PlanetZ forums - scopeusers.com

Immanuel wrote:

medway wrote:Bits don't equal resolution.

I suggest some basic reading and understanding of how bits work as building blocks. Here goes a simple explanation. Well, it is not an explanation, but ... 4 bit recording ...

8
8
8
8
84
84
842
8421

0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15


Now, removing one bit

8
8
8
8
84
84
842
842

000
001
010
011
100
101
110
111

0
2
4
6
8
10
12
14

Where did 1, 3, 5, 7, 9, 11, 13 and 15 go, when we removed a bit? Well it disapeared - simple as that ..... the resolution dropped.

Bits determine noise floor when properly dithered, nothing else. Again very easy to test. When you remove bits you don't lose resolution, you just lose low level signal into the increased noise floor. The key here is understanding how dither linearizes digital signals.

darkrezin wrote:

medway wrote:

darkrezin wrote:As to *why* you should do this - if you reduce the master fader in:

1. an analog mixer - you reduce the signal-noise ratio

2. a digital mixer - you lose resolution (bits)

Bits don't equal resolution. Bits define a potential to capture dynamic range. In a properly dithered system less bits only means more noise (which can be construed as a basis for quality but given the signal to noise ratio of most analog gear digital doesn't have to work too hard to match it). The whole analogy of "resolution" when talking about bits is a false one.

That being said -6db coming out of a master fader results in the same amount of bits regardless of if it was the master fader lowered or the track it self. It's the final outcome that matters. The master is just a multiplier.

Unfortunately this is one of those myths that will be very hard to shake as abotut 99.9% of engineers out there don't understand it. Largely due to manufacturers using it as a selling point.

This was very badly expressed by me - thanks for picking up on it. In my experience, all digital mixers have a limited amount of headroom, but it is not the same across all of them. I have found that I get better results by not overdriving the mix bus and just keeping individual channel faders nice and low. In a digital mixer there's no need to have every channel touching 0dB, there's certainly no benefit from this. I find that if I keep channel faders low so I don't need to bring down the master fader or put some ugly ultra-limiter on it, the end result sounds more natural. Sadly I'm not a scientist when it comes to digital systems (I have no knowledge of DSP programming) so I can only go on what sounds better to me.

Keeping levels low is deffinatly good practice, if anything just so you don't have to worry about clipping somewhere in the chain. Unless you spend a lot of time testing every plugin and mixer you use it can get you in trouble. This is evident for instance by the fact that the Scope mixers seem to handle overs differently based on each mixer. I do find that a bit odd on CW's part though. You'd expect them to behave similarly.

We have enough dynamic range due to high bit rates that there's no need to overdrive levels in the digital domain, or track recorded signals near to 0db.
The idea is then you make up the gain at the very last process of the mix.

valis wrote:I concur with what darkrezin says above. I'm limited on time atm but I'll try to dig out some of the references I recall that discuss gain structure (in analog and digital). Also I recall a nice discussion here on planetz on how floating point data can 'erode' through multiple steps. Memory tells me that it was Tom that posted it up in one of the discussions about why mixing in 32bit int might be better, but don't quote me on that 100%. I do recall it voiced something I learned in my programming days (15 years ago!) better than I can voice it now by memory.

Anyway I'm happy to see the discussion is a bit more back on track. I certainly meant no ill will towards medway, I just would like to get at the details if we can. And if we can't then just be clear where we're stating opinion or facts we can't back up.

Valis I too don't mean any harm in these discussions and I appologize if I came across otherwise.

My understanding is that 32bit float is easier to program as it sorts itself out in the way of how the numbers are scaled. But that 32bit integer is ultimately more flexible as you don't have to use all the 8 bits for scaling and lose some of your mantissa.

From what I remember the issue with float is that although it works great for scaling single tracks, when they are added some precision is lost. And in addition the mantissa gets reduced to 23bits when adding two 24bit signals.

That being said a properly designed mixing environment should work whether its integer or float. The error in float signals is so small it's not worth worrying about. I believe its down to about -140db or so which is way beyond what anyone could detect audibly.

