|
Home
/ FAQ
/ News Classifieds / Events |
Audio Asylum Thread Printer |
Get a view of an entire thread on one page |
124.189.36.153
| '); } else { document.writeln(''); } } else { document.writeln(''); } } else { document.writeln(''); } } // End --> |
In Reply to: RE: Please - skip dithering and provide 24/44.1 posted by soundchekk on July 02, 2009 at 01:25:21
Suggest you may want to re-download if you are getting decoding errors.
Hi Christine.
Dither: Ok you mentioned that you have applied POW dither. 24bit and dither?? This got me confused.
Decoding: Same Problem. Obviously FMAK has the same problem. Why don't you upload a .wav or .raw
Dithering is always a good idea when reducing sample depth.
The 24th bit is at -140 dB. Deep within the noise. Dithering isn't necessary.
"The 24th bit is at -140 dB. Deep within the noise. Dithering isn't necessary."
High end audio isn't necessary.
Are you suggesting that adding dither at 24 bits hurts the sound? When do you think dither should be used?
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
We found that going from 64 bits to 24 bits sounded better with rounding than dithering (+/-1 LSB TPDF). I was shocked that I could hear any difference at all, but there you have it. The improvement I heard from rounding instead of dithering did *not* want to make me spend a bunch of time trying different dithering algorithms.
Christine, if you still have the 64 bit master files, you might try it both ways and see what you think. Or even post them on the board and let other people see what they think.
Stereophile has an article on this same topic posted on their website. It was written by Keith Howard a few years ago. He reached the same conclusion, although he was going from 24 bits to 16 bits. That's what had us even try the experiment.
I have some ideas on dither algorithms you might try, if you decide to investigate them again for a new project. If you are interested send me an email. Perhaps at that time you will be looking at 32 bit chips and won't have any need for dithering.
My experience reducing hi-res material to 44/16 format for release is similar to the Stereophile article. To do the best job possible mastering, it is necessary to fiddle with both the resampling and the dither. You are trying to fit 10 pounds in a 5 pound bag and it is an artistic judgement what to keep. But at the high sampling rates and bit depth inside a DAC, perhaps one size fits all...
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
"Are you suggesting that adding dither at 24 bits hurts the sound?"
I stated it wasn't necessary. At best, it wouldn't impact the sound at all. At worst, if there is a rounding issue involved (unlikely), it could alter the signal. It could also raise noise levels under some circumstances. Because at these levels, the dither will be only acting on the ambient noise, not the signal.
My contention is, why use something if it isn't needed? In audio, if it isn't needed, it *might* (not necessarily will) degrade sound.
"When do you think dither should be used?"
I would guess truncating to 18 or fewer bits.
To get a better answer to this question, this is a case where DBT evaluation would be valuable. Determine, by blind listening evaluation, where the threshold of audibility is for dither, as a function of wordlength.
WinAmp plays the FLAC file just fine, but, DbPoweramp will not convert it to WAV. Some kind of error.
I think there must have been some problem editing the tags - I think the actual music data is fine.
I'll try and post a version without metadata shortly.
Post a Followup: