Head-Fi.org › Forums › Equipment Forums › Sound Science › 24bit vs 16bit, the myth exploded!
New Posts  All Forums:Forum Nav:

24bit vs 16bit, the myth exploded! - Page 3

post #31 of 1154
3/19/09 at 6:27pm
A quick look at wikipedia seems to discount the 160db sound is deadly thing.

If that were true, anyone who had fired a M1 rifle would be dead or deaf.
Sound pressure - Wikipedia, the free encyclopedia
Reply

Head-Fi's Best Sellers

post #32 of 1154
3/19/09 at 6:51pm
Quote:
Originally Posted by nealric View Post
A quick look at wikipedia seems to discount the 160db sound is deadly thing.

If that were true, anyone who had fired a M1 rifle would be dead or deaf.
Sound pressure - Wikipedia, the free encyclopedia
You can also draw your hand through a flame for a fraction of a second without any damage (I know this because I dragged my hand through a bunsen burner this morning, and all I got from it was burned off hair), that doesn't mean that the flame isn't damaging. 160dB from a rifle would last a microsecond. Whether or not an energy source causes damage will be a function of the intensity of the energy source, the length of exposure, and the ability of the energy's destination to dissipate that energy.
Reply
post #33 of 1154
3/19/09 at 6:57pm
Thread Starter 
Quote:
Originally Posted by nealric View Post
A quick look at wikipedia seems to discount the 160db sound is deadly thing.

If that were true, anyone who had fired a M1 rifle would be dead or deaf.
Sound pressure - Wikipedia, the free encyclopedia

There are no absolutes with these dB figures when it comes to the level at which someone feels pain, goes deaf or is killed, we are all individuals. For example, try firing an M1 rifle 1m away from your granny and see if she dies!!

It's usually accepted that at about 120dB - 140dB pain will be felt and permanent hearing damage is likely to occur. Persumably anyone firing an M1 rifle would have to be wearing some kind of hearing protection to avoid serious hearing damage. 180dB is usually the figure quoted for causing death but presumably someone with any kind of heart disorder could be killed by significantly less than this.

In my original post I mentioned that actually expeiencing 144dB above a 50db noise floor (194dB) would likely kill you.

G
Reply
post #34 of 1154
3/19/09 at 7:15pm
Saving that wall of text until I find myself to be a little brighter.
Thanks for your work.
Reply
post #35 of 1154
3/21/09 at 10:55am
Thread Starter 
Quote:
Originally Posted by ILikeMusic View Post
Ya think..?
Strange, I was expecting more dissenters!?

G
Reply
post #36 of 1154
3/21/09 at 11:50am
Here's a link to two music files, both from the same song and mastering session, in standard 16/44 and hi-res 24/96. You tell me if there's difference:

LINK


When comparing the files, be sure your system is not performing any sort of resampling and/or dithering to either one. Some will by default, either upsample the 16/44 or downsample and dither the 24/96.
Reply
post #37 of 1154
3/21/09 at 12:26pm
...
Reply
post #38 of 1154
3/21/09 at 12:38pm
Quote:
Originally Posted by gregorio View Post
Strange, I was expecting more dissenters!
You did too good a job with your explanation! But try writing a similar treatise on why copper conductors can't sound 'warmer' than silver and see what happens...
Reply
post #39 of 1154
3/21/09 at 12:57pm
Thread Starter 
Quote:
Originally Posted by nick_charles View Post
I have downloaded these samples in the past, they are not identical, they are not perfectly aligned and there is a recording artifact on the high rez sample. When you analyze them.....

By either measure there is an average 4db difference between the two samples.
Hi Nick, thanks for pointing that out. I have to say that most of the examples I've seen, where people are asked to compare 24bit and 16bit are in some way slightly 'loaded'. They usually have a point to prove or a product to sell.

I remember when the consumer demand for 24bit started. Most of us thought that consumers were a bit deluded but obviously there were/are those who feel it's an opportunity for a new marketing strategy and that there's money to be made. This whole hi-rez thing is a great opportunity for those who want to make money for nothing, not dissimilar to some of the cable retailers out there!

I've been using higher than 16bit recording since 1992 but we just called it 20bit or 24bit. 24bit only started being called 'hi-rez' nearly a decade later and co-incidentally when consumer demand started!!

Quote:
Originally Posted by ILikeMusic View Post
... try writing a similar treatise on why copper conductors can't sound 'warmer' than silver and see what happens...
I'd love to but unfortunately, my expertise is only in the field of audio, rather than psychology!!