A lot of these arguments seem to be based on technical and not perceptible differences. The psychological factor in how we hear is much stronger than some these errors brought to attention.

WOW!!!!


82 replies.


WOW!!!!

Okay. I defiantly did not mean for this to be a war of words, or a battle over numbers. However, it is good to see how passionate some of you are about mixing and the Scope platform, I hope to join soon.



To the heart breaker Medway, thanks man.

But Let me quote you....

medway wrote:If anything is making your sound "flat and glassy" its the plugins not the mixing engine...........

..........Once you stop worrying about the platform/bits/sample rate you can actually focus on what makes a difference.

Now when you were referring to the plugins here, it made me think about something, and oh god I really hate these debates, but what is so special about the plugins in Scope, that makes them better than those comparable in Cubase or Sonar or Logic, after all, its just programming right. ?


Are you saying that the Scope programmers are better than these programmers. (mixing engine aside)?

Or is the programming the same?

Perhaps the fact that a dedicated dsp system, design for audio is being used that gives the programmers more power and flexibility in there coding, compromising nothing.

Native systems are designed to share resources with a power hungry operating system that I believe handcuffs the programmers.

Now, to agree with some of you, 4 tracks mixed together on any daw will not show the hole in the wall. However, the productions that I have done in Cubase, and other platforms, usually have a high track count, I for one do not wish for any compromises in sound quality. The y say that the errors are barely audible, but the missing link that I believe I am missing is very much subtle.


Now to piss myself off!!!

I have taken the time to finally give in to what some of medway was saying about the way I mix. I work on mostly urban music, that being said, there is a fanatic obsession with certain frequencies in urban music. So I've gone to work with this in mind. I’ve listened, and listened and listened. And I was startled to find something so obvious yet so overlooked by me, that I feel somewhat ashamed.
The low end in some of my mixes was just too much. There I have said it.

Now, I have taken the steps to put a 40hz filter on all my channels with a recent project that I’m working on, and I can say that it does sound more...uh...hmmm...what’s the word I’m looking for?..... Even. (The mix that is) However, there is still that subtle quality issue that I am still not getting. And that my friends is the reason why I am hell bent on finding a solution that will not compromise my workflow as well as my quality. ITB composing and mixing is all I know now. And Id rather not than let this go. -Yish

yish313 wrote:Native systems are designed to share resources with a power hungry operating system that I believe handcuffs the programmers.

that's one.
Also, DSP is realtime processing, like an amp which's working, native is serial, buffered in processing units.
Realtime processing is said to be better, and who am I to deny

Good find, to eliminate everything below 40 Hz, it's common practice.
I love the praphic eq plug for bass eq-ing, in combination with a lowcut I can get it exactly as I want
The ISON eq by De-Vice is able to control the natural small peak at the frequency cutoff value, specially important when you cut off the low end.
The Psy-Q is another valuable instrument for the lows, they get cleared up, get more push and get to be more tight.

yish313 wrote:
Now when you were referring to the plugins here, it made me think about something, and oh god I really hate these debates, but what is so special about the plugins in Scope, that makes them better than those comparable in Cubase or Sonar or Logic, after all, its just programming right. ?


Are you saying that the Scope programmers are better than these programmers. (mixing engine aside)?

Or is the programming the same?

well I'm not a programmer but simply because something is "just programmed" instead of built in hardware doesn't mean there's only one way to do it and there can't be difference in quality and sound between different plugins on different systems. I don't think Scope programmers are neccesarily better than VST programmers, but I do believe that the fact that it's running on dedicated DSP's gives good programmers the opportunity to use more powerful algorythms and better streamline the programming.

yish313 wrote:WOW!!!!


82 replies.


WOW!!!!

Okay. I defiantly did not mean for this to be a war of words, or a battle over numbers. However, it is good to see how passionate some of you are about mixing and the Scope platform, I hope to join soon.



To the heart breaker Medway, thanks man.

But Let me quote you....

medway wrote:If anything is making your sound "flat and glassy" its the plugins not the mixing engine...........

..........Once you stop worrying about the platform/bits/sample rate you can actually focus on what makes a difference.

