Schiit Happened: The Story of the World's Most Improbable Start-Up

[Bowmoreman]

Everything is available at all times.

Well, sort of.

Loudness is real loudness, meaning it's coupled to the Mimir's volume setting. The lower you turn Mimir's own volume, the more its loudness curve will affect the signal Mimir puts out.

Mimir's EQ works independent of Mimir's volume, but if you keep it at full volume and use positive gain in any of its three EQ bands, you might run into digital clipping, depending on your source material. Forkbeard helps you avoid that issue by calculating the ideal volume setting for whatever EQ setup you have at any given time and will show you a red bar on Mimir's volume control to clearly visualize how much you'd have to turn down volume to be guaranteed to never run into digital clipping.
Yes, some digital EQs don't run into digital clipping when you use positive gain. But that's because they attenuate the entire frequency spectrum first to avoid that. We don't, which is why Mimir's EQ is actually as lossless as a digital EQ can be.

Balance control is independent from all of the above, you can leave everything else bypassed and only use balance, and you'll not run into any issues.

So now, after a couple of day’s with Mimir, I better understand the ^ comments, but it still begs several questions (for me, perhaps PEBCAK is involved, but…):

1. Given the Loudness is ONLY available with digital volume engaged, what is the tradeoff in resolution versus curve boostings? My hope had been that I could apply Fletcher Munson and then control overall system gain via Kara’s stepped attenuator, but it *seems* that I’d have to cut digital volume by quite a bit, to be able to get FM-level boosts in (especially) the bass realm… In initial playing, I cut the digital volume by an indicated 6dB, and then turned on loudness, in my room (26x16x7.5’) I really didn’t hear much of a difference.

Question 1: how much digital volume “down-ness” is needed to begin to get hearable FM loudness boostage? - i.e. must one suffer significant digital resolution loss in order to get significant Loudness benefit?

2. I did a lot of playing with the PEQ, and have come to a couple of conclusions. First, 3 bands simply aren’t enough for a big room with big speakers. I shall play with it in my office right shortly. Secondly, the behavior of boosts causing digital clipping did manifest itself. So this led to a situation where I’d try and “build my own” loudness curve “underneath” the PEQ flatline indicator, but it required going back/forth (a lot!) to not get “in the red”.

Question 2: might it be possible to update/modify the UI such that one “builds” a desired PEQ curve, and then “selects” the entire thing/curve, and “drag” the entire curve downwd to prevent the “red”, or have an “auto-level” to push digital volume to exactly as high as the currently chosen PEQ curve allows (mathematically)? That would be a huge ease of use improvement, and especially more-so when we get the 5 band PEQ in Gumby (and even more so should Byggy get more than 5).

Question 3: is it possible to have all three PEQ points be selected (frequency-wise) however we wish, or are there some restrictions/limits? (It might be my fat fingers and I just need to try harder, but I couldn’t figure out how to have my PEQ points be at (say) 30Hz, 80Hz, and 160Hz). Is this doable, or are any of them “range limited”?

 

[Paladin79]

Time to return work after my lunch break. Listening to this on Folkvangr with the Senns. Superb.

A favorite album, especially her version of Strange Fruit.

 

[golfbravobravo]

Having grown up in a town of 10,000 and gone to a town of 500,000 in college I certainly had some wild years that caused me to change my way of thinking. That was a time of social discord, free love, and occasional substance abuse. While bourbon, women, and gambling were all fun pastimes I got to the point where I could have been shipped to Viet Nam if I did not get serious about my education. There were also three incidents that could have changed my life drastically had I not had good friends to pull me from the brink. Ah the good old days.

All I can say is that it's a bloody good job there were no cell phone cameras in the 70s & 80s!

My maternal grandfather's business was bottling Guinness and other delectations in East Yorkshire. I only remember that we didn't drink at all in our household when I was growing up. Turns out that my dad had a few "adventures" early in their marriage and an agreement was made for no alcohol thereafter. That persisted until my mother died (very young) when my father resumed modest alcohol consumption.

