44.1 vs 48 vs 88.2 vs 96

@linuxdsp:

Thx for your response.

I totally got messed up there for a sec. :slight_smile: sorry 'bout that. Yes, I do understand what the dithering is for.

Still just wondering (for knowledge not that I would do a crazy thing like this), If the audio interface would work at another depth than the one I want to track to, would enabling the dithering in JACK dither that audio in real time for the recording? And with good quality? If not then what precisely is the dithering inside JACK for?

Thanks.

Generally (as I’m sure you know) you would always want to track at the highest resolution (bit depth) you can, and I can think of almost no benefits to dithering audio while recording (except for possibly saving storage space, but in these days of huge disk capacity and system memory that’s almost never an issue anymore - for pro audio anyway). As to what jack is capable of in that respect, I think its a question for Paul or the other jack developers (I’d have to look at the code anyway to see what dithering algorithm it uses, and I have no practical experience of ever using jack to dither audio at all - although it might be interesting to try out the options with baudline or similar audio tools and see what the effect is…)

@joegiampoli: there is no reason to use dithering in your scenario. as linuxdsp noted, its only use is when reducing the bit depth (e.g. from 24 to 16 bits). it exists inside JACK so that people working with a 16 bit interface (there are still a lot of them) can optionally turn it on for the final conversion of audio into 16 bit form as it leaves the JACK ALSA backend. it has no role to play within JACK clients, where all audio is and remains 32 bit floating point (though i suppose something that (foolishly) converted to and from 64 bit floating point internally might want to use some form of dithering as well). Clients that can export to a disk file (like Ardour) can also use dithering if asked to write a file with <24 bits of resolution, but they do not do this via JACK.

Ok, thank you so much. I just wanted to know why there would be a dithering option in JACK.

Well anyway, I’m doing some TESTING now to see what results I get tracking @48KHz (Fast track Pro native internal sample rate) obviously in 24 bit mode, mixing at same values, then exporting (upsampling) up to 96 KHz 24bit for mastering and then dither it down to 44.1KHz 16 bit (a la Bob Katz).

I’ll let you guys know if the results are really worth it. I’ve just seen too many posts on other places where they swear by this method, so I gotta give it a go to see for myself.

Cheers!

hey hello
sorry about my wild english
this is not the truth …only my experience

well this thing is very complex ( mathematically ) but in my experience , it depends of the kind of stuff you have… so if you have an
audio-interface that supports 44.1, 48 ,88.2 ,96 and 192 sample rate you must need to have a pair of speakers(monitors) that supports those sample rate …but the thing about is :

between 44.1 and 48 there´s not difference really (what really matters is :if you are working at 24 bit or 16 bit…"".24 bit is much better but when you finish a song you need to render or export to 16 bit, many people use dither to avoid “loss” but some times it is not needed you have to listen and decide what is better)

… 44.1 sample rate is for any kind of music-recording audio(standard) and 48 Sample rate is supported by video-editing-recording(cameras diginal-analog or video interfaces and soft. ) for this two options you only need a common pair of monitors … i mean analog(with rca inputs or input 1/2) it will sound good at 46 and 48 sample rate so although you increase your “interface audio sample rate” to 88.2, 96, and 192 you wont hear difference …(IF you have speakers such i mentioned before(analog) you will increase the work on your processor (computer) and nothing will become better)…

well when you use 88.2 ,96 or 192 sample rate you before have to know this:

44.1 *2 = 88.2 so this is for audio(music)
48 * 2 = 96 …(videos,films etc)
96 * 2 = 192…( digital optical audio environment) (between 88.2 an 96 Khz there´s not difference really , it depends of if audio is for music or video project )

so if you have an -interface audio- with “digital /optical out” and also you have speakers with “digital/optical input” and both support 88.2, 96 and 192 sample rate you really will hear the difference between “44.1, 48 sample rate playback” and 88.2, 96, 192 sample rate playback" (they have to be connected via tosh link optical cable""""" if you connect them via rca input or 1/2 input you wont hear the difference im talking) but it doesn’t matter really because ''not everyone has/ digital/optical audio-system/ …thus only
who has /optical/digital audio - system /will be able to hear the details of the sound at 88.2 or higher.

but the advantage of working at 88.2, 96, 196 sample rate for example during recording, you catch some details that you didn’t catch at 44.1 or 48 sample rate and when playback each sound has more detail.

the problem is that when you Finnish a track you can´t export it to 196 or 96 sample at 24 bit because there would be a lot of “loss” audio due to not everyone has optical audio system as i wrote before… normally we all export tracks to 48(video), 44.1(audio) sample rate and 16/24 bit so a kind of rule to avoid loss audio is/are:

for audio you work/record at 88.2Khz and 24 bit then you export tracks to 44.1 Khz and 16/24 bit(only few optical monitors supports 88.2Khz)

for audio-video you work/record at 96Khz and 24 bit then you export tracks to 48Khz and 16 bit/24 bit

about 192Khz and 24 bit can be used at sound installations or digital theaters for HD films with HiFi optical Audio(very expensive technology )) rare¡
also if want to record birds (or other animals) 192 khz is the best sample rate to do that ,96Khz as well. also if you want to record some inaudible/low.hi noises.textures or freqs to make experimental music works ok,96Khz as well.

also
for video you can work/record at 192Khz and 24 bit then you export tracks to 96Khz and 24 bit and then to 48Khz and 16/24 bit (not common//only me)

(But If You Really Want To Avoid Any Kind Of " Loss Audio"… Record And Playback Like This …Music 44.1Khz,24/16 Bit And Video-Audio 48 Khz, 24/16 Bit… Better 24 Bit For Recording And 16 Bit For Exporting

well its all i can say about … bye bye