Forum Replies Created
-
Posts
-
Hi MrKlorox,
Headphone listening fatigue — that makes perfect sense, actually!
But, then it needs to go in the speaker controller (speaker out, monitor out etc), not inside the main processor.
I'm already planning to add a Mono switch to the speaker controller (customer request for another product based on the same framework) so that'd be a good time to put a Mono Bass switch into the speaker controller, too.
Best,
///LeifOops.. I believe this is actually a BreakawayOne problem.
For MP2/MP3/AAC streams, metadata is inserted by the server into the normal stream at regular intervals (the metadata interval).
For anything that uses OGG bitstream format (FLAC, Vorbis and Opus), metadata must be inserted into the encoder.FLAC is integrated into BaOne (as of the 3.18 versions), so I handle that part.
Vorbis and Opus are not — they're still executables called through stdin/stdout (and only show up in BaOne if you have those executables in your path somewhere), and there's no way to insert metadata when running this way.
So, in order to make this possible, I'll have to integrate Opus into BreakawayOne directly.That's of course possible, but I can't say when I'll have the time to do it — it isn't commonly used for streaming, nowhere near as common as aacPlus, which sounds better at a given bit rate.
Opus strength is that it's reasonable quality at low bit rates, and incredibly scalability and error resiliency, low latency. So, it's great for a contribution feed when you have a limited bit rate available.. but normal streaming to listeners (where you'd need metadata) is not such a common use case.
That being as it is, it's not a high priority compared to all the other urgent things on my to-do list. But, I'll get to it eventually.
Best,
///LeifHD doesn't have multipath problems like analog FM, what's the reason for wanting to mix the bass to mono if it's not already mono in the recording?
This is actually part of the Multipath Mitigation feature now, you'll find the button under FM Processor / Misc / I/O / Main.
///Leif
Hi Timmy,
16-bit to 16-bit needs no dithering unless you're attenuating (at which point it's no longer 16 to 16). It should not matter, and FB2K is probably smart enough not to do anything differently no matter what you select.
Hey, there's a phenomenal video by Chris "Monty" Montgomery explaining sample theory and dithering in a really riveting and cool way. His personality helps make it great too — I've watched it like three times, he gets every single bit right. I recommend you watch it if you're interested.
https://www.youtube.com/watch?v=cIQ9IXSUzuM&t=1040s
Best,
///LeifHi Paul,
If you come up with a good, completely different name for it (something that doesn't make people pick it just because they happen to be a radio station playing contemporary hits, which would result in making a bad situation worse by overcompressing already overcompressed material) I'll port it over for you. 🙂
///Leif
Hi Timmy,
It's true that music is generally recorded in 16 bit resolution, and if you ask me, this is plenty for a recorded piece of content.
The problem with music isn't the resolution, it's the fact that they beat the audio to death and master it as if they had another three bits of resolution on top. (See: Loudness War)
The noise floor was not a problem in properly mastered late 80s 16-bit recordings.
The most laughable thing is that the 24-bit version of Imagine Dragons latest album (yes, there's a 24-bit release) has exactly the same amount of clipping as the 16-bit version.. they sound identical.So, why did I bother with making BreakawayOne fully 24-bit?
Because, if the input path is 16 bits, you'd only be able to preserve the 16-bit resolution of the original content by feeding it into BreakawayOne completely without attenuation … leaving you no headroom for any other elements.
The standard level for digital mixers is probably somewhere between -12 and -20. That's 2 – 3 bits.. if the input path was 16 bits, you'd be left with 13 or 14 bits. Sure, it'll still sound fine, but the noise floor will be too close for comfort. By extending the WDM I/O resolution, we moved the noise floor 48 dB's down.. so no we can really truly forget about it and worry about things that matter, like making sure we never ask Windows to do sample rate conversion by running the input pipeline at 48 kHz while playing 44.1 content. 🙂
BreakawayOne has been 24-bit (well, 32-bit actually — 24-bit audio + 8 bit headroom) internally since the very beginning, and so have the ASIO and Livewire I/O sections.
