Another thread about summing in scope

[ReD_MuZe]

it supports recording in 32bit integer? ive seen only float options in the project settings.
where is it?
maybe i should just move to reaper instead of logic

(PCM is just sampling)

[voidar]

It supports recording 32bps with WavPack and AIFF. Sadly not in WAV. Try it .

Then there is Wavosaur which handles 32 bit PCM WAV.

[ReD_MuZe]

i dont thin pcm means integer.

besides what is
32bps?

[voidar]

bps = bits per sample.

Well, they generally refere to PCM as the linear form. Anyway, in REAPER, unless the resolution is marked with "float" it is 32 bit linear.

I have requested 32 bit integer for WAV, but nothing yet. So you can in effect load tapes from VDAT, but not save.

[ReD_MuZe]

worth giving a try, altho reaper still feels kinda raw sometimes..

[voidar]

I like it as an recording/editing/mixing application. I don't find much inspiration with it when composing.

It works very well with SFP using ReaInserts. Then I just mix with the internal engine.

[Warp69]

you need to set your project to be 32bit float in project settings

That did it! Now it no longer truncate to 24bit for 32bit float.

32bit float and 24bit int has the same resolution at high levels, but 32bit float has betters resolution at low levels - Nothing new here.

intfloat.jpg (53.45 KiB) Viewed 2040 times

But why do 32bit float have the same resolution as 24bit int, when 32bit float can have greater precision even at high levels as shown earlier?

Thats because Scope use 'non-rounded 32bit float' and not 'rounded 32bit float'.

Non-rounded :

0,66666667 = 1431655765 = 1010101010101010101010101010101 (31bit) => E/-1 M/01010101010101010101010 (32bit float)

That gets translated back as 1010101010101010101010100000000 = 1431655680 (just like 24bit int)

Rounded :

0,66666667 = 1431655765 = 1010101010101010101010101010101 (31bit) => E/-1 M/01010101010101010101011 (32bit float)

The last digit is different in the mantissa, since we rounded up because the next digit in the 32bit is "1".

That gets translated back as 1010101010101010101010110000000 = 1431655808 (which is different from 24bit and non-rounded 32bit float).

1431655765 (32bit) - 1431655680 (non-rounded 32bit float) = 85
1431655765 (32bit) - 1431655808 (rounded 32bit float) = -43

[Mike Goodwin]

I would like to take a moment at this point in time to say that this thread is now officially way over my head.

Thank you

Carry on....

[dawman]

I enjoy seeing Warp69 and ReD_MuZe engaging here.
I don't know what they are talking about but they make great shit.
I have seen several well versed guys popping in & out of the developement threads, announcements, and here.
6 days until NAMM.......... all of this is vet positive, and yes Mike. you kicked the tires and lit the fires with this thread.
I was falling off my chair from boredom of this thread until page 7, now I stop in every few hours and yes I am doing the Reaper test also, as every demo I use I have chosen Reaper for.
I use VDAT for live playback and recording also, because it is painfully simple and it sounds better. Don't know why and don't care, but I am fascinated with the mathematical reasons now.
I have no idea my ears were so good at Math. ............

One of the best threads in a while though.
I want j9k to get drug in here sometime though.............where art thou ?

[Warp69]

Well - I just got a new toy - the Quantec 2496 - it's like late christmas So Im going back to my cave.

[dawman]

Rub it in yeah.........
I can't even find a studio in the states that has one of those.
I guaranfuckintee you didn't find that on ebay.

Bloody Lucky You........maybe a DSP emulation for Scope 5.0 & XITE-1......sorry,I don't have a begging avatar,.......Stardust to the rescue.

[voidar]

Warp69:

Any reason for not trying the ASIO 32 drivers?

[Mike Goodwin]

Well - I just got a new toy - the Quantec 2496 - it's like late christmas So Im going back to my cave.

Good god that thing costs 3600 POUNDS! It better put you in the dam room.

[Warp69]

Warp69:

Any reason for not trying the ASIO 32 drivers?

No, so I just did it - they're perfect - no errors.

Good god that thing costs 3600 POUNDS! It better put you in the dam room.

It's expensive, but not that expensive - I got mine for 2118 pounds.

[Mike Goodwin]

Warp69:

Any reason for not trying the ASIO 32 drivers?

No, so I just did it - they're perfect - no errors.

Good god that thing costs 3600 POUNDS! It better put you in the dam room.

It's expensive, but not that expensive - I got mine for 2118 pounds.

Ah! a bargin
So is that little understated sentence up there saying "they're perfect - no errors" to say that if I use the 32-bit flt ASIO drivers in Scope I can record into my host (Ableton Live) that uses 32-bit flt with no errors? I keep asking these simpleton questions because I keep assuming wrong! It is a little embarrassing but hey sometimes if you want to learn you have to shed the ego.

[Warp69]

So is that little understated sentence up there saying "they're perfect - no errors" to say that if I use the 32-bit flt ASIO drivers in Scope I can record into my host (Ableton Live) that uses 32-bit flt with no errors?

No - only the pure 32bit int drivers is without errors.

The 32bit flt drivers is second best.

[Mike Goodwin]

No - only the pure 32bit int drivers is without errors.

The 32bit flt drivers is second best.

So the only way to not loose data is to use a host that supports int. Otherwise no matter what when you go from the internal int. environment of scope into flt. you loose some data. And as previously determined in this thread between 24-bit and 32-bit it is just different data that is lost. 24-bit favoring the loud portions of audio and 32-bit favoring the quite portions. I hope that I have this all right this time :l

[wayne]

Let's rock, yea verily

[Warp69]

So the only way to not loose data is to use a host that supports int.

Yes, 32bit int and only 32bit int.

Otherwise no matter what when you go from the internal int. environment of scope into flt. you loose some data.

Correct.

And as previously determined in this thread between 24-bit and 32-bit it is just different data that is lost. 24-bit favoring the loud portions of audio and 32-bit favoring the quite portions. I hope that I have this all right this time :l

Wrong. 32bit flt can represent every 24bit number AND more.

So for best representation of Scope audio :

1) 32bit int
2) 32bit flt
3) 24bit int
4) 16bit int

[astroman]

unfortunately I forgot about the source, but someone wrote that (in his opinion) the crucial difference between int and float was the continuity of the numeric space.
While float can represent individual values with an incredible precision, it's 'full resolution' doesn't apply to to the entire range of numbers.
I'm not enough math guru to judge the relevance in this specific context, but I found the concept convincing...

cheers, Tom