how to stop multiparameter updating two values in different increments

[babbid]

I setup one Multiparameter on MLimiter to adjust Threshold and Ceiling at the same time and set range mode to "interval" and set value to 0dB and Max to -24dB (on both) and when I move my multiparameter to 0 or -24 then both Threshold and Ceiling are updated correctly but anything in between the values are different. For example setting my multiparameter to -20dB the result is that ceiling is -20dB but Threshold is -17dB. Why is that? As far as I can tell Im doing everything correct.
Thanks,
-David
PS: I have no transformation shape in place

Code: Select all

$eNqNVdty2jAQfecrNJvHJvEFm8uM7Q7Q0GYGGganTfqo2gvWREiMJJKQr+9Ixi7Qdho-ycd70zm76+Tj64aTZ1SaSZFCcO0DQVHIkol1CjuzuhrAx6yTzB8fHzUaw8Rak4sUAiATyTndaizHVJQ3JTNSHXBhlOQclb4R9CfHMgUfyERJreUzqrksMYUwKseeLMzRh-v9FlMYCcrlGsgntX9AM1qtsDD6Vmx35jNlwmX4-O02Z28YphB04-hyEA5a7IbqfZhCP4gu+70ekBkTTy79bJlvOTMGlTULUgj9XngOhynEAZA5fZ1N78ZcFk8WTaEbApmPTyq625njkuY7btiCKrpBg0q33Ljc3xt6+4Nh7EeQJe2d7W33b6hIvpHSVAK1TsG-9mMgD0yU8qUm5QsVArJkoXDFuEHFxDpLFpIJQ6acrvU4hSAKgLxaCbt+4IfD4TDu9cD7l1l0buadBVeyQK2l8kmrYgBkyjh3h7ySLxPJpfpO+Q61w+4rVji2vSP-gNw9o1KsRGf+b8PwvYbd9xpG-zH0-pAh6yQ-qSj9Q4v-tRFy5FgYLBd8t7bq+5ehlca96eZApjvOv9INpjBjG2ZQkQDICQDk0XXHjxR+B8hPZ+yiQMbdKF4Fw+uhe+LQVnFhKoW6krwMbMN0+-Fw6MexP+j2gYyMocVTPWhTqs3RsP5u1Nn0bkEFchu7HpRJRYVAruuBkfJpVCEtPyGnewf9lY4lcqQaj5PllvGKMrFoPFskx0KKkqr9IVejRH379qCzxHNC2NVzOlrZGeDb0a-XT13DV6k2lNcbgW5wKV-q4k8+1krcOw4-TJBxIPkWm3m1LW1ZHXTdE3SjXnyA6zDjPVkxpQ3ZNmVYBZuzE39JxRpTiBxtTUQgrVFdgd2dJCBXpO0SckUmtehtIYHr6v9Hvw6jbi+MBsPQ74fvS1UzIHl5msw74-gPRGedZFQY9owLhRoNsQpY5yyZUINrqfa3Bje3YiW9LJnslEJhDqy7BMfONpbakiU19scAnhX9oPUDK001z5Jc7lSB5HhHNa3TmHgHn07iHf2tdNb5BUEXNmQ=

[DarkStar]

That is because the ranges of the target parameters are different:
-- Ceiling is from -24 to 0 dB
-- Threshold is from -80 to 0 dB
All values are sorted internally as decimals, in the range 0 ... 1.

I get pretty good correspondence using an 8 bank mode:


More banks will give you better correspondence. And, incidentally, here you can see that the corresponding values are at different places in the range.