G
Reply
post #40 of 1154
3/21/09 at 1:42pm
Wow, that cleared a lot up. Thanks
Reply
post #41 of 1154
3/21/09 at 1:50pm
Quote:
Originally Posted by gregorio View Post
Hi Nick, thanks for pointing that out. I have to say that most of the examples I've seen, where people are asked to compare 24bit and 16bit are in some way slightly 'loaded'. They usually have a point to prove or a product to sell.
oops, I noticed an error in my comparison which is why I removed my post. However if you load both up in Audacity as 16/44 (downsampling the 2496) and then again as 24/96 (upsampling the 1644) (I used triangular dither) and plot the spectra of both samples in both cases using a 2048 or 4096 FFT, export the figures to a text file load up in Excel and graph them and compare you get some interesting artifacts...
Reply
post #42 of 1154
3/21/09 at 2:12pm
Thread Starter 
Quote:
Originally Posted by nick_charles View Post
oops, I noticed an error in my comparison which is why I removed my post. However if you load both up in Audacity as 16/44 (downsampling the 2496) and then again as 24/96 (upsampling the 1644) (I used triangular dither) and plot the spectra of both samples in both cases using a 2048 or 4096 FFT, export the figures to a text file load up in Excel and graph them and compare you get some interesting artifacts...
There are always going to be slight measurable differences between the 24bit and 16bit versions. However, if the conversion is done correctly these differences will be undetectable when listening. Now, how they got to 16bit from the 24bit master is another issue. Has it been truncated, rounded or dithered, has a good quality noise-shaped dither been used? Has the 24bit version been mastered separately from the 16bit version? Has noise-shaped dither been used on the 24bit version and again on the 16bit version? How has the sample rate conversion been handled? A professional would not use 96kFs/s for a product that will be used on CD. If a higher sample rate is required (due to weak 44.1k filers) then 88.2kFs/s should be used as there is less math involved in getting back to 44.1k and less chance of errors.

There are all kinds of issues which could (theoretically) make a 16bit version sound (potentially) worse than a 24bit version but none of them due to an inherent weakness of 16bit format compared to 24bit.

G
Reply
post #43 of 1154
3/21/09 at 3:02pm
Quote:
Originally Posted by ILikeMusic View Post
You did too good a job with your explanation! But try writing a similar treatise on why copper conductors can't sound 'warmer' than silver and see what happens...
I cannot say with any certainty that any copper cable cannot have a more rolled off high frequency response than any silver cable.

What I can tell you with some certainty is that the Frequency response for the cables (several stranded copper, several solid copper, one Silver plated copper and one stranded silver cable) I have empirically tested in my system using my CD player and my ADC have not shown significant measurable deviations either from a notional flat frequency response or from each other's frequency reponses either in amplitiude or frequency terms.

This allows me to conclude that in my system none of these cables exhibit notably different frequency "signatures". Make of that what you will.
Reply
post #44 of 1154
3/21/09 at 3:44pm
Quote:
Originally Posted by oldschool View Post
Here's a link to two music files, both from the same song and mastering session, in standard 16/44 and hi-res 24/96. You tell me if there's difference:

LINK
Thanks for the link! Finally some native 24/96 files to play with...

I was using my E-MU 1212M for auditioning. The two files vary slightly in maximum amplitude as well as with arbitrarily tested sample points, but the deviation is within 0.04 dB. To exclude any difference other than such induced by the format, I additionaly upsampled and downsampled (...downsampled and upsampled) the two files, using WaveLab Lite. For valid listening tests the E-MU 1212M has to be set to either 44.1 or 96 kHz.

At 16/44.1 both files sounded identical to me: the (low-rez) original and the file downsampled from 24/96. At 24/96, the hi-rez original sounded more 3-dimensional and had a finer overtone sparkle on the upmost treble. It's certainly not a night-and-day difference and only noticeable after intensive trial, but it's there. It's the same (kind of) difference I notice by comparing DVD-As with CDs.

With respect to the thread topic -- I don't think the difference has much to do with the higher bit depth, rather with the increased sample rate (actually I forgot testing dynamic resolution separately).

Headphone used: AKG K 701, connected directly to the (modified) E-MU line out.
.
Reply
post #45 of 1154
3/21/09 at 5:00pm
Just asked an hearing aid researcher and expert I know, he asked around a little to be sure. The generall consensus is that there is very little reason to believe spatiel cues we hear should exsist above 20-22KHz since there is not really any natural source which would provide the high energy sound needed to for the ear to physically be able to receive it. The very same reason is why we don't hear that well compared to some animals. While generally sound energy is manly generated within our hearing spectrum many animals communicate at much higher frequency - so they have to do both well.

How ever as its writte by many experts in the field and a few said here already there is techincal reasons both in recording and playback which can make high sampling rate easyer to sound better.

Quote:
Finding out how high a sample rate you need to get all the spatial cue's with you is hardly something anyone have been able to conclude on due to its on unconscious level and spatial perception is not fully understood at all. I'm studying this topic and I haven't at seen any research into this which concentrated on this at all.

But I don't know everything and there is certainly many who knows more about this - I'm actually going to contact some experts and see if they know anything.
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science

Head-Fi's Best Sellers

Head-Fi.org › Forums › Equipment Forums › Sound Science › 24bit vs 16bit, the myth exploded!