An introduction to computer audio
Does what is say, applying Digital Signal Processing to the audio.
Beside the normal stuff like tempo & pitch, effects, etc. it also offers advanced features like Digital Room Correction, active crossovers, parametric equalizer, etc.
All processing is done in 64 bits. This will keep quantization errors down.
By definition DSP is altering the bits.
As a consequence, the signal can become too loud (digital clipping, the numerical value generated exceeds the maximum possible value)
In all DSP dialogues you can choose between Flat line overflow (clipping) and Clip Protection.
The latter scales down the volume.
You can chance the order of the modules by using drag/drop.
This is also the order of execution.
If you hoover the DSP icon on the toolbar, the audio path pops up.
It will tell you what DSP is active.
All kind of enhancements quite rightly called effects here.
The equalizer is a nice example.
You want a more dynamic sound so turn up the bass and the treble.
As you boost the bass, you run the risk of clipping as a lot of audio is produced at a very high level.
Sounds counter intuitive (you want more) but you can obtain the same result by lowering the mid-range.
As you don’t boost anything, you avoid clipping.
As stereo recordings are produced to be played over speakers, on a headphone the stereo becomes very pronounced. A crossfeed reduces the stereo image.
This is an important part of DSP studio.
If you are using WASAPI set the bit depth to match your audio device.
For each sample rate you can set the conversion.
If you have a DAC with 2 clocks and it won’t play above 96 kHz you might decide to down-sample 176 to 88 and 192 to 96.
If you are using DS (Direct Sound) beware of double re-sampling.
Win will re-sample the audio to the setting in the audio panel.
If you use the re-sampler of JRiver, set it to match the setting in the Win audio panel.
I am inclined to call this placement correction as in general room correction is correcting the system response using a target curve based on measured response. See Convolution.
A bit technical description for digital room correction.
The idea is simple:
You measure the response of your system using a calibrated mike.
You might see some nasty peaks or dips instead of a linear response.
You create a target curve, in essence one that correct the response (if it is a peak, make a dip in the target curve).
On playback the signal of the audio is merged with the target response (convolution)
This is the way correct for anomalies due to the interaction between speakers and room.
Pretty advanced but not easy to implement properly.
This allows you to add all kind of filters.
The low-pass and high-pas filters are Butterworths filters.
If needed you can use this to build an active crossover.
You need a multi-channel soundcard to drive a set of amps.
The results can be viewed in the analyzer.
A new toy added to JRiver 18.
It allows you to change the bit depth.
Play a 24 bit recording and lower the bit depth until you hear an audible degradation.
A nice toy to learn what bit depth does on your system.
You can use the space bar to toggle.
Shows you the magnitude (probably in dB) over the frequency range.
A nice tool to test your filters.
You can add all kind of plug-ins