Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Samples with high DR (Read 16607 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Samples with high DR

Hi.

Now I'm trying to find differences between hi-res and 16/44.1 (down-sampled hi-res). And I've heard the differences only on a simple synthetic samples like -60 dB sine, etc. And all my hi-res records even don't have dynamic range or noise floor higher (lower) than 80 dB (or even 60 dB). And I wonder is there any real records (e.g. instrumental music with live instruments) with dynamic range comparable to at least CDDA - 96 dB? Or maybe there exist some records with noise floor lower than -96 dBFS? I think they must be somewhere... If you know something like that, please make me know.

It would be ideal to find something adequate to DVD-A quality - DR more than 100 dB, and with high frequency range (but I'm completely not shure if frequencies >18-20 kHz have a real influence on a percieved quality).

So only thing I need is a HQ sample. Thanks.

P.S. I expect you will advice me to look at electronic music, but I repeat: I don't want synthetic samples, only live records (I mean live instruments).

🇺🇦 Glory to Ukraine!

Samples with high DR

Reply #1 –
Interesting topic. You may know about the nicely recorded and freely available http://www.2l.no/hires/ . I am really curious what your analysis reveals...

Samples with high DR

Reply #2 –
Downloaded 1 Gb Mozart: Violin concerto in D major - Allegro. Not the best results

Original:
Quote
Left   Right
Peak Amplitude:   -0,77 dB   -0,42 dB
True Peak Amplitude:   -0,77 dBTP   -0,42 dBTP
Maximum Sample Value:   7551362   7552040
Minimum Sample Value:   -7679330   -7994183
Possibly Clipped Samples:   0   0
Total RMS Amplitude:   -23,23 dB   -23,48 dB
Maximum RMS Amplitude:   -9,83 dB   -8,79 dB
Minimum RMS Amplitude:   -75,23 dB   -74,44 dB
Average RMS Amplitude:   -28,40 dB   -28,50 dB
DC Offset:   0,00 %   0,00 %
Measured Bit Depth:   24   24
Dynamic Range:   65,40 dB   65,66 dB
Dynamic Range Used:   39,60 dB   38,55 dB
Loudness:   -17,81 dB   -16,91 dB
Perceived Loudness:   -13,95 dB   -14,59 dB
ITU-R BS.1770-2 Loudness: -19,61 LUFS

0dB = FS Square Wave
Using RMS Window of 50,00 ms
Account for DC = true


Filtered to 44.1 samplerate:

Quote
Left   Right
Peak Amplitude:   -0,83 dB   -0,51 dB
True Peak Amplitude:   -0,81 dBTP   -0,49 dBTP
Maximum Sample Value:   29235,04   29366,94
Minimum Sample Value:   -29788,55   -30911,73
Possibly Clipped Samples:   0   0
Total RMS Amplitude:   -23,24 dB   -23,49 dB
Maximum RMS Amplitude:   -9,84 dB   -8,78 dB
Minimum RMS Amplitude:   -92,03 dB   -89,55 dB
Average RMS Amplitude:   -28,85 dB   -28,96 dB
DC Offset:   0,00 %   0,00 %
Measured Bit Depth:   32   32
Dynamic Range:   82,19 dB   80,77 dB
Dynamic Range Used:   60,00 dB   59,30 dB
Loudness:   -17,85 dB   -17,64 dB
Perceived Loudness:   -13,32 dB   -14,33 dB
ITU-R BS.1770-2 Loudness: -19,58 LUFS

0dB = FS Square Wave
Using RMS Window of 50,00 ms
Account for DC = true


Still not enough for 24vs16-bit testing
🇺🇦 Glory to Ukraine!

Samples with high DR

Reply #3 –
Unless the peak sound pressure is at a level damaging to human ears, there are few places on the planet where the background noise is 100dB below peak level. Hence there are few (if any) natural recordings with this dynamic range. That's before you start to think about microphones, their pre-amplifiers, humans moving and breathing etc.

Cheers,
David.

Samples with high DR

Reply #4 –
Hmm, interesting statement

But what about sound isolated rooms? I'm not sure it's impossible to remove from a sound recording room all sources of persistent noise. Also I don't consider human's breathing as noise, I think it is a part of a signal. It's not persistent and appears occasionally so must not affect the dynamic range.
🇺🇦 Glory to Ukraine!

Samples with high DR

Reply #5 –
What about the freely available NIN The Slip recordings? I don't know how it was made, but perhaps part is not sampled and therefore has no inherent noise (all-digital)?
Music: sounds arranged such that they construct feelings.

Samples with high DR

Reply #6 –
Quote from: ktf on
What about the freely available NIN The Slip recordings? I don't know how it was made, but perhaps part is not sampled and therefore has no inherent noise (all-digital)?


