Login

The most colorful preamps

Officially Licensed 3rd Party Developer Libraries
Free 3rd Party Programs

Re: The most colorful preamps

Postby cdsoundmaster » Fri Jul 08, 2016 8:59 pm

Testing the same programs at the same buffers will generate unique invert results as well- it is the positive complexity of the signal not a difference in timed buffer.

Buffer is about giving more time for the system to process, reducing the bottleneck before processing.

The points at which bottleneck can occur are:
speed of hard drive
speed of ram
speed of cpu
# of cores and # supported by OS and DAW
max ram


Then of course sample rate bit depth number of instances.

The more buffer, hypothetically, the more instances.
But this equation changes when you get beyond a standard 'average' computer from the past 3 years versus top of the line i7 with fastest motherboard options.

When you get to a certain speed, lowering to less than say 1024 can actually give more performance boost.

The same goes for Acqua instances. They are not processing different from NebPro Reverb at least up to a certain point. I don't know if later cores that claim zero latency are making changes to that affect but I personally do not use them for other reasons. The ones I use are stable to their NebPro preset counter-parts to the same quality and settings.
You can go into CDS Acqua xmland adjust dspbuffer for the identical results, and there will be small variation in every signal regardless of buffer- for the right analog-sounding reasons.

Brian wrote:This is what I thought - program rate and audio buffer size are two very different things.

Small audio buffer sizes (say 256 samples and under) will need a proportionally higher CPU load, but theoretically shouldn't have any influence on the program quality.

However I'm not sure about Nebulas internal dsp buffer parameter.

Was there anything random or with modulation in these audio tests? Not at my computer now so unable to check.
cdsoundmaster
Vip Member
Vip Member
 
Posts: 447
Joined: Tue Apr 20, 2010 2:18 am

Re: The most colorful preamps

Postby cdsoundmaster » Fri Jul 08, 2016 9:05 pm

This is why it should be based on setting one main quality setting in NebPro Reverb that works best for your system, and never touch any other settings except:
dspbuffer to tune to your speed for performance
rate cnvt to set to some high number- I set all the way because I do not care about latency for my work-
This guarantees you do not accidentally load from one rate to another with any errors.

Everything else should be left to the developer's choice for what is the highest quality choice for the program under the widest range of user options. Every change in the enginer, timed/freqd, kernels length, hipass, type of timing element rms evf17 etc, make a change to the actual program.

If you change something in MAST I recommend change while at INIT, save. close, reload. It IS possible, depending on NebPro version, that DAW will either not update immediately, Neb won't update immediately, or the PROGRAM won't update immediately. I have had ALL of these happen. The only consistent answer has been to save, close, reload, to be sure it is resetting itself to the identical condition only with the change you intended to make.

theheliosequence wrote:
jfjer379@gmail.com wrote:so what is the recommended settings for nebula exactly ? :)


Yes, It would be nice to have a list of programs that their quality is affected by the DSPbuffer amount. Or is there an easy way to see which programs are Timed and which are Freqd? It would also be nice to have a lower DSPbuffer amount available for rendering. This also brings up the question of whether or not there are any Acqua plugins also effected by this setting. I think a lot of people would want to know about this and I'm surprised that I haven't heard more about it.

Thanks to Shibata for bringing this to our attention and TimP for more info.
cdsoundmaster
Vip Member
Vip Member
 
Posts: 447
Joined: Tue Apr 20, 2010 2:18 am

Re: The most colorful preamps

Postby cdsoundmaster » Fri Jul 08, 2016 9:07 pm

I can't speak to latest core changes made to acqua for AA store products, but none of my CDS acqua plugins are affected by changing dspbuffer setting. You will get the same intentional variation from instance to instance at the same buffer as you do from changing buffer.


theheliosequence wrote:
jfjer379@gmail.com wrote:so what is the recommended settings for nebula exactly ? :)


Yes, It would be nice to have a list of programs that their quality is affected by the DSPbuffer amount. Or is there an easy way to see which programs are Timed and which are Freqd? It would also be nice to have a lower DSPbuffer amount available for rendering. This also brings up the question of whether or not there are any Acqua plugins also effected by this setting. I think a lot of people would want to know about this and I'm surprised that I haven't heard more about it.