Now when you were referring to the plugins here, it made me think about something, and oh god I really hate these debates, but what is so special about the plugins in Scope, that makes them better than those comparable in Cubase or Sonar or Logic, after all, its just programming right. ?


Are you saying that the Scope programmers are better than these programmers. (mixing engine aside)?

Or is the programming the same?
. -Yish

Yish,

Yes I understand it is a bit heartbreaking, I've been through it quite a few times myself, so I've been trying to save you the trouble of getting a new system only to find out it wasn't really the answer.

You are right it is ultimately down to the programming. I would say that overall Scope has a higher average of top quality plugs. Now Native has just as good plugs but you have to wade through a lot more choices.

With Scope if you stick with Vinco, Sonictimeworks Reverb, the CW synths etc... then you are getting some of the best stuff available in plugin format. With Native you have to be a little more selective.

At one point I do think DSP had an advantage but with todays CPU's its not a problem getting high quality with Native. I haven't done an A/B yet but the SonicTimeWorks reverb is now available in Native as the CSR reverb from IK Multimedia. It's getting alot of high marks as one of the best reverbs in Native. This shows that with todays CPU's you don't have to compromise quality comparable to a DSP system.

I think a little too much is made here on float vs integer and DSP vs Native. A properly programmed plugin shouldn't rely on the math used as long as the math is sufficient.

And all you have to do is compare some high quality Native plugins to see that they do stand head to head with their DSP counterparts.

I'm glad you found an answer already in how you were mixing. Hipass filters are essential to getting a clean mix thats loud and with clear bass response.
Trust me the more you focus away from the tools and more to your technique your mixing skills will vastly improve.

If you haven't already check out Gearslutz forum, they have a good one for R&B/Hiphop.

At one point I do think DSP had an advantage but with todays CPU's its not a problem getting high quality with Native. I haven't done an A/B yet but the SonicTimeWorks reverb is now available in Native as the CSR reverb from IK Multimedia. It's getting alot of high marks as one of the best reverbs in Native. This shows that with todays CPU's you don't have to compromise quality comparable to a DSP system.

If you would have done your A/B tests properly, you would have noticed that the STW Plate has the much better filter implimentation than the CSR, which results in a more realistic room simulation.

I mean the P100 does simulate the absorbing much more prefectly than the CSR counterpart.... The resulting reverbs sound cleaner and more spacious, while the native version always let some noisy high freq pass through. If you set the filter to extremer settings to cut away these frequencies the reverb is sounding more dull.

With the correct filter settings the P100 reverb sounds as "heavy" as a hardware reverb.

I aggree that the CSR is one of the best native reverbs. But compared to the STW version it cannot hold up against it. ( my personal opinion)

With open filter settings they both nearly sound the same.

scope plugins are hardware. they sound better than vsts in general. there are some excellent vsts that are as good as the best scope plugins, but they are extremely resource heavy for the most part. scope is a bargain for quality/price/usefullness. it will extend the capabilities of your in the box mixes.

hifiboom wrote:

At one point I do think DSP had an advantage but with todays CPU's its not a problem getting high quality with Native. I haven't done an A/B yet but the SonicTimeWorks reverb is now available in Native as the CSR reverb from IK Multimedia. It's getting alot of high marks as one of the best reverbs in Native. This shows that with todays CPU's you don't have to compromise quality comparable to a DSP system.

If you would have done your A/B tests properly, you would have noticed that the STW Plate has the much better filter implimentation than the CSR, which results in a more realistic room simulation.

Try reading my post again, I said I didn't a/b it. It was speculation.

I'd be interested in what the programmer has to say about the difference in filtering though.

yeah sorry ...

I misread it.

hifiboom wrote:yeah sorry ...

I misread it.

