It is currently Fri Mar 29, 2024 1:05 am


Loading in a different bit rate, I noticed a difference

Existing and forthcoming Hauptwerk instruments, recommendations, ...
  • Author
  • Message
Offline

1961TC4ME

Member

  • Posts: 3144
  • Joined: Mon Feb 07, 2011 1:45 pm
  • Location: Lake Minnetonka, Minnesota

Loading in a different bit rate, I noticed a difference

PostFri Jan 20, 2012 1:54 pm

Hello All,

I know there's been endless debate on this here, and the intent is not to argue which is better or open up the can of worms again, more to simply report what I noticed and suggest trying it for yourself if you have not yet done so.

I was reading through the Sonus Paradisi site the other day, and read with interest that loading the Forcalquier (which I have) in a higher bit rate than 16 would be of advantage to make the full use of it's sound. I've never tried it on this or any other instrument and have always loaded in 16 bit only. Since I have the newer computer with 8GB of RAM, I figured why not? Give it a try and see what happens. I decided my computer would probably be limited to load in "20 bit 32 bit aligned" (and I was indeed close with about .8 GB left after it loaded), so away I went.

Here's what I heard in differences and I wouldn't say it neccesarily made an improvement, more just a few subtle changes and would be one of those what sounds good to one, might not to the next scenarios.

It sligtly softened the sound, as if a person turned down the treble some, just slightly less hiss to the sound, a little cleaner souding. The biggest change I noticed is how it brought up the low end, in fact so much I had to go in and do a bit of re-voicing to tone it down some. The bass was really brought out and again for some this might be desireable and for others not, in my case I was actually looking for a little more low end, so for me this part worked.

Other than that, not much of a difference to my ears and I could actually use it either way.

And there you have it, my 2 cents worth.

Regards

Marc
Offline
User avatar

schantzplayer

Member

  • Posts: 485
  • Joined: Tue Jun 22, 2010 4:50 pm
  • Location: Mifflinville, PA USA

Re: Loading in a different bit rate, I noticed a difference

PostFri Jan 20, 2012 2:55 pm

I usually load instruments in at full 24bit resolution and the difference in signal to noise ratio is apparent. With 32 gb Ram I will have to compromise should the samples get any larger.
Offline

1961TC4ME

Member

  • Posts: 3144
  • Joined: Mon Feb 07, 2011 1:45 pm
  • Location: Lake Minnetonka, Minnesota

Re: Loading in a different bit rate, I noticed a difference

PostFri Jan 20, 2012 3:06 pm

Yea, I'm going to have to go out and purchase more memory one of these days, I have Win 7 Pro so I can go to 32GB, just haven't done it yet.

I was curious to see what 24 bit would do but I know I'd run out of memory before the set would fully load. It definitely takes a lot more memory! 20 bit made an interesting change, but as I reported I found myself once again revoicing some ranks. The Forcalquier is a rather bright organ out of the box, and loading in a higher bit rate for this set I can see may have it's advantages depending on the expected outcome, and might also make it easier to voice.

I wonder from set to set what differences there might be? For this set it was much more in the lows that were changed. I suppose for others it could be competely the opposite? In any case it was interesting. :D

Marc
Offline

ludu

Member

  • Posts: 998
  • Joined: Wed Nov 02, 2011 8:52 am
  • Location: Tournai (Belgium)

Re: Loading in a different bit rate, I noticed a difference

PostMon Jan 23, 2012 1:58 am

Playing big instruments needs a lot of memory and sometimes 24 or 20 bits are not possible in stereo, even with a ram of 32 Gb. What do you mean the most important in this situation, 24 bits mono or 16 bites stereo?
Offline
User avatar

mdyde

Moderator

  • Posts: 15444
  • Joined: Fri Mar 14, 2003 1:19 pm
  • Location: UK

Re: Loading in a different bit rate, I noticed a difference

PostMon Jan 23, 2012 6:15 am

Hello Ludu,

In my personal opinion loading a wet sample set in 16-bit stereo will give a *much* more realistic result than loading it 24-bit mono. Although 20/24-bit may give less background noise than 16-bit with wet samples (and is thus probably worth doing if you have ample RAM), loading in stereo vs. mono gives a huge difference in realism, so is the more important of the two.
Best regards, Martin.
Hauptwerk software designer/developer, Milan Digital Audio.
Offline
User avatar

coupler

Member

  • Posts: 182
  • Joined: Sat Feb 14, 2009 8:57 am
  • Location: Perth, Australia

Re: Loading in a different bit rate, I noticed a difference

PostMon Jan 23, 2012 10:41 am

1961TC4ME wrote: I have Win 7 Pro so I can go to 32GB


The physical memory limit for Windows 7 Pro is 192GB. The limiting factor will most likely be your motherboard or your wallet. (Mine is both! :mrgreen: )
Cheers,

David Russell
Offline

1961TC4ME

Member

  • Posts: 3144
  • Joined: Mon Feb 07, 2011 1:45 pm
  • Location: Lake Minnetonka, Minnesota

Re: Loading in a different bit rate, I noticed a difference

PostMon Jan 23, 2012 11:39 am

Hi David,

Same here!

