Bitwig 6

[coroknight]

According to the latest steam hardware survey, 95% of people have cpus that support AVX2. Sure, gamers probably have slightly better computers than music producers but it's not like gamers are all rich, they have to buy affordable hardware too.

So, I just checked, and according to the Steam Hardware Survey, only 1.88% of users are on Mac. That's fewer than users which have an NVIDIA GeForce RTX 4080. Bitwig should drop all of the Mac and Linux support, and make Windows require a GTX 4080 or higher


firefox_5D1beyYvWT.png

Another attempt at a stupid gotcha. Maybe instead of dismissing real statistics you could focus on where they are applicable which is clearly windows computers.

Point is, AVX2 probably isn't as uncommon as you implied. Like I said, gamers probably have better computers than music producers but I highly doubt it's as bad as a 30% drop in customers which is what you said. Clearly Bitwig thinks it's fine.

[psycha0s]

So what? Those of us with computers that actually support AVX2 get a performance boost, while those that emulate it don't which is the same as today except now Bitwig only has to support one instruction set.

Can't wait to hear what other bullshit excuses people come up with.

You forgot about ARM Neon, so it's more than a single instruction set anyway.

[tumface]

Another attempt at a stupid gotcha. Maybe instead of dismissing real statistics

I literally used the same statistics you did to support your argument, using the same reasoning you did to support your argument.

[coroknight]

So what? Those of us with computers that actually support AVX2 get a performance boost, while those that emulate it don't which is the same as today except now Bitwig only has to support one instruction set.

Can't wait to hear what other bullshit excuses people come up with.

You forgot about ARM Neon, so it's more than a single instruction set anyway.

Clearly I meant one less instruction set within the domain of applicability. Obviously AVX2 doesn't apply to ARM processors. Great point! Always fun when people think litigating language is a good argument.

[coroknight]

Another attempt at a stupid gotcha. Maybe instead of dismissing real statistics

I literally used the same statistics you did to support your argument, using the same reasoning you did to support your argument.

No you're trying to sound smart by over-analyzing casual discussion on a music forum. I'm not going to write an essay outlining my argument. If you can't use critical thinking to figure things out then that's on you.

[tumface]

When the people replying to your arguments, which you chose to post, have both technical and reasoning based refutations to your arguments, and your only reply is "stop over-analyzing, I'm not going to write an essay, you can't think critically" then it might be time to stop.

[coroknight]

I give in, you all are right. How silly of bitwig to rely on an instruction set that's almost 15 years old.
I'm going to stop arguing this and email bitwig support immediately. How dare they.

[tumface]

I give in, you all are right. How silly of bitwig to rely on an instruction set that's almost 15 years old.
I'm going to stop arguing this and email bitwig support immediately. How dare they.

AMD 3DNow! is a 27 year old instruction set, and Bitwig hasn't chosen to rely on it.

[coroknight]

When the people replying to your arguments, which you chose to post, have both technical and reasoning based refutations to your arguments, and your only reply is "stop over-analyzing, I'm not going to write an essay, you can't think critically" then it might be time to stop.

You're totally right. I forgot to think about all the ARM Mac's they'll stop supporting by switching to AVX2. Very valid of you to point that out.

[coroknight]

I give in, you all are right. How silly of bitwig to rely on an instruction set that's almost 15 years old.
I'm going to stop arguing this and email bitwig support immediately. How dare they.

AMD 3DNow! is a 27 year old instruction set, and Bitwig hasn't chosen to rely on it.

Lol what a stupid thing to say, as if that's relevant.

[pdxindy]

It looks like Add and Multiply relative options are gone and there's no way to do the same.

Personally, I won't be missing this one. It was fussy and complicated, and never worked properly once you started editing and moving stuff around. I will just automate a macro knob if I want to do it.

I'll miss it a bit (for maybe a week), but overall the improvements vastly outweigh the loss.

[koalaboy]

The new (better) way in 6 is to create an automation clip of a different length of the content clip, then set the automation clip to loop. When you resize the content clip, the automation clip will resize to match, but with different loop lengths. Perfect.

Apologies - the 'better way' we're looking for isn't actually the free-running (I understand how that works now) but how to reproduce the 'relative' automation clips from v5.

It looks like Add and Multiply relative options are gone and there's no way to do the same.

Hopefully they will enable multiple automation lanes for the same parameter, and the ability to specify if an automation clip is absolute (default), additive, or multiplicative... in a future update. It would be great to have overlapping/aggregating automation clips for a single parameter on the arranger view, for more ambient/longer pieces.

[tumface]

Hopefully they will enable multiple automation lanes for the same parameter, and the ability to specify if an automation clip is absolute (default), additive, or multiplicative... in a future update. It would be great to have overlapping/aggregating automation clips for a single parameter on the arranger view, for more ambient/longer pieces.

Yeah, that would be sweet. Right now, you have to set that up manually with macro knobs, and it's not ideal.

[psycha0s]

Clearly I meant one less instruction set within the domain of applicability. Obviously AVX2 doesn't apply to ARM processors. Great point! Always fun when people think litigating language is a good argument.

I'm not trying to argue. I'm just noting that the devs have to support multiple instruction sets in any case. I highly doubt they write SIMD intrinsics or assembly manually for each one, so dropping SSE4 wouldn't significantly simplify their work. You can even test SSE4 code on an AVX2 machine, it's testing ARM Neon requires separate hardware.

[Xbucket]

Personally, I won't be missing this one. It was fussy and complicated, and never worked properly once you started editing and moving stuff around. I will just automate a macro knob if I want to do it.

thanks for this - with this there is no need for the other options. didnt thought about it!