I never had a real blackout drunken session. The closest I came was in Bangkok where we were guests at the Walton Hotel (despite its name, a US forces club) and I awoke in house of very ill repute. I was awoken by a noisy cockerel, as was my fellow NAAFI manager from another ship, Horace. He created an interesting kerfuffel by strolling, naked except for his socks, to deal with said cockerel. We left shortly afterwards.

I was proud that, for me, this must have been a love encounter as I had, wisely, gone ashore with minimum cash and therefore there could have been no exchange of specie. My colleagues (I won't call them friends) revealed that they had negotiated "good terms" and demanded repayment. To this day I do not know if they told the truth or not.......

Needless to say, duty free cigs didn't tempt me back to smoking, and duty free good booze didn't make me into any more than a modest drinker. The only remnant of those days is that the only spirit I regularly drink is gin in a G&T.

 

[Bowmoreman]

Recioto della Valpolicella Amarone Reserva was my favorite wine prior to health issues forcing me to stop drinking alcohol (your sympathies are noted and appreciated but I've managed to move on ). Barolo was my second favorite.

Quintarelli has long produced the greatest Amarone but the prices are now astronomical: $600-900 a bottle! We used to buy cases for $240 back in the eighties. Those Amarone I tasted several years ago (ok I cheated) in the $40-50 range weren't much, if any, better than Masi's Campofiorin, a mini Amarone (grapes not dried as long) that still sells for around $20 a bottle.

Edit: Bertani used to be the second best producer: their wines now seem to go for between ~ $100-150. Masi's Amarone was good also and is a bit less expensive than Bertani. There is also a producer by the name of Dal Forno (whose wines I've never tasted) that seems highly regarded but they're in the multi hundred dollar range as well.

We long have called Masi Campofiorin “poor mans Amarone”, as it’s made in the Ripasso style (taking the fremented lees from real Amarone and fermenting fresh Valpolicella with them)… long been in the $20 range and a good bargain… regular Masi Amarone goes on sale a few times a year and I can get it in the $50 range… some other good ones in the 50-75 range… my days of paying more - no matter HOW good - are basically over. TBH I’ve not found an Amarone that wasn’t at least good…

 

[golfbravobravo]

Recioto della Valpolicella Amarone Reserva was my favorite wine prior to health issues forcing me to stop drinking alcohol (your sympathies are noted and appreciated but I've managed to move on ). Barolo was my second favorite.

Quintarelli has long produced the greatest Amarone but the prices are now astronomical: $600-900 a bottle! We used to buy cases for $240 back in the eighties. Those Amarone I tasted several years ago (ok I cheated) in the $40-50 range weren't much, if any, better than Masi's Campofiorin, a mini Amarone (grapes not dried as long) that still sells for around $20 a bottle.

Edit: Bertani used to be the second best producer: their wines now seem to go for between ~ $100-150. Masi's Amarone was good also and is a bit less expensive than Bertani. There is also a producer by the name of Dal Forno (whose wines I've never tasted) that seems highly regarded but they're in the multi hundred dollar range as well.

I love Amarone and tested quite a few in the Veneto. I was introduced to Ripasso (perhaps your mini Amarone?) there and found that more acceptable to my wallet. We visited quite a few wineries and the one that sticks in mind (despite being the 800lb gorilla of the region) is Allegrini. Their Palazzo della Torre has become an affordable staple over the years, though vintages can be a little more variable than I would like.

 

[ArmchairPhilosopher]

1. Given the Loudness is ONLY available with digital volume engaged, what is the tradeoff in resolution versus curve boostings? My hope had been that I could apply Fletcher Munson and then control overall system gain via Kara’s stepped attenuator, but it *seems* that I’d have to cut digital volume by quite a bit, to be able to get FM-level boosts in (especially) the bass realm… In initial playing, I cut the digital volume by an indicated 6dB, and then turned on loudness, in my room (26x16x7.5’) I really didn’t hear much of a difference.