Thanks to Shibata for bringing this to our attention and TimP for more info.
cdsoundmaster
Vip Member
Vip Member
 
Posts: 447
Joined: Tue Apr 20, 2010 2:18 am

Re: The most colorful preamps

Postby Shibata » Sat Jul 09, 2016 10:59 am

Duplicate post.
Last edited by Shibata on Sun Jul 10, 2016 10:27 am, edited 1 time in total.
User avatar
Shibata
User Level VII
User Level VII
 
Posts: 77
Joined: Tue Jul 09, 2013 8:06 pm
Location: Odessa Ukraine

Re: The most colorful preamps

Postby Shibata » Sat Jul 09, 2016 11:00 am

cdsoundmaster wrote:If you zero 128 to 128 or 1024 to 1024 from multiple instances you will find the same variation results.


From my previous posts

Shibata wrote:Look, i render the tape preset with 128 internal buffer two times, reverse phase and get zero. You see that RME mixer shows -оо ( infinity )
http://i.imgur.com/hGfBxF5.jpg


cdsoundmaster
The problem is not in the buffer and not timed vs freqd and not with quality.
But with RTE. It not compensated! It miss the rendering signal part by an amount equivalent to RTE ms time.
Therefore the compresors don't catch the transients and if you change the RTE at Edit-Glob-Rate to 0.180 ms for example - they catch much more signal.
Likewise with preamps, consoles and tapes, we get holey rendering signal by amount RTE time, because it not compensate. Thats why the bigger buffer lighter on CPU - it renders less signal, coz RTE time bigger. I can set buffer 128 and than change RTE at Glob page to 40 ms lets say, and i got less stress to CPU and ASIO.

Another problem - if i want to change RTE manualy at Glob page, i must change kernels to Timed engine, because RATE settings locked when kernels at freqd engine. And compressors usually have 10/10/10 ms length but preamps, consoles and tapes usually have 50/50/50 ms length. And changing it to Timed - very heavy to CPU and so i cant change the RTE)
User avatar
Shibata
User Level VII
User Level VII
 
Posts: 77
Joined: Tue Jul 09, 2013 8:06 pm
Location: Odessa Ukraine

Re: The most colorful preamps

Postby cdsoundmaster » Sun Jul 10, 2016 11:48 pm

The issue of RTE and DSPbuffer are two separate things.
I agree there are some things to watch out for with RTE, but...

This is why I highly, highly recommend letting that be based upon the editing of the developer. If, and this is a strong if, the program is designed to give the proper response for compression (and that is a much longer topic than you want here I'm sure), it has to have other things locked into place correctly in the program's settings in the n2p or, no matter what is done at the MAST page or global settings will get the same response on timing.

The Rate in MS on Global is tied to the program, the program type, AND the sample rate of the program AND the program material. This is all what has to be factored in when actually making the program. You will see that this number adjusts based on the program and program type as well, not only with the buffer size or length.

This is a setting that is recognized by the program. If you change it- it isn't about changing how Neb is responding to signals- as far as better or worse- but completely changing the intended character of that preset all together.

You should see the RTE length change based on 44.1kHz versus 96kHz to adjust at its best to the correct number of samples.

Shibata wrote:
cdsoundmaster wrote:If you zero 128 to 128 or 1024 to 1024 from multiple instances you will find the same variation results.


From my previous posts

Shibata wrote:Look, i render the tape preset with 128 internal buffer two times, reverse phase and get zero. You see that RME mixer shows -оо ( infinity )
http://i.imgur.com/hGfBxF5.jpg


cdsoundmaster
The problem is not in the buffer and not timed vs freqd and not with quality.
But with RTE. It not compensated! It miss the rendering signal part by an amount equivalent to RTE ms time.
Therefore the compresors don't catch the transients and if you change the RTE at Edit-Glob-Rate to 0.180 ms for example - they catch much more signal.
Likewise with preamps, consoles and tapes, we get holey rendering signal by amount RTE time, because it not compensate. Thats why the bigger buffer lighter on CPU - it renders less signal, coz RTE time bigger. I can set buffer 128 and than change RTE at Glob page to 40 ms lets say, and i got less stress to CPU and ASIO.

Another problem - if i want to change RTE manualy at Glob page, i must change kernels to Timed engine, because RATE settings locked when kernels at freqd engine. And compressors usually have 10/10/10 ms length but preamps, consoles and tapes usually have 50/50/50 ms length. And changing it to Timed - very heavy to CPU and so i cant change the RTE)
cdsoundmaster
Vip Member
Vip Member
 
Posts: 447
Joined: Tue Apr 20, 2010 2:18 am

Re: The most colorful preamps

Postby Thierr7 » Sat Jul 16, 2016 1:02 am

Thanks for these very Interesting Posts Michael :)
User avatar
Thierr7
Member
Member
 
Posts: 260
Joined: Sat Jan 28, 2012 8:52 am

Previous

Return to 3rd party libraries

Who is online

Users browsing this forum: UrbSgt and 9 guests