Looks like it was released on CD and vinyl only.
🇺🇦 Glory to Ukraine!

Samples with high DR

Reply #7 –
Many people here have accepted the fact that it is not possible to hear any difference between CD spec music and music recorded at a greater bit depth or sample rate (or between dithering and not dithering (properly) when reducing bit depth). Even with synthetic clips it is possible only on very low level signals that must be greatly amplified to be heard.

I came to this tentative conclusion ten years of more ago by my own experiments. On a number of different audio forums I began asking for MUSIC samples where differences might be heard every time someone started proclaiming an advantage to higher resolution of any sort. I continued doing so for years. I never got any.

There have been a few occasions where some one person claimed to be able to successfully ABX some particular piece of music. I eventually became bored and stopped following these discussions in detail, so I don't know for certain, but their perceived differences always seemed to come down to some equipment or resampling difficulties creating a real difference in their own experiments, and thus not experienced by anyone else.

Samples with high DR

Reply #10 –
Quote from: Steve Forte Rio on
But what about sound isolated rooms? I'm not sure it's impossible to remove from a sound recording room all sources of persistent noise. Also I don't consider human's breathing as noise, I think it is a part of a signal. It's not persistent and appears occasionally so must not affect the dynamic range.


Have you looked at the SNR specs of microphone input stages? You won't find many with more than 96 dB SNR...

Samples with high DR

Reply #11 –
Quote from: Kees de Visser on
Quote from: Steve Forte Rio on
Measured Bit Depth:   32
It looks as if you forgot to dither the 44.1 kHz version to 16 bits. If you want to compare 24 to 16 bit, this is crucial.


It was just resampling with 32-bit float output - to filter the high-frequency noise and see what the DR became.
🇺🇦 Glory to Ukraine!

Samples with high DR

Reply #12 –
Quote from: Steve Forte Rio on
Quote from: ktf on
What about the freely available NIN The Slip recordings? I don't know how it was made, but perhaps part is not sampled and therefore has no inherent noise (all-digital)?

Looks like it was released on CD and vinyl only.

Nope, you can download it for free here, 44/16 and 96/24 files: http://theslip.nin.com/
Music: sounds arranged such that they construct feelings.

Samples with high DR

Reply #13 –
Well, thanks. But nothing special, maximum 75 dB of DR.
🇺🇦 Glory to Ukraine!

Samples with high DR

Reply #14 –
Quote from: Steve Forte Rio on
Hmm, interesting statement



Not just interesting but highly relevant.

Quote
But what about sound isolated rooms?


Not very many of them are available for making musical recordings.

The noise floor of rooms is often rated using the NC system:

http://www.engineeringtoolbox.com/nc-noise...rion-d_725.html

Quote
I'm not sure it's impossible to remove from a sound recording room all sources of persistent noise. Also I don't consider human's breathing as noise, I think it is a part of a signal. It's not persistent and appears occasionally so must not affect the dynamic range.


There is a law of signals analysis that roughly says if you add up a large number of uncorrelated sources, you end up with Gaussian noise. 100 musicians breathing would appear to be a large number.

The noise floor of good quiet  commercial recordings ranges from 65 to 85 dB below peak levels. The Swedish Bis is a label that has a number of very quiet recordings.  http://www.bis.se/index.php  I know their Beethoven Symphony series has a noise floor that is better than 80 dB below peak levels. Among popular recordings, several by Rickie Lee Jones have noise floors at least 75 dB below peak levels which is pretty good for pop.

Samples with high DR

Reply #15 –
Quote from: Arnold B. Krueger on
The noise floor of good quiet  commercial recordings ranges from 65 to 85 dB below peak levels.
Let's not forget that most of the noise in acoustical recordings is in the low frequency region, below say 300 Hz and can be up to 40 dB higher in level compared to the 3 kHz region where our ear is most sensitive. Peak meters are usually broadband, so when you see them move (unmusically) around -60 dBFS in quiet passages, it's most likely LF rumble. Luckily the ear isn't very sensitive in this frequency region.
I don't know how sophisticated DR calculations are but if we're interested in perceived loudness, different weighting curves should be used for loud and soft passages.
If the OP wants to hear differences between 24 and 16 bit, I'd suggest to search for noise differences (due to added 16-bit dither) where the ear is most sensitive (3kHz region). Some people claim differences like "depth", "width", "musicality". If that works for you in a double blind test, fine, but it never did for me.

Anecdote: many years ago I did a classical piano cd recording spanning 3 days. The digital 8-track Tascam recorder had a 24-bit mode which had to be activated with a button. On the second day the assistant forgot to press the button and so the 24 bit ADC's output was truncated and recorded to 16 bit. Needless to say I got mad at him when we found out, but during the editing I couldn't hear a difference. There's no excuse for sloppy practice and it has never happened again, but it made me realize that the difference between 16 and 24 bit is hardly relevant as long as no further processing is done.