Question 1: how much digital volume “down-ness” is needed to begin to get hearable FM loudness boostage? - i.e. must one suffer significant digital resolution loss in order to get significant Loudness benefit?

Depends on what you're going for. I'd just play around with it a bit more.

What you can do to get a better feel for what Mimir's loudness compensation sounds like you could turn Mimir down to 0% and your preamp up to 100% (in that order!).
Start play something, and then start to turn Mimir up to where you would your volume usually have.
Now go to the screen with the loudness curve and swipe vertically to scale it up or down (goes all the way down to 0% loudness).
That way, you can get a feeling of what the loudness compensation sounds like.

But yeah, 6dB attenuation isn't a ton, so I wouldn't expect too much of an effect.

2. I did a lot of playing with the PEQ, and have come to a couple of conclusions. First, 3 bands simply aren’t enough for a big room with big speakers. I shall play with it in my office right shortly. Secondly, the behavior of boosts causing digital clipping did manifest itself. So this led to a situation where I’d try and “build my own” loudness curve “underneath” the PEQ flatline indicator, but it required going back/forth (a lot!) to not get “in the red”.

Question 2: might it be possible to update/modify the UI such that one “builds” a desired PEQ curve, and then “selects” the entire thing/curve, and “drag” the entire curve downwd to prevent the “red”, or have an “auto-level” to push digital volume to exactly as high as the currently chosen PEQ curve allows (mathematically)? That would be a huge ease of use improvement, and especially more-so when we get the 5 band PEQ in Gumby (and even more so should Byggy get more than 5).

The pEQ is done this way on purpose.

Analog signals can be boosted as much as the amplification circuit allows, even beyond the original "100%" of the signal strength. As long as you don't "hit the rails," essentially, you can boost to your heart's content without getting into analog clipping.

But you can't do that with a digital signal. A 16, 32, or 64 bit number covers a final range of values. You can't go beyond that range. So if you want to "boost" a digital signal that's already getting close to the edge of that range, you'll run into digital clipping.

To get around that issue, a lot (most?) digital (p)EQs tend to normalize their output based on the resulting curve. Meaning they calculate the final EQ curve based on your inputs, determine the highest point on that curve, and then drop the entire signal by about the same amount that you're actually trying to boost the signal by, and only then will they apply the EQ curve to your signal.

That's fine, and it works. Kind of. This way you get your boost (or so you think, anyway), but you'll never run into digital clipping. But since your entire signal gets attenuated first to create the headroom that you will need for the boost you want, you lose detail. The more you boost, the more detail you lose.

I don't think that's the best approach, for two reasons:

First: Not all music is mixed so that it's so close to the edge of the range of the digital signal that you actually get into digital clipping if you boosted it by a few dB. Classical music is one of those. Some acoustic is another. And so with the approach we're taking with our digital pEQ, you get to have your digital signal remain as intact as possible and still be able to boost a bit via your pEQ settings.

And second: Lack of transparency. With our approach it is more intuitive to understand what's going on, and you have a maximum of control over what's happening to your digital signal.

Besides, as a rule of thumb with any EQ, analog and especially digital: Always try to attenuate and avoid to boost.

None of this applies to the loudness compensation, by the way, even though it boosts parts of the signal, precisely because it is coupled to Mimir's volume. Because you have to lower your volume to get loudness compensation to begin with, it's always guaranteed that you never get into digital clipping even though some boost is applied to the signal.
Which is also one of a handful of reasons why we don't couple that loudness compensation to your Kara's or Freya's attenuation ladder, even though we could technically use their volume levels to calculate a corresponding loudness compensation curve in the DAC. But then you'd either run into digital clipping from the boost that's applied, or we would have to drop the overall digital volume first before the loudness compensation gets applied, which would degrade your signal slightly behind the scenes. I don't think anyone would be all too thrilled about that.

