|
|
Post by ragan on Aug 24, 2017 13:38:51 GMT -6
I'm certainly not "anti-dithering" by any stretch. Nor am I a conversion tech wizard.
I'd just like to know if there is an answer to the question: "is this demonstrably real or is it placebo?". Maybe there isn't. If not, that's when I'd totally spend some time comparing for myself.
In this case, testing it out is so stinkin' easy I'll just try it anyway.
|
|
|
|
Post by thehightenor on Aug 24, 2017 14:31:57 GMT -6
Same here. When I insert a dither it's a "night and day difference"  When a friend double blind ABX tests me I couldn't pick it anymore than 5/10. .... and that special little "pro audio phenomena" applies to lots of other tests as well like plugin EQ's vs hardware EQ's sample rates, and fancy pants pre-amps. Of course I'm not saying other people can't hear these things or to diminish the value to others. I'm convinced that blind testing imposes a peculiar listening mode which causes many people to tend to second guess themselves. Related to that is my belief that the accuracy of ones selections in a blind test decreases with the number of trials. In other words I'm inclined to believe that your first choice is the most accurate and they decrease in accuracy the more additional times you go. I suppose that wouldn't apply to blind testing the pitch C or C# - I'd get that 100% 'till the cows come home. So then you'd be into C or C pitched up by 75 cents? 50 cents? 25 cents? 10 cents? 1 cent! We all have a cut of point with a test like that - I'm pretty sensitive to pitch, but there comes a point it's irrelevant as natural deviation on pressing a string or vocal pitching can be greater than any need to resolve pitch variation to 1 cent. I think I'm just trying to echo a point I read by Hugh Robjohns (the SOS ex BBC engineer) that if you're hotly debating the difference between two things that are separated by a very small variation that can be lost in a blind hearing test - probably for the very reason you stated - then maybe it's not worth worrying too much about if you can't hear it, after all there are so many large fish to fry on the path to a great song and production. On the other hand if one can't hear the difference between C and C# then I hear fishing is an excellent hobby |
|
|
|
Post by lcr on Aug 24, 2017 14:49:51 GMT -6
I think we have to consider the workflow and the amount of dithering were talking about. Im sure dithering 60-80 inserts effects things more than say 8 inserts. Im looking forward to my own test, I think I'll test the Pro L and something else I have (hmm.. what do I have?)
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 24, 2017 15:46:51 GMT -6
Sometimes I dither. Other times I just dilly - dally.
|
|
|
|
Post by johneppstein on Aug 24, 2017 16:21:51 GMT -6
^^^ This A-B-A-B-ing conundrum is absolutely true. At least for me it is. Especially if things are "close". And I think in dithering they are, although the benefits can be cumulative when you're rocking 60-80 hardware inserts like I often am. The answer is simple. Mathematically, it's the "right" thing to do. Two of the Bob's say it's the right thing to do. I can hear a "positive" difference when I instantiate the ProL dither in front of a hardware i/o. Why would I sit around all day A/B-ing things. It just works. So I do it when I'm not being lazy or working fast. One thing is for sure, it certainly isn't NEGATIVELY affecting my mixes. What's the big anti-dithering contingent's reasoning? It takes an extra 1/1000th of my CPU to dither all my outputs? You've got to "prove" it to me? "I just don't believe it"? Whatever. I say do whatever works for you. I didn't dither for years. Now I do. It's made a positive difference in my mixes. If you don't believe me, prove it wrong.  NOTE : BTW, as mentioned I'm using ProL's dither. I still think it's doing "something" with the dither. I hear magic dust. But that said, if anyone has a CRAPPY dither that doesn't cost me $$$ to test that they would like me to try, go ahead and hit me up with the crappy dither plugin. AAX please. Well, they're not mine, and they're not "crappy", but Chris J at Airwindows has a number of different dither plugins. I'd be curious to hear what you might think of them. |
|
|
|
Post by drbill on Aug 24, 2017 21:47:23 GMT -6