The streaming sections and WDM I/O was 16-bit though (in 3.16.84).. and in fact I wasn't even dithering the output properly. My bad. So, I took care of the few missing pieces, to achieve a truly 24-bit chain.
In contrast, the entire previous generation of Breakaway products (Breakaway Broadcast, Breakaway Live, Breakaway Audio Enhancer etc) was 16-bit throughout (16-bit audio + 16-bit headroom internally).
Now to actually answer your question:
CDs are 16-bit! There is absolutely no benefit to adding eight zeroes for every sample. The player is perfectly capable of adding those eight zeroes in real-time as you play your 16-bit files — there is absolutely no benefit to adding the zeroes to the file itself. Believe me, adding zeroes is very easy. Look, I can do it right here: 00000000000000000000000000000000000Best regards,
///LeifFixed in 3.19.19.
https://www.dropbox.com/sh/1anat3m9eaz6hu4/AACbO6619-lrPU82fUuZXuLfa?dl=0
Changelog is there too. Will update the public web site later, no sense in doing it for every little update. 🙂
Best,
///LeifHi Bluebird,
I remember there being a reason why I did this during development, but you're right — this isn't a good thing, I don't know what I was thinking. I'll look into it again and find a way.
Best,
///LeifHi Mike,
BreakawayOne should be smoothness itself!
What processing core (HD, FM etc) and which output are you using?
If you are using FM Output, without De-emphasis on, that would explain the extremely bright sound you would be hearing.The best output to use for auditioning is the Common Monitor Output. That way you can switch between hearing different processing cores from within the program.
The Input level meter is an industry standard ITU BS.1770 loudness meter. Any level is fine (although most presets are designed to be in the green-yellow-orange section). If you overload the input, the entire meter will flash red. You can increase "Input Gain" without overloading the input.
Red bars in the multiband means peak limiting — this is normal operation.
On the input oscilloscope, you don't want to see it being pegged to the rails (square wave), but if you are playing modern music, this will pretty much always be the case. At least the input oscilloscope will let you see where the distortion is coming from, as well as to ensure you're not adding to the problem by overloading your input.
BaOne can process at either 44.1 or 48 kHz, and the FM MPX output is at 176.4 or 192 kHz.
///Leif
It's a whole new composite back-end actually, completely separate code path :). But yes, it uses more CPU, has more headroom, and less distortion.
Dear Jess,
Ram cluttering is not a term I am familiar with. There are two types of memory allocation issues that I know of:
– Memory Fragmentation. Imagine filling up the available RAM with small allocations, and then deallocating half of them at random. Even though you then have free memory, you wouldn't be able to allocate a big block because there would not be enough contiguous space free — only lots of smaller pieces.
– Memory Leaks. Continuously allocating things and then forgetting to deallocate them, meaning more and more wasted memory.I've ran BaOne FM+HD for months and months on a system with 2 GB ram and have never noticed any memory leaks over time. Memory fragmentation is not an issue as the large allocations happen on startup, not as it's running. The dynamic allocations/deallocations as it's running are all pretty small.
That said, even if there were memory leaks in BaOne, you wouldn't have to restart the whole machine to get rid of them — you'd only have to restart BaOne. Everything a process has allocated goes away when the process is closed.
16 gigabytes of RAM is a huge amount of overkill. No way will RAM be an issue. 🙂
Best regards,
Leif ClaessonI have a new free tool that can do exactly this. It requires the management package in BaOne, other than that it's completely free.
https://www.dropbox.com/sh/1lwusvkuion6322/AAD7IYmujzOAolfQU9fjBT6La?dl=0———–
NfAudioTool
———–A robust and reliable tool to connect to any NF-based unit (whether Omnia.9, Linear Acoustic AERO, or Breakaway One) the exact same way NfRemote does, and stream audio.
The received audio can be sent to local sound cards, sent over livewire, and written to files in FLAC, Wave or MP3 format.
Thus, it is a convenient way to do mandatory logging, pull audio from an Omnia.9 HD processor and feed the Exporter (without using up a physical output), feed a pair of speakers in the lobby, etc.///Leif