No problem. And I realize I shouldn't be making any types of insinuations before doing my own tests anyways.

medway wrote:...I'd be interested in what the programmer has to say about the difference in filtering though.

if I remember correctly his words were ... in SFP you can do filtering that's simply impossible in native programming...

cheers, Tom
(not the programmer, btw)

astroman wrote:

medway wrote:...I'd be interested in what the programmer has to say about the difference in filtering though.

if I remember correctly his words were ... in SFP you can do filtering that's simply impossible in native programming...

cheers, Tom
(not the programmer, btw)

Interesting, I wonder what the limitation is. I can understand DSP having some benefits but to say that something is "impossible" in Native makes me wonder why. Or does he strictly mean SFP here?

astroman wrote:

medway wrote:...I'd be interested in what the programmer has to say about the difference in filtering though.

if I remember correctly his words were ... in SFP you can do filtering that's simply impossible in native programming...

cheers, Tom
(not the programmer, btw)

yep, i remember that, too. and it was warp69. another big advantage in scope is the possibility for inserts within a device. you can't do this in vsts, dirk from brainworx (bx_digital plugin) told me...

medway wrote:

Immanuel wrote:

medway wrote:Bits don't equal resolution.

I suggest some basic reading and understanding of how bits work as building blocks. Here goes a simple explanation. Well, it is not an explanation, but ... 4 bit recording ...

8
8
8
8
84
84
842
8421

0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15


Now, removing one bit

8
8
8
8
84
84
842
842

000
001
010
011
100
101
110
111

0
2
4
6
8
10
12
14

Where did 1, 3, 5, 7, 9, 11, 13 and 15 go, when we removed a bit? Well it disapeared - simple as that ..... the resolution dropped.

Bits determine noise floor when properly dithered, nothing else. Again very easy to test. When you remove bits you don't lose resolution, you just lose low level signal into the increased noise floor. The key here is understanding how dither linearizes digital signals.

You are so wrong. Take a nice 24bit recording and do some bit crushing. Your ears should tell you what happens.

Immanuel wrote:

medway wrote:

Immanuel wrote: I suggest some basic reading and understanding of how bits work as building blocks. Here goes a simple explanation. Well, it is not an explanation, but ... 4 bit recording ...

8
8
8
8
84
84
842
8421

0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15


Now, removing one bit

8
8
8
8
84
84
842
842

000
001
010
011
100
101
110
111

0
2
4
6
8
10
12
14

Where did 1, 3, 5, 7, 9, 11, 13 and 15 go, when we removed a bit? Well it disapeared - simple as that ..... the resolution dropped.

Bits determine noise floor when properly dithered, nothing else. Again very easy to test. When you remove bits you don't lose resolution, you just lose low level signal into the increased noise floor. The key here is understanding how dither linearizes digital signals.

You are so wrong. Take a nice 24bit recording and do some bit crushing. Your ears should tell you what happens.

Emmanuel bit crushing doesn't utilize dither so yes of course it will incur distortions. As I've said before dither randomizes distortion present when reducing bit depths so that it becomes non correlated to the signal, in other words it becomes noise. The signal is still harmonically pure, but with a higher noise floor.

I'm not sure I can put it in any simpler terms. Perhaps you should take your own advice and do a little reading on the subject:

http://www.sonyoxford.co.uk/pub/plugins ... htm#dither

That link explains it very nicely (thanks to Paul Frindle)

Alternatively if you have Wavelab you can set the master output to 8bit, notice the distortion, then turn on dither and hear the distortion disappear along with the resulting higher noise floor.

(OT) are we going to quote that lengthy list of numbers with every new post again in this thread
except a 'quote' knob there's also a simple but nice 'reply' knob...

medway wrote:bit crushing doesn't utilize dither so yes of course it will incur distortions.

I see you describing loss of resolution there.

As I've said before dither randomizes distortion present when reducing bit depths so that it becomes non correlated to the signal, in other words it becomes noise. The signal is still harmonically pure, but with a higher noise floor.

...

Alternatively if you have Wavelab you can set the master output to 8bit, notice the distortion, then turn on dither and hear the distortion disappear along with the resulting higher noise floor.

But that 8bit signal will not have the same resolution = the same level of details and pureness as the original 16 or 24 bit signal. Therefor the bit count is not only about noise level.

hubird wrote: except a 'quote' knob there's also a simple but nice 'reply' knob...