Question 3: is it possible to have all three PEQ points be selected (frequency-wise) however we wish, or are there some restrictions/limits? (It might be my fat fingers and I just need to try harder, but I couldn’t figure out how to have my PEQ points be at (say) 30Hz, 80Hz, and 160Hz). Is this doable, or are any of them “range limited”?

Each band has its own range of frequencies it can cover:
Band 1 covers 20Hz through 190Hz, band 2 covers 200 through 1,900, and band 3 covers 2,000 through 20,000Hz. You can freely set each band to whatever you want, as long as it's within their respective ranges.

Selecting more than one band at a time is problematic from a usability perspective. Not impossible, but problematic, because things get a lot less intuitive quickly.

If sliding your finger isn't accurate enough for fine control, try rolling your finger tip. That's an decades-old trick with touch screens in general.

 

[bagwell359]

Yes, my mother was an alcoholic, and not wanting to turn out like her was part of my "get more sober" mental discipline. Well, that plus being threatened with military jail by my training officer....

I saw my father sober only rarely, but never drunk. He was the penultimate drink every 2 hours from morning until night, buzzed I guess is the best description. It takes a lot to get me over my skiis and I feel poorly well before. My health/budget/lack of drinking buddies keeps me to about 15 times a year and never more than two.

@ArmchairPhilosopher "Each band has its own range of frequencies it can cover:
Band 1 covers 20Hz through 190Hz, band 2 covers 200 through 1,900, and band 3 covers 2,000 through 20,000Hz. You can freely set each band to whatever you want, as long as it's within their respective ranges."

Hmmm... So I assume thats going to be similar with the Gungnir 2 when it comes? 20-100, 100-440, 440-1800, 1800-6500, 650-20k? They are all bell curves w/ Q adjustments I believe.

 

[classfolkphile]

I love Amarone and tested quite a few in the Veneto. I was introduced to Ripasso (perhaps your mini Amarone?) there and found that more acceptable to my wallet. We visited quite a few wineries and the one that sticks in mind (despite being the 800lb gorilla of the region) is Allegrini. Their Palazzo della Torre has become an affordable staple over the years, though vintages can be a little more variable than I would like.

Yes, Campfiorin is made via the Ripasso method.

 

[Ripper2860]

Happy Monday (or Tuesday if applicable)!!

(Nice music for a rainy day)

** Who says accordions can't do Jazz.

@ArmchairPhilosopher "Each band has its own range of frequencies it can cover:
Band 1 covers 20Hz through 190Hz, band 2 covers 200 through 1,900, and band 3 covers 2,000 through 20,000Hz. You can freely set each band to whatever you want, as long as it's within their respective ranges."

Bummer. So I assume thats going to be similar with the Gungnir 2 when it comes? If so, its of limited use compared what I use on UAPP, often using 3 bands in the bass and 2 in the treble or visa versa. Are they just bell curves, or is the bass and treble shelved?

As I understand it Gungnir 2's EQ will be 5 bands as it has a more powerful SHARC DSP, unlike Mimir's PIC processor . That is unless I'm getting it confused with Byggy and Singular's FPGA processing power or am just flat out wrong due to blacking out.

 

[ArmchairPhilosopher]

So I assume thats going to be similar with the Gungnir 2 when it comes?

5 bands for G2 and Byggy, but also clamped to their respective frequency ranges.

Are they just bell curves, or is the bass and treble shelved?

Peak / dip.

 

[JamminVMI]

I fully agree that a sink is exactly where Kraft mayo belongs.

How is that considered mayo in the first place?

 

[Bowmoreman]

Depends on what you're going for. I'd just play around with it a bit more.

What you can do to get a better feel for what Mimir's loudness compensation sounds like you could turn Mimir down to 0% and your preamp up to 100% (in that order!).
Start play something, and then start to turn Mimir up to where you would your volume usually have.
Now go to the screen with the loudness curve and swipe vertically to scale it up or down (goes all the way down to 0% loudness).
That way, you can get a feeling of what the loudness compensation sounds like.

