R/E/P Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: [1]   Go Down

Author Topic: "Even" Dithering ?!?  (Read 2987 times)

leeberkgrad08

  • Newbie
  • *
  • Offline Offline
  • Posts: 4
"Even" Dithering ?!?
« on: April 05, 2007, 02:22:48 am »

       

 So Ive heard from a reliable source that if one's intended destination sample rate of a project is say... 44.1khz (redbook cd), then it would be advantageous to record in an even multiple of that (44.1, 88.2, 176.4) rather than having to dither down from say 48, 96, or 192.

 Since hearing this, I have also heard that this is mainly a problem within pro tools and that if a good external dithering device is used, the sample rate conversion does not become an issue.  

 Any thoughts?

  -Adrian Olsen
Logged

JDNelson

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: "Even" Dithering ?!?
« Reply #1 on: April 05, 2007, 01:38:16 pm »

I don't think that dither is a factor in these cases, what you're asking is whether the sample rate conversion is performed adequately.  Given that it's just mathematics I would think it doesn't matter... assuming that you use a good SRC algorithm of course.

PookyNMR

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1991
Re: "Even" Dithering ?!?
« Reply #2 on: April 05, 2007, 04:31:07 pm »

And if you're smart, you'll let your mastering engineer worry about doing the SRC...

Logged
Nathan Rousu

flanger_folly

  • Full Member
  • ***
  • Offline Offline
  • Posts: 129
Re: "Even" Dithering ?!?
« Reply #3 on: April 05, 2007, 05:04:27 pm »

You are talking about sample rate conversion, as opposed to changing the word length/bit depth- which may sound better when dither is added to the signal.

As far as src goes, the logic of the integers matching up isn't really accurate across the board b/c of "oversampling"... I believe that most boxes basically bounce the signal UP to a rate that can then be divided back down evenly. May be wrong on that, but it is what is in the ol' memory banks...

There has been many a long drawn out post on this... some probably numbering 30+ pages back in the old musicplayer.com website. There have been arguments back and forth. IMO, each situation can be unique, b/c you can never be too sure about the way that various peices of gear interacts. The only real answer is to actually test your particular setup. Anything else is hypothesis.
Logged
George Toledo III
X The Unknown
www.myspace.com/xsongs

littlehat

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1257
Re: "Even" Dithering ?!?
« Reply #4 on: April 05, 2007, 05:07:43 pm »

I think another angle is looking at plugin math and DAW mixers, which have to crunch far more numbers than a stereo converter.

The myth is that using whole numbers for your levels and settings in a plugin make the numbers coming out easier to compute. I'm not a big believer since the signal coming in is of unmeasured level and bit depth anyway.

I WOULD say that you should mix to 24bit SDII in PT when you know it's going to a good mastering engineer though.

When in doubt...

http://www.digido.com/

Ask the man.

Logged
Push the RED button!

Will Russell

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 55
Re: "Even" Dithering ?!?
« Reply #5 on: April 06, 2007, 07:42:37 am »

Hi,

Your reliable source is mis-informed. The same math is applied regardless of the source sample rate. 88.2->44.1 is no better than 96->44.1.

Dither has nothing to do with sample rate.
Logged
Will Russell
Electric Wilburland Studio
http://www.wilburland.com

jigsawlogic

  • Newbie
  • *
  • Offline Offline
  • Posts: 6
Re: "Even" Dithering ?!?
« Reply #6 on: April 07, 2007, 08:28:45 am »

Sounds like a question you should put up in Bruno's forum.

Firstly, you shouldn't confound terms.  Sample rate and bit depth are two different aspects of the same thing.  Rate being how often a sample is measured, and bit depth the level of detail of each measurement.  Dithering only has to do with the latter, when converting from a higher to lower bit depth.  

An AD converter measures a voltage at regular time intervals (ideally).  Each measurement is quantized to the nearest bit value, and the time at which the measurement happens is correlated to a clock.

To answer your question, you are half right.  Only half, because you are taking two variables into account when only one is necessary.  Bit depth has nothing to do with your question, so put that aside for now.