Code: Select all

$eNqVVF1v4jAQfOdXWNt76wFJCCGVklQ0QFUJ2hQo5R5dshQXY1Pb4aO--kQSuB49Vbp98+xkd8a7cXC9W3GyQaWZFCHYNQsIiplMmXgNITPzqg-XUSUY9NmKGVRTjcYw8arJRQg2kIsZMp5zq17NbzhALsxCoV5IntohWDWv5Xue7bmu27CBtI2hs+VAphhCj2oDJJbCKMk5Kt0V9IVjGoIFpJsyI9WIfWAItmNZP1238RntUr0Pwba9RpHpM7GMF1QI5Dov0Jdy2V4gTTvI6b6AhqM1Z8ZgXsEOwbE85xx2QmjaQAZ01+893HA5WxYiDtYGGTcsoYqu0KDSseScrnWpeIgcqcbC24jLLZARSzFeUCaSo4QTMsKZFClV+1J0np0cx+A37SvPhyg463gOWOSGimUsM2FC8IHcPt0VV1SouJdqRTmQ4nSgaiD3dIUhjPMhXcbIOJDRGksTE8ozPIyt1fTKU-nxnsyZ0oasj70hCk46SHdnhqjzjehlnA+peMUQ3PwejzWBnPiFhrhYnVNXG6If7VPcTIqY9pJtid23qyU4scd+CXZeE7sE8an6Xob3dspvk20YBvVT8--VXXPchue4-pVjtZwvJsbHbf-LxvSt4ZrdcvPx4U-3LWl5ejNOnGXz1uf+yPyq2vHHy7Aqn53x5VMy3T82dWeebO-fJ9mk080eq5e4azwsmtQbf2OjfrYMXxAdVYL2zLANJgo1GhJnSqEwn7E7MZchwL9ShcHirzvYqh-egbyuEwXPLDWLQRSMZKZmSBKFc8YNqvwtsKAeBfUjpf-0h1ds00CuUBiq9jkvz8cKtfm+3pFSP4qoBPXzd0lHld-lWJHW

[DarkStar]

And here's my follow-up question.

Let's say we want to use a Transformation Shape for the Threshold for the above case. So that when the Multiparameter is set to -X dB, then both Ceiling and Threshold are set to -X dB? What is the easiest way to create that transformation curve?

[babbid]

Thanks for the answer DarkStar and sorry for responding that late! Life got in the way.

I understand what you are saying and possible workaround using Banks. But ranges you are referring to are original full ranges which I overridden in my params definition by saying that both should be
-- Ceiling is from -24 to 0 dB
-- Threshold is from -24 to 0 dB
in which case they should increase in the same increments no? Otherwise you cant rely on multiparam values no? I mean if I set value of multiparam to 50% and I know that I limited threshold range to -24 to 0 dB (and NOT its original full range of -80 to 0 dB) then I would expect that 50% means -12 dB. But right now it is -8.15 dB.

[DarkStar]

That's why I went for the Banks approach.

If the parameter ranges are set to be the same (-24 dB to 0 dB in this example) but their full ranges are different, then the MP value (which is in the range 0 to 1) cannot set the same value in both targets. If we knew the transformation curves used for each (full) range, then we could create the appropriate Transformation Shape graph(s).

[DarkStar]

Give these two a try; they both use a Transformation Shape graph for Threshold. One covers the range -24 to 0; the other is for -80 to 0.