^^^ This A-B-A-B-ing conundrum is absolutely true. At least for me it is. Especially if things are "close". And I think in dithering they are, although the benefits can be cumulative when you're rocking 60-80 hardware inserts like I often am. The answer is simple. Mathematically, it's the "right" thing to do. Two of the Bob's say it's the right thing to do. I can hear a "positive" difference when I instantiate the ProL dither in front of a hardware i/o. Why would I sit around all day A/B-ing things. It just works. So I do it when I'm not being lazy or working fast. One thing is for sure, it certainly isn't NEGATIVELY affecting my mixes. What's the big anti-dithering contingent's reasoning? It takes an extra 1/1000th of my CPU to dither all my outputs? You've got to "prove" it to me? "I just don't believe it"? Whatever. I say do whatever works for you. I didn't dither for years. Now I do. It's made a positive difference in my mixes. If you don't believe me, prove it wrong. NOTE : BTW, as mentioned I'm using ProL's dither. I still think it's doing "something" with the dither. I hear magic dust. But that said, if anyone has a CRAPPY dither that doesn't cost me $$$ to test that they would like me to try, go ahead and hit me up with the crappy dither plugin. AAX please. Well, they're not mine, and they're not "crappy", but Chris J at Airwindows has a number of different dither plugins. I'd be curious to hear what you might think of them. AAX only please. |
|
|
|
Post by drbill on Aug 24, 2017 21:49:55 GMT -6
I'd just like to know if there is an answer to the question: "is this demonstrably real or is it placebo?". Maybe there isn't. Evidently not. For you it seems. If both Bob's say it's essential, and if various people on this thread say it's good, and you still doubt,,,,,,then...... It must be GOOD -- Or, as you say, you'll have to do the test yourself to prove it's worthlessness. If it's a placebo, I'd like to upgrade all my gear with the placebo option so that I can be happier sitting at my mixing desk! Let me put it a bit more pragmatically -- on certain material (ac guitars seem to love ProL dither) it's a bigger difference to me than the difference between a cheap Neve style transformer and a Carnhill. That good enough to try? |
|
|
|
Post by wiz on Aug 24, 2017 21:51:40 GMT -6
I use the Airwindows Plug In... Bob put me onto that one
cheers
Wiz
|
|
|
|
Post by ragan on Aug 24, 2017 22:35:39 GMT -6
I'd just like to know if there is an answer to the question: "is this demonstrably real or is it placebo?". Maybe there isn't. Evidently not. For you it seems. If both Bob's say it's essential, and if various people on this thread say it's good, and you still doubt,,,,,,then...... It must be GOOD -- Or, as you say, you'll have to do the test yourself to prove it's worthlessness. If it's a placebo, I'd like to upgrade all my gear with the placebo option so that I can be happier sitting at my mixing desk! Let me put it a bit more pragmatically -- on certain material (ac guitars seem to love ProL dither) it's a bigger difference to me than the difference between a cheap Neve style transformer and a Carnhill. That good enough to try? I can't quite figure out your angle in this response. I simply don't take things as gospel truth based on reputational authority alone if there seems to me to be a reasonable chance the 'authority' has it wrong. I have no idea if any of these Legendary Sacrosanct Bobs have it wrong. They probably don't. But since the mathematic/programming side of this seems up for debate, I'm interested in those concrete answers, if they exist, to consider alongside the anecdotal testimony. I can't imagine why this would ruffle anyone's feathers. Then again, people are very, very weird. Myself included. Edit: PS - I have no dog in this fight. If there's something as easy as plopping a dither plug before I/O that makes things sound better, I'm all for it. It's usually a lot more work to improve sonics. |
|
|
|
Post by Bob Olhsson on Aug 24, 2017 22:36:20 GMT -6
If dither is properly programmed, it should be less audible than the distortion it prevents. It is not masking anything while the distortion will mask low level detail such as depth and imaging cues.
Outputs to converters or fixed point files are what needs to be dithered. Virtually all A to D converters output 24 bits so the only time you'd want to dither an input would be to print a lower bit depth file which is pointless.