But yeah, 6dB attenuation isn't a ton, so I wouldn't expect too much of an effect.


The pEQ is done this way on purpose.

Analog signals can be boosted as much as the amplification circuit allows, even beyond the original "100%" of the signal strength. As long as you don't "hit the rails," essentially, you can boost to your heart's content without getting into analog clipping.

But you can't do that with a digital signal. A 16, 32, or 64 bit number covers a final range of values. You can't go beyond that range. So if you want to "boost" a digital signal that's already getting close to the edge of that range, you'll run into digital clipping.

To get around that issue, a lot (most?) digital (p)EQs tend to normalize their output based on the resulting curve. Meaning they calculate the final EQ curve based on your inputs, determine the highest point on that curve, and then drop the entire signal by about the same amount that you're actually trying to boost the signal by, and only then will they apply the EQ curve to your signal.

That's fine, and it works. Kind of. This way you get your boost (or so you think, anyway), but you'll never run into digital clipping. But since your entire signal gets attenuated first to create the headroom that you will need for the boost you want, you lose detail. The more you boost, the more detail you lose.

I don't think that's the best approach, for two reasons:

First: Not all music is mixed so that it's so close to the edge of the range of the digital signal that you actually get into digital clipping if you boosted it by a few dB. Classical music is one of those. Some acoustic is another. And so with the approach we're taking with our digital pEQ, you get to have your digital signal remain as intact as possible and still be able to boost a bit via your pEQ settings.

And second: Lack of transparency. With our approach it is more intuitive to understand what's going on, and you have a maximum of control over what's happening to your digital signal.

Besides, as a rule of thumb with any EQ, analog and especially digital: Always try to attenuate and avoid to boost.

None of this applies to the loudness compensation, by the way, even though it boosts parts of the signal, precisely because it is coupled to Mimir's volume. Because you have to lower your volume to get loudness compensation to begin with, it's always guaranteed that you never get into digital clipping even though some boost is applied to the signal.
Which is also one of a handful of reasons why we don't couple that loudness compensation to your Kara's or Freya's attenuation ladder, even though we could technically use their volume levels to calculate a corresponding loudness compensation curve in the DAC. But then you'd either run into digital clipping from the boost that's applied, or we would have to drop the overall digital volume first before the loudness compensation gets applied, which would degrade your signal slightly behind the scenes. I don't think anyone would be all too thrilled about that.


Each band has its own range of frequencies it can cover:
Band 1 covers 20Hz through 190Hz, band 2 covers 200 through 1,900, and band 3 covers 2,000 through 20,000Hz. You can freely set each band to whatever you want, as long as it's within their respective ranges.

Selecting more than one band at a time is problematic from a usability perspective. Not impossible, but problematic, because things get a lot less intuitive quickly.

If sliding your finger isn't accurate enough for fine control, try rolling your finger tip. That's a decades-old trick with touch screens in general.

Thanks, @ArmchairPhilosopher exactly the deets I needed to play more intelligently. What will the 5 pEQ ranges be on Gumby? Will we have ability to have, say 3 of them in the 200Hz and below realm, perchance? Or perhaps fewer range restrictions (possibly due to more computes being available on the DSP than on the Unison CPU?).

When you say “set Kara volume to 100% do you literally mean cranked to the clockwise max? Or do you mean the max normal volume I’d listen to were no Mimir be in the chain?

 

[fidgeraldo]

Happy Monday

Indeed!


Cheers!

 

[DougD]

This is my favorite version

re: "The Lion Sleeps Tonight" ... I'll counter with this non-traditional version. Hint: things start to go a wee bit crazy around 2:30.

 

[bagwell359]

5 bands for G2 and Byggy, but also clamped to their respective frequency ranges.


Peak / dip.

Well, actually if the bands are spaced about 2-2.25 octaves apart it turns out to be pretty useful for my speakers. But hunting planar headphone CSD's not happening.