Code: Select all

$eNq9WFtz4jYUfs+v0Cg7nXZaEku+TzGdBEJKC7tMWPbyqNgnoMVIrCwnIb++4xs3E+-utFRPcPRZ5zvf0dGR3f7jeRmjR1AJlyLA5MLACEQoIy5mAU71Q8vDf3TO2qMhX3IN6lMCWnMxS9B5gAlG5yHwOMe2qJX9nTPFQg0qwN0YmMDofMa4CHCLGBidr4At5jKOAmxcGIZhY3SezNkKljKCAA-ljCmu50seYnSu5wqSDEwyNLWpaduO7RHfw+hKaxYuRvlTfZZojLoyjtkqgeiOz+Z6zATEOcGbiGupJvwFAuz45DfLMneNNyxZB5iQambIxaI7Z0JAnATYwGgo5eJqDizqQczWAaYUo+HdZBVzrSFfggSYGk7NTANsE4xG7HnYf3cdy3BRkDApRqM01nzMFFuCBpVsuOce7yAGlkAR3Agini4xmvAIunPGxThnYexYJhBKETG1Lnnnsx+qjHo28R0Pd9oHPg8NBrpmYtGVqdABpjj-95YtwQiwgaLrrYEE2CwsPXhgaaw-sDiF3OmtkukqwLexvGdxgtHtdFAoXcRyzWe5bSjDBbuPIc9POcXEIsEoWz-APcUfAaPJCkpFcg8lcI3umVggwZaAuNCgVjJmGiLcaW9iQTfP+g6SfP1+Gsd3TMwgwFaejYruhU+cargYbR4uOAzEKtUo27oYlY8bF67hmdUoWeVbc7uSiztvvpbD+OJ-KMYnPX4KgvblxkenfXmg-mE6yG46zL10lHW1k49xKsL5emuiAf6TqeibCZlADKGWCv18z2e-HM1Ft6rnWj6aMvPjyTCJQ6nrOrbv1ZNxFT0yEUKEWmgMbIGyQ2EnLYZBdrJhEnu7UudNz4gHn+cjn-S88WQ65zcv5nT8dLUZe1n5DtKvU8tk2GFl486byonX88pfvacxnfldDV-Ct0brR53vKkaoXWMzYTpVTHMp6nxy-CZzW25XVz3915ebgbp++fs0wuzsoSP+t9rsj3-pf0+K91UrOcpgbwy+Kufdyl-9+rnVVLDksGDp4XF44bvVuUC9b5+NqdZSvHY8vpVqyeJNTRYtt1aRmzp84CrRaFVR26vGRtE2K29UuqxFTg8j915tBJPyrlCLtorne6ItWv0Jgt3ZoZWPTbFYPnVt06Em8bxGNbxDNfz-To3b6WDClnAnn14Vp7wrnFadjZMfl8ev9TXjtLtlW-Yn0GTvTOmp9eVH0I3hk3pfJ7VzYr-1nXCzZFfZFsvusqfdL8d6tGW5lJi2b9vNgtXOVWIe2TDTQVfGUgXYM-r9fq-ff1WmYzp00U9sufodvf+-TtDvfHanS02yl6IbkWU3yvO5bVvtfA69ewSleARTwXXyKXuFwWgi2Oq9vFU8ygwZacmFRv2YzZLrABPLwGid31gtyzNMx7eJnfOrw57zK0MF94hrZ+8RptGIpyXepKZJbce0iWE14c0K71rENH3LJz5twlsl3jKJ5xqO47uW04S3K7zv2Kbt+o7ruE14p8TbLnE82yeUNtJ3S3jxUuoQyyR+E94r8ZWWlteopl-Cq1Bt3-Zexxe5InlN5Tuk03h9MWumpHPWvgo1f4SxggQ06qZKgdC7toF4kAHGx6aKLTxiz3zJX7KrXlaNBaGz9ih3QDvtjzzS81GnPZGpCgGNFTzwWIPKvyAUB0IFGU63uPJdWC5BaKbWOS6f7ypIdPN6FeSyInHWvjz8mpF0zv4BWYxb+A==