|
|
|
|
Post by drbill on Aug 24, 2017 22:54:53 GMT -6
Evidently not. For you it seems. If both Bob's say it's essential, and if various people on this thread say it's good, and you still doubt,,,,,,then...... It must be GOOD -- Or, as you say, you'll have to do the test yourself to prove it's worthlessness. If it's a placebo, I'd like to upgrade all my gear with the placebo option so that I can be happier sitting at my mixing desk! Let me put it a bit more pragmatically -- on certain material (ac guitars seem to love ProL dither) it's a bigger difference to me than the difference between a cheap Neve style transformer and a Carnhill. That good enough to try? I can't quite figure out your angle in this response. I simply don't take things as gospel truth based on reputational authority alone if there seems to me to be a reasonable chance the 'authority' has it wrong. I have no idea if any of these Legendary Sacrosanct Bobs have it wrong. They probably don't. But since the mathematic/programming side of this seems up for debate, I'm interested in those concrete answers, if they exist, to consider alongside the anecdotal testimony. I can't imagine why this would ruffle anyone's feathers. Then again, people are very, very weird. Myself included. Edit: PS - I have no dog in this fight. If there's something as easy as plopping a dither plug before I/O that makes things sound better, I'm all for it. It's usually a lot more work to improve sonics. It's all just discussion man. Try it for yourself to see if it's worth it. I'm happier doing it than not. I personally don't care if it's a placebo or not. I'm happy using it. |
|
|
|
Post by ragan on Aug 24, 2017 22:59:14 GMT -6
I can't quite figure out your angle in this response. I simply don't take things as gospel truth based on reputational authority alone if there seems to me to be a reasonable chance the 'authority' has it wrong. I have no idea if any of these Legendary Sacrosanct Bobs have it wrong. They probably don't. But since the mathematic/programming side of this seems up for debate, I'm interested in those concrete answers, if they exist, to consider alongside the anecdotal testimony. I can't imagine why this would ruffle anyone's feathers. Then again, people are very, very weird. Myself included. Edit: PS - I have no dog in this fight. If there's something as easy as plopping a dither plug before I/O that makes things sound better, I'm all for it. It's usually a lot more work to improve sonics. It's all just discussion man. Try it for yourself to see if it's worth it. I'm happier doing it than not. I personally don't care if it's a placebo or not. I'm happy using it. Maybe I was misinterpreting your last post to have more attitude than was intended. I couldn't quite make out what you were trying to get across. Anyway, just tried ProL dither on some acoustics (patching in the Harrison EQs) and I heard no difference. Tried hard to, but I got nothing. I'll try more though of course. |
|
|
|
Post by drbill on Aug 24, 2017 23:04:38 GMT -6
? ? ?
I can't comment on what you did or didn't hear. For me, it's fairly obvious 60+% of the time. Again, Ears, Monitors, DAW, Converters, Program material, & Room all come into play big time when making these subtle comparisons.
The time I first used it I was on a transient heavy acoustic project with about 10 delicately picked acoustic guitars, mando, dulcimer, harp, and other stringed instruments. Going out into comps was anything but subtle in that situation. Other things, like a dense heavy tracks, it's not nearly as "obvious". At least for me.
Draw your own conclusions.
|
|
|
|
Post by ragan on Aug 24, 2017 23:19:41 GMT -6
? ? ? I can't comment on what you did or didn't hear. For me, it's fairly obvious 60+% of the time. Again, Ears, Monitors, DAW, Converters, Program material, & Room all come into play big time when making these subtle comparisons. The time I first used it I was on a transient heavy acoustic project with about 10 delicately picked acoustic guitars, mando, dulcimer, harp, and other stringed instruments. Going out into comps was anything but subtle in that situation. Other things, like a dense heavy tracks, it's not nearly as "obvious". At least for me. Draw your own conclusions. Well we can take several of those variables out of the equation easily. Uploading files now. Maybe it's actually there and I just can't hear it. If you guys can, and it's an improvement, I'd still start doing it. |
|
|
|
Post by ragan on Aug 24, 2017 23:25:53 GMT -6
|
|
|
|
Post by ChaseUTB on Aug 25, 2017 1:42:09 GMT -6
Thanks for the files. I posted my results from testing myself and got passively aggressively shit on and told my ear suck and so does my room and monitoring and DA ...
So to follow along with my shitty ears I will use my shitty earbuds and say that I don't hear this night and day difference at all.
Waiting for the ppl to chime in and say yeah you idiot you need 24 inserts to hear the difference and you need this brand of DAC and bla bla bla
Then if you ask them to pick the difference they can't or won't ...