I probably should have said from what I learned, anything less than Win 7 Pro the max is 16GB, and Win 7 Pro will support 32GB. How much beyond that I had no clue. 192GB? Wow! I haven't looked in awhile but I think I have 4 memory slots. Not that I would ever attempt it (or afford it), but are the memory sticks even available to get to 192 or is it more a matter of having a certain number of slots, then the available memory?

Marc
Offline
User avatar

toplayer2

Member

  • Posts: 1071
  • Joined: Mon Jul 23, 2007 2:15 pm
  • Location: Michigan, USA

Re: Loading in a different bit rate, I noticed a difference

PostMon Jan 23, 2012 12:20 pm

24 bits has a theoretical noise floor of-144 dB vs -96 dB for 16 bits. The higher bit rate will produce less hiss provided that the individual samples have a signal-to-noise better than -96 dB. My own experience (having sampled many organs) is that I have yet to find a sample with anything approaching -96 dB even after Noise Reduction has been applied. This being the case, and since embedded noise is additive in the same way as quantization noise when many notes are sounding, it is hard for me to offer any rational argument favoring higher bit rates. Perhaps an extraorniarily quite wet sampled organ will have reverb tails with inherent noise below -96 dB and if so there would be an audible reduction in hiss in the reverb tails.

Beyond reduced noise, there is no quantifiable benefit to higher bit rates. This may seem counter-intuitive for us when we think about how much sharper a high megapixel digital photograph looks, but it is a false analogy to think of higher bit depths as being higher resolution.

Perhaps another cause is behind perceived changes in bass response. One phenomenon that we all share is what Ethan Winer calls "expectation bias". A very entertaining series of demonstrations of this are here:
http://www.youtube.com/watch?v=BYTlN6wjcvQ

Best regards,
Joe Hardy
Offline

etcetera

Member

  • Posts: 100
  • Joined: Thu Dec 03, 2009 4:56 pm

Re: Loading in a different bit rate, I noticed a difference

PostMon Feb 06, 2012 9:17 am

toplayer2 wrote:24 bits has a theoretical noise floor of-144 dB vs -96 dB for 16 bits. The higher bit rate will produce less hiss provided that the individual samples have a signal-to-noise better than -96 dB. My own experience (having sampled many organs) is that I have yet to find a sample with anything approaching -96 dB even after Noise Reduction has been applied. This being the case, and since embedded noise is additive in the same way as quantization noise when many notes are sounding, it is hard for me to offer any rational argument favoring higher bit rates.


Hi Joe!

So true.
Below is an image showing a waveform in db for one note of a published organ sample recording originally recorded at 96 kHz/ 24 bit. I tend to record with enough latitude mainly to be able to do post processing adjustments safely.
As seen on the image the initial noise level is pretty high, as it is in most churches. I have seen noise levels well beyond this sample. Some of the noise is very low frequencies, and on the upper part I did a hi-pass filtering removing the frequencies below this notes fundamental (much of it was really low frequency rumble below 100Hz), but the section labelled "Hipass 320Hz" still shows a noise level that is not acceptable for publishing.

Image

The lower part is after the main Noise Reduction is applied. In front of and behind the sounding note you see that most of the Noise seems to be gone. To see what still remains, I increased the volume with 50db in a silent part behind the note. On the first part of that amplified section I also removed the frequencies below the fundamental, while on the second part of the amplified section I did no such filtering, so the low freq. rumble is still there too.

As can be seen from the charts, the signal/noise level after Noise reduction is not anywhere near -96 db. Getting the Noise level any better than this sample shows will IMO start reducing the quality of the recorded sample. Which some are willing to do, but I do not think that gives an improved fidelity, IMO it rather starts picking apart the "soul" of the recorded sample.

With a remaining noise level like this, I must admit that publishing a 24 bit organ sample is not much of a point.
Yes - there are recording situations out there where it "could" have had a purpose. IMO an organ sample recording is not one of those situations...

regards
-etc

Return to Hauptwerk instruments

Who is online

Users browsing this forum: No registered users and 12 guests