Code: Select all

$eNq1WG1z4jYQ-p5foVFuOu20BEl+n2JuCISUFu6YcNzLR8VWQIctcbKchPz6jm1sXkx8d9OiT8l6H+3us9pdic7b5zgCj0wlXAof4isEAROBDLlY+DDVDy0Xvu1edCZjHnPN1OeEac3FIgGXPsQQXAaMR7lui+ArC5mZaEkVDTRTPuxHjAoILheUCx+2MILgcs3oaimj0IfoCiFkQXCZLOmaxTJkPhzLBVVcL2MeQHCpl4olmTLOtIlHTMMmDibYhqCnNQ1Wkxw1pImGoC+jiK4TFt7xxVJPqWBR7uRNyLVUM-7CfGh7+A-TNPaFNzTZ+BDj8suYi1V-SYVgUeJDBMFYylVvyWg4YBHd+JAQCMZ3s3XEtWb5FtiHBNk1MfGhhSGY0Ofx8P11JINV4YRBIJikkeZTqmjMNFNJ5Xtu8Y5FjCasCG7CQp7GEMx4yPpLysU09wLtSWYskCKkarP1O--6scyqa2HPdmG3c2TzWIDANRWrvkyF9iGB+X-vaMyQDxEIr3cC7EOjkAzYA00j-ZFGKcuN3iqZrn14G8l7GiUQ3M5HBdNFLNd8kcvGMljR+4jl+dl+omKVQJDt78OB4o8MgtmabRnJLWwVN+CeihUQNGaAC83UWkZUsxB2O1Us4OZZ37Ek33+YRtEdFQvmQzPPRunulYftcjkQVODCh5FYpxpkRxeCLRxdOcg1yrX1qjiaeLdR98237UJfvY-F+qynT77faVc2up32EfvH6cD76TAO0rGtq718TFMRLDc7EfHhX1SF303IjEUs0FKBX+-54reTueiX9VzLR1Nmfj4ZBrYJcRzb8tx6MnrhIxUBC0ELTBldgawp7KUFIbyXDQNbu526bwYoGn1ZTjw8cKez+ZLfvBjz6VOvWgdZ+QGnX3cto2HPKwt235RG3IG7-WvwNCULr6-Z1+Adav2s8X3GMLFq3syoThXVXIq6P7l+lbmdb73eQP-99Wakrl-+OQ8xe2fohP0dN4frP9o-oOJDOUpOenCwRt+U-X7trX--0moqWHxcsOS4HV55TtkXiPv93phqLcVr7fGdVDGNqposxm6tIqs6fOAq0WBdunZQjY2kVTtXLLVrkZPjyN1XB8Fse1+oRVvG8yPRFqP+DMHundDSRlUspkccy7CJgV23kQ33mA3v-2Pjdj6a0ZjdyadXydneFc7LTmXk5+nxanMNnfe07Mr+DJwc9JSB2rQ-Md0YPq7PdVzrE4ej74yHJbvKtmh2lz3veTk1o03TIdiwPMtqJqzWV7Fx4sDMR30ZSeVDFw2Hw8Fw+CpNp3jog19ovP4TfKjRkM-J7B1zzn76g9i9mTXLnkg3Ist1mGd3d-XM3kSm6xHkENjt5Irg-SNTiodsLrhOPmevG3go+1LIZoKuP8hbxcNMKYtKcqHBMKKL5NqH2CzcrQnBc3Wf2OT3L9smyHY927Kw0wQhFQRjz7ANw7MyUQPC2CFcDzmWbSBikSaEWSGIbbvIdgixXa8JYVUIw7KQSRC2XdQYh10hTMvxXAN7BnEaEU6FsMogMMJNCLdCZE9e13Qdw2ukyqsArm1Zpl2AXkfgXDv3oZ2fm27jfceoiZLuRacXaP7IpoolTIN+qhQTel82Eg-Sh-DUp+KUT+gzj-lLdjfMyrdw6KIzyQ2QbucTD-Vy0u3MZKoCBqaKPfBIM5X-7FB0kFJlPN-pbR-PMmZCU7XJ9fLvfcUS3bxfqdIunbjotI9-Akm6F-8CtBJpsg==

I worked them out by trial and error heuristics using the Expression Evaluator in the Transformation Shape graph, then nudging the points to get a closer fit. But it was all a bit painful.

[babbid]

First, thanks again for your help! I can see how transformation shape can help here but as you said it looks like quite an effort. To be honest I solved my need by simply using a different limiter.

But generally speaking, wouldn't you agree that this is a bug in the product? If you simplify my original usecase and change multiparam to update only single parameter, the threshold, and limit the range to -24 to 0 dB and set Value Mode to "Percent" then as I said setting multiparam value to 50% leads to unexpected threshold db level of -8.15dB:

See M5 param in:

