have to speak of that entire band as "indeterminate". The system cannot resolve it.

The 24 bit point, at -144dbFS, can similarly be thought of as a band but just a narrower one.

Better, I suggest, not to speak of "going through zero" on each cycle but crossing into the "indeterminate zone" - and every audio system has such a zone - because any point within that zone is beyond the resolving power of the system.

Once you get into this unresolveable zone, it's not possible to know at any given moment whether the audio wave is either positive or negative, let alone give it a value.

The midway line can only be spoken of in actual level terms as "minus infinity dbFS." or in digital terms "infinite bits". It's a theoretical construct, not a reality we can measure.

Speaking practically, the system, like a chain, can only ever be as good as its weakest link. In noise terms, if you are recording at 24 bit resolution and with converters equivalent to 20 bits but any other part of the chain, including the aircon noise in the building, is at say 12 bits, then 12 bits is the maximum dynamic range you can resolve. In this case, recording at 16 bits would sound no worse. You would lose nothing.

Hope this helps.

Tim