Pantalones wrote: The scale is really great, super easy and intuitive, but a little laggy for fast notes. Is that scale v1 or v2 or both you are referring to?I'm looking to buy my first dedicated quantize (I got a Listing now) and I need it to be able to quantize 1/32 notes, in 120-150 bpm.
Is this accessible and editable to everyone? Figured it'd be a handy thing to have when this conversation comes up, but I've got no experience with much outside of the Listing and A156, please do stick info in if you know it. In that case you will want to test the quantize rs listed for accuracy and delay and not assume they all work well.
Eldradthemad Common Wiggler Posts: 183 Joined: Sun Mar 08, 2015 8:05 pm Location: United States Postby eldradthemad Thu Jul 30, 2015 2:54 pm Cheap option I got was a Blue Lantern dual Barton clone.
Eldradthemad Common Wiggler Posts: 183 Joined: Sun Mar 08, 2015 8:05 pm Location: United States Postby eldradthemad Fri Jul 31, 2015 7:01 pm I know BL is hated by some here but his cheap Barton style quaint works great for me.
While the BL is kind of funky, not great build quality, and not terribly precise, I find myself using it quite often. www.soundcloud.com/volodwww.youtube.com/user/volod3mir/videos dope fiend wrote: The first step to recovery is: “We accepted that we were powerless over synthesizers, that our budgets had become unmanageable”.
Mdoudoroff Super Deluxe Wiggler Posts: 4127 Joined: Fri Jul 17, 2015 1:10 pm Location: New York City Postby mdoudoroff Tue Sep 24, 2019 2:24 pm Continuum and Timmy : I totally understand that these are all important considerations for designing a high quality product.
So, presuming I am willing to take a stab at #3, what would be an acceptable, concise standardized format for summarizing performance in ONE (1) new column, that I could document for readers in a short paragraph or two above the comparison? Take the voltage range and divide by the number of digital values.
This would be the best case value since the analog stages can alter this performance a great deal, but it would be equal best case for all devices. I guess you could convert to cents from that too with the caveat that quantize rs only provide voltages and pitch tracking and tuning is a separate issue. For example an 8 bit DAC with a 0-10V output would have accuracy of 39mV (10 / 256) which is almost half a semitone.
Pld Super Deluxe Wiggler Posts: 1034 Joined: Thu Mar 05, 2015 10:15 am Location: Germany Postby PLD Tue Sep 24, 2019 3:49 pm Yeah, I can see the difficulty in providing a useful metric because it's easy to get into the weeds, but I'd agree that some measure would be useful.
Mdoudoroff wrote: If, for example there are any 8-bit ADC's/Days in any of these products, I can see how that would be worth calling out. But it is designed so that on paper the error should be < ~20mV at each semitone just from the DAC specs.
Mdoudoroff Super Deluxe Wiggler Posts: 4127 Joined: Fri Jul 17, 2015 1:10 pm Location: New York City Postby mdoudoroff Tue Sep 24, 2019 4:24 pm Listing the bit depth is probably fine with a note about the accuracy with a couple of examples.
I'm really interested in Intro Harmonies but the smaller ADDAC207 seems tempting as well. Do you guys know if ADDAC207 has the Harmonies ability of setting diatonic notes in a pre-programmed scale according to the root note? I now it's not a dedicated quantizer, but the algorithms are quite advanced, like having the ability to load and parse Scala files.
Mdoudoroff Super Deluxe Wiggler Posts: 4127 Joined: Fri Jul 17, 2015 1:10 pm Location: New York City Postby mdoudoroff Wed Sep 25, 2019 11:53 am Continuum wrote: Listing the bit depth is probably fine with a note about the accuracy with a couple of examples.
A bit depth is a very familiar term in audio and people know how to compare it. Mine were so bad that I was going to replace them with 12-bit Days and write a calibration UI.
Then there is the fact that voltage only very roughly correlates to actual pitch, but that is a much larger issue in Euro. Hm, a couple of things- without testing, the DAC res/ nominal accuracy column is super misleading and isn't worth much.
I haven't read all the comments so bear with me. First- There are a number of ways one might improve output accuracy with limited hardware, ***especially** in such a limited application of a DAC as a 12-tone, equal temper pitch quantize, that might only need to span 5-6 octaves. Well it's not as good as some others, but it's well within the acceptable, at least for me. I think it's unfair to give the impression (or am I misreading it) that is +/- 46.8 cents, which obviously would render it completely useless.
Mdoudoroff Super Deluxe Wiggler Posts: 4127 Joined: Fri Jul 17, 2015 1:10 pm Location: New York City Postby mdoudoroff Fri Sep 27, 2019 12:42 pm Toppobrillo wrote: Hm, a couple of things- without testing, the DAC res/ nominal accuracy column is super misleading and isn't worth much.
Well it's not as good as some others, but it's well within the acceptable, at least for me. I think it's unfair to give the impression (or am I misreading it) that is +/- 46.8 cents, which obviously would render it completely useless. The 12 bit has the resolution to hit everything within a MV or so which could be within a cent in an actual product once the analog stages have proper adjustment for offset and scaling.
Between offset and drift the analog stage can move things several MV if the components are not precise and don't have any adjustments. For example, the Quantize has some pricey parts and full adjustment for offset and scaling, but in practice that only gets the error to about 1mV over a 5 volt range which is many times the ideal DAC resolution.
Seems that assumes using all 256 codes, but could/should be design-specific and be optimized as @toppobrillo suggests. For Penrose, the output scaling is such that 1LSB = 1/2 semitone, so only using codes 0-240 for 0-10V and an ideal DAC would produce exact values for the whole range. Which is why ate gets reasonable results but apparently I had horrible bad luck.
That's where the range/resolution value becomes not-quite-so-useful, especially at the low end, but also hard to quantify better without analyzing the hardware.