Code: Select all

$eNq1V9ty4jgQfc9XqJR52K0dQJbvVZgtAslsdmHGBclcHhW7gzXYskeWk5Cv37KxucNkapd+glZb3X1Oq9Xq-vmSxOgJZM5T4WGtTTACEaQhFzMPF+qx5eA-exfd8YgnXIH8moNSXMxydOlhDaPLAHhc2bZI2yS1uNQplyImWaBAengQAxMYXc4YFx5uaQSjywzYPErj0MOkTQgxMbrMI5ZBkobg4VE6Y5KrKOEBRpcqkpCXxlppbVHTdA2daqZtYtRXigXzcfXVDcsVRoM0jlmWQzguYsV9JlkyuvnkMwFx6XzDYMJnkaoXCEYf7m+n-BWohzXNdt4bjrbSXbN8QT1s6857w7IxGnExH0RMCIjz6ttRms77EbBwCDFbeJgaGI0m0yzmSoEst9A8bLn6rrZ0RiwLo3WwoEDmqxir3ScQA8thmeUYQl4kGE15CIOIceFXHsmGZgpBKkImF3WM1ernhmbbcU1i4F73jidQMTtVLMnQI5e58rClW45tWQ61LYxiVqtcauuUGgbulPWwHWxvR0HQFRPzQVoI5WGKq38fWQLEwwSFV2uF5mF9qRnCIyti9ZnFBSzJkGmRefhDnD6wOK+IuA65SuUShCs+q3SjNJizhxiqcqyXmJjnGJX7e3go+RNgNM2ghrLyUBsu0AMTcyRYAogLBTJLY6YgxL3uKhd0-aImkFf73xRxPGFiBh42MBqzlybctqtZjdgYrT5exnArskKhsvgxqj8nbZs4eiN1VKWarneyce-dj1rId-fzUr4q-9nzup2Vj163s4P+Lh3aJh36Fh31ydzgwy9EEC3WKurhv5gMf0rIFGIIVCrRbw989vtBLgZNR9jj4xQzv06GrlmU2rZlus4+Gf3wiYkAQtRCPrA5KtvKBi2EaBts6Jq53qn3bkji22-R2NWGjj+9j-j1q37vP-dXssXKG4I+HloJw0ZUJu69a5w4Q6f+NXz26cwdKPgefCStX3W+iZhGzb1opkwVkimeiv14KvsVc+vY+v2h+vv79a28ev3nPMBs1NAB-2tstuU-+t+C4q65jI4gsCG3P6T1KXOzP761Th1YbffA0t122Hbtpi+Ud+vPemOhVCqOtcePqUxYvDqTywv8eIesLgWUNaFtncaToK12XqHU2cuc7mZu7GWuO65LialTm+h7mTa5HMpsvKipOtRt2uZWnj7IAITK357dgSpoU0O3qOG4lNj0QLLGbrLO0VtvWo9ZRxN+C7XLyegMzG4cx8bHqjMYLrVN3aK65jgnqXd20XD-PzTKoY0lMEmfj4JTT1TnRWfl5NfhcfcucXLealn3uDNgstVAh3LR+QLqZPra-hCj7beGrXv+jMVSDvctVk73562XQwOJYdhU003XNE8Dpu2p8t5Ftx8o-gS+hBwUGhRSglCbulvxmHoYH1qquyh74Ql-Le-b8g00DSSAGO8+UTSMSqjLH9XzoH4VfOGhisa97jQtZADIl-DIYwWyejEus2hMRvdru3qgjFKp7kAmlV21PpCQq9P7NSYfGBcSwiIoi+6ufPKl0sM3lZDK9FOhskLFaREKyPM3mMiSkhN2-mhyYjUoAzu43qnhuuh2dt-Zee-iX05MCMg=

Don't get me wrong, I love Melda products, but this behaviour of multiparam breaks basic logic. And consequently leads to unpredictable results. Which sometimes is OK, but if you know what you want this is confusing.
-David

[MeldaProduction]

babbid: What makes you think this is a bug? MPs are controlling are parameters equally. You are trying to control parameters, which have a completely different meaning and structure, so... Taking it just a bit further - attach an MP to frequency of some kind and a gain, both full range. Move it to the center, the frequency became say 1000Hz, what should the gain be? Well, the value in middle of course.

[babbid]

re. "the value in middle of course" - thanks Vojta, now I got it! There are knobs which change value logarithmically(?) (eg Threshold) and there are knobs which change value linearly (eg Ceiling)(in MLimiterX). So "middle" for Ceiling is exactly 50% of its full range - 12dB. And I was arguing that Threshold should do the same but I overlooked the fact that Threshold operates in logarithmic scale and hence "middle" value is NOT half of its range, it is not 40dB but 11.9dB. And that also clarifies for me that having one multiparam controlling params from two different scales is not going to produce the same increments. Unless I enforce it via transformation shape as demonstrated by DarkStar. Makes all sense to me now.
Thanks for your time man,
-David

[MeldaProduction]

Great! And my pleasure