PlanetZ forums - scopeusers.com

Just a little tip here.

With this free VST plug-in you no longer need to do the ASIO-loopback (i.e. merger module) for recording your XTC-synths or XTC-effected audio in your favourite host.

Simply place Voxengo Recorder last in the chain, set bitrate, set file and hit "start".
When you hit play (or the recorder detects a sample-stream) it will start to tick.
The recorder will also stop once the stream has ended - like after you hit stop on the transport.
You've now recorded what you (or rather the recorder) heard.

Voxengo Recorder records to both 32bps as well as 32bps float. It is probably wise to choose the 32bps float option as this is the standard internal resolution for most hosts today. Also, this is what our XTC plug-ins put out due to the VST-standard.
You can do your own experiments, but I always choose to record to 32-bit float these days unless I am recording directly in the Scope OS-realm.
Results should be equal or better. The less conversion, the better.

... but still it gives worse quality in comparison with playback in project yet. Tape-It gives little more better result, than Voxengo Recorder, but also is worse than in playback mix in project. PEOPLE, I FOUND ONE AMAZING THING TO SOLVE THIS BIGGGGG PROBLEM. Needs simply to record at high sample rates. Its possible to make with help of NI Reaktor recorder. Hang Reaktor FX (tapedeck recorder) in master insert (JUST BEFORE MAXIMIZER, since maximizer as a rule does dithering, it might to carry damage quality in Reaktor recording!!!). Then set sample rate(SR) in the same Reaktor window SR more than in project in 2 times, at further record a playback in it. Although reaktor audio recording format has only 16 bit (but not 32 f.p.) still it is best way to record mix without loss of quality.

Tape-it (or the commercial one) surely looks like a better deal. Less hassle, but doesn't do 32bit int. It's not a big deal though since I advice on just using 32-bit float. _And this is while mixing down/freezing single tracks you have processed_. Once you pass several 32-bit float signals through a 32-bit float pipe-line you will obviously start to destroy the finer details of your sound.

Prototype:

Your other advice makes no sense - to record 16-bit w/ no dither and 88.2KHz+ ?

16-bit will truncate your data and over-sampling will do nothing but waste space when just recording. You need to process/generate at those rates too for any theoretical effect.

Voxengo recorder or tape-it @ 32-bit float should record what you hear.
I think you should examine your signal path if you get strange results with either.

No, bits influent only on gain changes. After recording at 16 bits you always convert to 32 bit float for further maximizing proccess. And then to dither this to 16 bit. Bits is division by Y-axis (amplitude), if quantity of division by X-axis is very high (as in occurance with high SR), then changes in bits (16 or 32 fp) is unheared by ear. I tried already variuos versions of these ones.

Gain changes, synthesis, modulation, distortion - these all affect bit-depth, thus I make it a big point to store tracks I process in 32-bit float. Anything below that and you will lose quality.

So you are saying that recording a 32f-bit/44.1Khz audio-stream at 16-bit/88.2KHz sounds the same or better than 32f-bit /44.1KHz?

I don't see how this is possible in theory. You are processing at a lower sample-rate than what you are recording. In effect you are recording the same sample twice without any interpolation and at reduced resolution.

If you would do both, when you listen back to the files this is how they will be processed (at 44.1KHz), viat ASIO-FLT drivers and 24-bit DAC:

32f-bit/44.1KHz -> converted to 32-bit int (all bits are used) through ASIO -> truncated to 24-bit through DAC = output is 24-bit (all bits are used) @ 44.1KHz

16-bit/88.2.KHz -> converted to 32-bit float and downsampled to 44.1KHz in sequencer -> converted to 32-bit int (16 lowest bits are empty) through ASIO -> truncated to 24-bit through DAC = output is 24-bit (8 last bits left empty) @ 44.1KHz

While you might not hear any difference at most setups (SN is 120dB at most), there is a difference. And this difference is important during processing for smooth results. Your method is doing nothing but removing the low level detail while adding nothing.

You're right, but just number of readings per time unit is different, thus converting to 32 bit float at 96kHz or 88 from 44.100 is more precisially describes original wave curve than at 44/32 float format since there are more readings per time unit. Difference between 16 and 32 float bits is in appearense of a digital noise when you change gain of wave since 16 bit is integer format and values are rounding itselves up to nearest integer values.

No.

Converting a 44.1KHz file to 88.2KHz does nothing but waste space! You have to actually create some usefull information to store in it.

I.e.: I clock my system at 48KHz even though I work with 44.1KHz files. This gives me a Nyquist frequency of 24KHz which gives more sampling points making room for more detail in the high end, shifting the aliasing noise higher in the inaudiable range.

BUT I get no benefit from unless I actually create some frequencies up there by the use of saturation, modulation etc. You can't create something out of nothing!

If you clock your system at 44.1KHz, but somehow record at twice that, you gain nothing. If you clock your system at 96KHz and record at the same, you gain nothing still if you are working with 44.1KHz material. Unless you create that material by artificial means.

If you do a mix i.e. Cubase and record that using Tape-it at 32-bit float, then what you hear is what you will get recorded. There is no way 16-bit/88.2KHz will sound better without doing something for it.

>Converting a 44.1KHz file to 88.2KHz does nothing but waste space! You have to actually create some usefull information to store in it.

Mate, you really didn't understand me! I told about RECORDING AT DOUBLE HIGH SR set in recorder!!! but this is not converting !!! The matter is even when you playback project at 44.100 all frequences included in this mix will not be reflected in the recording at the same 44.100 kHz, that's why playback quality at the same 44.100 is always better, than exporting audio at the same SR or recording audio at ones.

Please explain to me how you believe recording a 96KHz file while clocking your system at 44.1KHz magically will create information above 22050Hz.

You can not capture anything above the nyquist frequency which your audio card is clocked at...!
Even if you insert your Reaktor recorder at the mix bus and tells it to record in 96KHz your system is still working at 44.1KHz. You would have to clock your system to 96KHz during summing for this to have any effect.

Except theory a practice exists (that is our ear). Everybody says that native mixdown is much worse than project playback, it also concerned recording using external rec. plugins such as voxengo or tape-it. I hear myself real difference between voxengo record (at 32 bit float) and simply playback. Record is dark and NOT BRIGHT in comparison with original !!! And i'm not alone the same. Do your conclusions, mate. Maybe (although i dont know exactly) result will be much better when to use the other recording way, thats external equipment. I tell about ADAT recorder having optics and self hard disc.

Ok, please explain your full signal path.

Full path is illustrated in the Scope Manual pdf file.

Could you please be more informative? On what page?

'VDAT' partition of pdf manual to Scope cards since 4.0 (as i have). I myself didnt find vdat module in routing window menu. Probably it's only for Scope Professional. But very pity