Samples with high DR

Reply #16 –
There are a few audio tracks that "need" more than 16-bits, if you are to keep the digital noise floor well below the recorded signal's noise floor at all frequencies, and if you don't use noise shaped dither. However, people only manage to hear this difference by increasing the volume control during the quiet passages, and then reducing it during the louder ones. Keeping the same listening level (which doesn't cause discomfort/damage during the loud passages) means the 16-bit noise floor is virtually inaudible. Using noise shaped dither makes is quite comfortably inaudible.

I'm surprised no one has posted a reading list, e.g.

why you don't need more than 16-bits...
http://www.hydrogenaudio.org/forums/index....showtopic=93853

A search for samples where 16-bits sounds inferior to 24-bits...
http://www.hydrogenaudio.org/forums/index....showtopic=49843

typical dynamic range...
http://www.hydrogenaudio.org/forums/index....showtopic=79139

...and a recent fun discussion on whether 8-bits is enough...
http://www.hydrogenaudio.org/forums/index....howtopic=102940


I think it's worth mentioning that audiophiles who claim 24-bits is superior don't say this because they want to perfectly experience the tiny handful of recordings where the native noise floor is lower than -100dB. They say it because they believe 24-bits make all recordings sound better. Good luck finding objective proof of that, but that it a typical audiophile belief.

Cheers,
David.

Samples with high DR

Reply #17 –
Thanks for all your replies.

But anyway, maybe there is some sample of at least electronic (fully synthesized) music with very low noise floor (and wide DR of course)?
🇺🇦 Glory to Ukraine!

Samples with high DR

Reply #19 –
Hardly any of us have quiet enough rooms to need anything more than 16 bits, regardless of the nature of the music being acoustic or otherwise:

http://www.audiocheck.net/audiotests_dynamiccheck.php

Samples with high DR

Reply #21 –
According to a book I've read there's no need to go beyond 16 bits in the first place if jitter is not <= 200 ps, including among other specifications.  One bit and multi bit converters behave differently to jitter.

Samples with high DR

Reply #22 –
Quote from: IgorC on
Yes, I was gonna to post about one sample of electronic music.  http://www.hydrogenaudio.org/forums/index....showtopic=81467


Nothing special, just 78 dB of DR and 61 dB used.


Quote from: mzil on
Hardly any of us have quiet enough rooms to need anything more than 16 bits, regardless of the nature of the music being acoustic or otherwise:

http://www.audiocheck.net/audiotests_dynamiccheck.php


Thanks. Have tried it with my headphones the last digits I heard was 72 dB. But that's a voice, a complex signal. With tones about 1 kHz we can get much higher results. For 1-6 kHz frequencies I get results up to 100 dB from maximum non uncomfortable to minimum audible.
🇺🇦 Glory to Ukraine!

Samples with high DR

Reply #23 –
Quote from: Steve Forte Rio on
Quote from: IgorC on
Yes, I was gonna to post about one sample of electronic music.  http://www.hydrogenaudio.org/forums/index....showtopic=81467


Nothing special, just 78 dB of DR and 61 dB used.


Quote from: mzil on
Hardly any of us have quiet enough rooms to need anything more than 16 bits, regardless of the nature of the music being acoustic or otherwise:

http://www.audiocheck.net/audiotests_dynamiccheck.php


Thanks. Have tried it with my headphones the last digits I heard was 72 dB. But that's a voice, a complex signal. With tones about 1 kHz we can get much higher results. For 1-6 kHz frequencies I get results up to 100 dB from maximum non uncomfortable to minimum audible.



I presume that you may obtain more sensitive results if you move the tone's frequency to 4 KHz where the ear's sensitivity peaks according to Fletcher and Munson.

That all said, the results with voice or perhaps the worst case musical instrument might be more relevant to those with pure tones, which are relatively rare in nature.

Samples with high DR

Reply #24 –
Quote from: Steve Forte Rio on
With tones about 1 kHz we can get much higher results. For 1-6 kHz frequencies I get results up to 100 dB from maximum non uncomfortable to minimum audible

Be careful, my friend. Don't mistake "maximum that's not uncomfortable" with "safe to endure". Even very modest headphones are capable of playing loudly enough to cause hearing damage, yet without the typical distortion that warns us to back off when using loudspeakers in a room. It's all a function of exposure time [scroll half way down to "Permissible Exposure Time Guidelines ? Sound Pressure Level - SPL (Dose)]. It's also difficult for us hobbyists to accurately measure the actual SPL at the eardrum, when using headphones, even if we know the manufacturer's published sensitivity and the headphone amp's power. We can only get a rough estimate from such specs.