- This topic has 22 replies, 10 voices, and was last updated 10 years, 8 months ago by
.
-
Posts
-
Hi guys!
I will be creating Block Diagrams and writing documentation over the next few days. They will eventually be posted on the main site, so I figured I might as well post them here first.
Documentation is not the most fun part about developing products, but somebody's gotta do it. <!– s:) –>:)<!– s:) –>
Best,
///LeifHi Leif,
[quote author=”Leif”]I will be creating Block Diagrams and writing documentation over the next few days. They will eventually be posted on the main site, so I figured I might as well post them here first.[/quote]GREAT!!
Gee, I’ve been away on a Production Project for some time I come back to the site and there is all this new stuff to take up all that free time I was planning to have. 🙂
All the new updates look excellent and I am VERY keen to try the ASIO version.
Scott
Where is the pre-emphasis inserted?
In Breakaway Broadcast, the Final Limiter threshold is raised a few dB to lift it out of the way. The audio signal is then sent to the FM back-end, where pre-emphasis and clipping takes place.
I have found that applying pre-emphasis before multiband compression (in order to dynamically vary the amount of pre-emphasis), while an effective way to keep from overloading a primitive back-end clipper, has some serious fidelity drawbacks. With Breakaway’s FM back-end this wasn’t an issue, so I add pre-emphasis after the multiband processing, thus pre-emphasis is the constant curve it’s supposed to be, regardless of program-material. Thus, it matches the de-emphasis in receivers, which is also constant regardless of program material.
All this will be shown in the BBP block diagram, which I haven’t made yet. 🙂
///Leif
[quote author=”Leif”]
Documentation is not the most fun part about developing products, but somebody’s gotta do it. 🙂
[/quote]I know what you are talking about. some one needs to do the dirty job 🙂
but it looks brilliant so far !cheers
tomWhere is the 15khz filter applied (providing of course that you have set it to 15khz)? If BBP is set to output @0db, how sure can we be that between BBP and the line output, noise and crap above 15khz isn’t being added?
[quote author=”Leif”]Documentation is not the most fun part about developing products, but somebody’s gotta do it. 🙂[/quote]
No kidding. I’ve written a couple of helper apps for my automation software. A couple of other people found them useful so I had to write a help file. Takes all the fun out of programming! 😆Don’t for get to fix this one Leif…… 😆
[quote author=”knigget”]Where is the 15khz filter applied (providing of course that you have set it to 15khz)? If BBP is set to output @0db, how sure can we be that between BBP and the line output, noise and crap above 15khz isn’t being added?[/quote]
BBP doesn’t have a diagram on this topic, that being said… I’ll answer your question anyways.
BBP’s low-pass filter is before the L/R clipper. Nothing at all above the low-pass filter is getting added except for an extremely low level high frequency noise that’s part of the dithering algorithm. (we’re talking under -84dB, and up to half is only quantization noise under -90dB)
The BBP clipper is not followed by any protection clipper for digital sample loudness, so it’s possible that now and then you could get a slight clip if your output is exactly at 0.0db. If you go into "Calibration" and set the "L/R Ref Level" to -0.2dB, you should completely remove all chance of not only digital sample values clipping, but TruePeak (4x oversampled) clipping, and "overs" out of the analog output, assuming your output is to spec and has a flat frequency response. (but that’s not too horribly important, it should be darn close anyways)
This assumes that you have NO needed tilt or EQ correction. If you’re using those features too then you’ll have to compensate further as the peak control will get biased based on the inverse of your exciter/transmitter/etc bias.
Check out the output on a spectrograph like MPX Tool, or open a stereo recording of the L/R output in an audio editor that has a spectral view like Adobe Audition. There really isn’t anything up there but slight dither noise, despite the clipper being after the low-pass filter, where it needs to be.
So yeah, final answer, no. No "crap" is getting added at all. The clipper really is that pristine. 🙂
Very true. Because tilt or EQ calibration is often necessary anyway, and the default is -3, I didn’t worry about the occasional hard-clips at 0.0dB — it’s not a good setting anyway 🙂. Perhaps I should limit that control to -0.5dB to make it a non-issue.
Indeed the clipper must be after the low-pass filter, but with a properly band-limited overshoot-compensated clipper such as the ones in Breakaway and Orban processors (but not Omnia), adding another phase linear low-pass filter after the clipper makes no difference, provided the cutoff is at or below the cutoff of the previous low pass filter + clipper.
///Leif
Nice post,The invention is incorporated in an electronic circuit board with a break-away portion and a processor which is programmed to recognize whether the break-away portion is present or broken away. The break-away portion has a physical and electrical arrangement which secures the electronic circuit board against tampering to defeat the break-away operation. The break-away portion with its electrical arrangement functions as a tamper-resistant key.
[quote author=”Roygbiv”]Helo
In Breakaway Block Diagram is bug, low pass filter is not after high pass, but is after multiband procesor, I tested it. I don’t understand this idea, because it is very very important tu put low pass direct input to process only target range of audio signal in muliband compressors and limiters.[/quote]
You’re right about the diagram being wrong.
The low-pass is after for a good reason though. For instance; with the low-pass before the multiband, if you have a top band that only has say 2kHz of bandwidth running through it… it’s likely the high-end will sound too dense because that band’s AGC will be way higher than normal, and the limiter will still see the same height of peaks… so it’ll end up making the treble way too dense. (you have probably heard this in certain other processors).
With the low-pass after the multiband, you get a naturally processed high frequency sound that’s consistent to what the preset (or you own settings, with the hardware box) were designed for… and the sound of the treble won’t change when you adjust the low-pass.
– – –
The bass on the other hand, you do want the high-pass to be right away since there’s wide-band dynamics directly after that, which while… they do have sidechains that avoid very low bass "thud" type problems, it is designed to be effected somewhat. So that’s a nice option to allow people to filter out actual audio that they may be having through with, and also to the benefit of the wideband dynamics within.
[quote author=”Roygbiv”]I think input low pass is very good option, because last band in multiband is high pass filter, so if I want only 20-15000Hz then 15000-21000Hz may reduce gain of this last band and produce not optimal output sound.[/quote]
It’s always possible, but in the grand scheme of things… the vast majority of tracks sound much more natural when the whole signal is processed up until the final peak limiting, and when the low-pass setting doesn’t effect the intended sound of the preset.
Basically I’m just repeating myself tho. 😆
- The forum ‘Breakaway Professional Products – [discontinued]’ is closed to new topics and replies.