When it comes to sample rate, you're right.  If you are going to downsample to CD quality, it makes much more sense to sample at an integer multiple of 44.1 (like 88.2).

It helps to think of two grids with the same bit depth, but different sample rates.  Think of bit depth as resolution in the y direction, and sample rate as resolution in the x direction (where x is time, and y is a measurement of voltage)

If your sample rate is an integer multiple of the destination rate (88.2 vs 44.1 in this case), then there are no mathematics which have to be imposed on the waveform.  Since every other sample lines up perfectly between grids in the time domain, the SRC only has to drop the sample in between.  If your sample rate is 3 times 44.1, then you would drop every 2 samples.  The reasoning is that if your integer is n, then every 1/n will line up perfectly with the waveform you are downsampling.  Ditch the extra samples, and the analog circuitry of your DAC constructs a smooth waveform which lines up exactly with your original file up to a certain frequency.

To see why a non-integer multiple grid doesn't work as well, it helps to think in only 1 dimension, namely time.  For instance, let's take 96K and 44.1K.  Let's say it takes 2,000 samples before our grids land on the same point in time (a rough figure for the sake of argument).  Then, since only 1 in 2000 points on our source grid matches temporally with our destination grid, every other point has to calculated mathematically.  This is where bit depth comes in.

Instead of using what was actually measured, we have to calculate a value for the destination grid which corresponds to its location relative to the source grid.  If the destination grid sample point is exactly half way between two source grid sample points, then we would take the average of the bit depth between those points.  However, this is rarely the case.  Which number is taken (measurement) has to be weighed against how close the measurement is to the source grid sample point, so as to provide as accurate a picture as possible (a weighted average).

Obviously, this requires computation of a value that may or may not have actually existed in the original wave being measured.

In short, sampling the original wave at a multiple of the destination sample rate requires no interpolation, since the grids line up perfectly at the multiple.  All that is required is dropping the samples in between.  But converting between sample rates which require a number of cycles between each grid before they line up again requires computation to approximate the values that don't line up in the time domain.

So, stick to a sample rate multiple of your destination.  Or, record at your destination sample rate (if your release is CD quality, then record at 44.1 instead of 48).  You will end up with real measurements instead of estimations of what occurs between them.
Logged

thephatboi

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 271
Re: "Even" Dithering ?!?
« Reply #7 on: April 09, 2007, 04:13:20 pm »

Excellent post above.
BTW, different sample rate converters do sound different, and different dithers sound different too. I do a fair amount of mastering and have A/B tested a bunch of both: all sound slightly different so it DOES matter. Try a bunch and choose the correct flavor for the song (oh no, another fricken variable to totally affect the final outcome of the sound of your song...) and then of course what does it sound like in MP3 or MP4...?

The best way to deal with this issue is to record high (I'm a fan of 24 bit 88.2k) then  have an excellent mastering engineer master through analog gear, recording the final outcome to a second DAW in 16 bit 44.1k, thereby bypassing all computer SRC and dithers; let analog processing do it for you. This is how I do it, but this is  only beneficial if your converters aren't crap, otherwise you might as well stay digital.
Logged

johnR

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 923
Re: "Even" Dithering ?!?
« Reply #8 on: April 09, 2007, 05:51:22 pm »

Another factor is whether the final destination is CD only or you want to be able to release it on DVD at some point. When downsampling from say 96 kHz 24-bit to 44.1 kHz 16-bit the conversion artifacts are likely to be below the 16-bit noise floor (assuming your SRC is decent) and therefore undetectable. Going from 44.1 kHz 16-bit to 96 kHz 24-bit, SRC artifacts could become a problem.
Logged

imagineaudio

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 92
Re: "Even" Dithering ?!?
« Reply #9 on: April 13, 2007, 11:59:00 pm »

...or just record at 44.1....
Logged
"'Cause life's to short to sound like shit." -Smart guy from the music dept of San Jose State University.
Pages: [1]   Go Up