I would literally be guessing which file has the dither and that is consistent with my testing...
going to 16 bit was different ...
|
|
|
|
Post by Quint on Aug 25, 2017 7:13:45 GMT -6
If dither is properly programmed, it should be less audible than the distortion it prevents. It is not masking anything while the distortion will mask low level detail such as depth and imaging cues. Outputs to converters or fixed point files are what needs to be dithered. Virtually all A to D converters output 24 bits so the only time you'd want to dither an input would be to print a lower bit depth file which is pointless. Does it hurt to dither even if you didn't have to? Is there a downside to adding dither when it wasn't needed? |
|
|
|
Post by Bob Olhsson on Aug 25, 2017 7:28:03 GMT -6
The only time dither isn't needed is when the numbers are being copied unchanged. Even if the distortion is inaudible at one point, subsequent processing can make it audible. My authority is Jim Johnston of Bell Labs where digital audio and signal processing was invented. In his words: "If you don't dither, the math is wrong!" The idea of it being optional is a myth. It is lots less noticeable in floating point handoffs because they are the equivalent of turning everything up to full scale and then back down again but dither is still required for conversion to fixed point. My experience has been that if the audio has ever been truncated, the effect of dither becomes less noticeable. I assume this is because the distortion is masking it.
|
|
|
|
Post by ragan on Aug 25, 2017 8:36:17 GMT -6
Thanks for the files. I posted my results from testing myself and got passively aggressively shit on and told my ear suck and so does my room and monitoring and DA ... So to follow along with my shitty ears I will use my shitty earbuds and say that I don't hear this night and day difference at all. Waiting for the ppl to chime in and say yeah you idiot you need 24 inserts to hear the difference and you need this brand of DAC and bla bla bla Then if you ask them to pick the difference they can't or won't ... I would literally be guessing which file has the dither and that is consistent with my testing... going to 16 bit was different ... Well I was literally guessing when I blind tested myself in my studio. I simply don't hear a difference. I have zero interest in proving myself right though, I'm just genuinely curious if there's something I just don't hear. Bill said it was particularly noticeable to him on acoustic guitar so that's what I tried first. Like I said before, I'll be happy if it turns out to be real, this audible improvement, because it's so easy. But in this one test, I don't hear anything. |
|
|
|
Post by Bob Olhsson on Aug 25, 2017 9:06:22 GMT -6
It's most noticeable when the sources are all microphones picking up room tone with the musicians.
|
|
|
|
Post by ragan on Aug 25, 2017 9:18:58 GMT -6
It's most noticeable when the sources are all microphones picking up room tone with the musicians. I don't have that to try at the moment. Curious, do you hear a difference in those acoustic guitar examples I posted last night? Would you expect to in that context? |
|
|
|
Post by drbill on Aug 25, 2017 9:45:51 GMT -6
I think I'll bow out of this now before it gets really ugly. But just to clarify -- I'm not here to prove anything, my only thoughts were to share my personal experience in my studio with my gear and my ears. Which of course are different than all of yours. ChaseUTB - I certainly hope you weren't referring to my comments about room, D/A, ears, program material etc. - which of course are all different - no putdown or offense meant in anything I said. Anyway, all, do what sounds best to each of you. ragan - sounds like you got your answer. <thumbsup> Cheers, bp |
|
|
|
Post by jcoutu1 on Aug 25, 2017 10:05:35 GMT -6
I slapped Pro-L across a mix of about 60 tracks last night and starting running into my CPU wall. Couldn't really bypass on and off the way I needed to, to form any opinions. I'll have to try again with a lighter session.
|
|
|
|
Post by Bob Olhsson on Aug 25, 2017 10:11:01 GMT -6
Where I hear a difference is taking it off! JJ told me that it would require as much as 10 dB. of extra noise to mask truncation distortion. You can actually hear detail below the truncation level in the dither. Daniel Weiss posted a demo of that at one point.
|
|
|
|
Post by ragan on Aug 25, 2017 10:38:12 GMT -6
I think I'll bow out of this now before it gets really ugly. But just to clarify -- I'm not here to prove anything, my only thoughts were to share my personal experience in my studio with my gear and my ears. Which of course are different than all of yours. ChaseUTB - I certainly hope you weren't referring to my comments about room, D/A, ears, program material etc. - which of course are all different. Anyway, all, do what sounds best to each of you. ragan - sounds like you got your answer. <thumbsup> Cheers, bp  Ugly? You lost me. I was just trying what you suggested and said I couldn't hear it and you expressed a lot of surprise at that so I uploaded the samples for you to hear. I'm just interested in what the truth of the matter is. Maybe that truth is that you can't hear it in this sample but can in others. I have no idea. It is always telling to me though when opinions seem to evaporate once there is actual audio present. |
|
