... but I have to ask:
When I make music completely 'in the box', using only vsti's and midi, do I have to concern about the quality of my soundcard for converting 24 bit audio into 16 bit or converting 48 kHz samplerate into 44.1 kHz? Everything is done by software, right?
Stupidest question ever on kvr, sorry...
-
- KVRAF
- 1800 posts since 10 Feb, 2007
-
- KVRAF
- 5913 posts since 17 Aug, 2004 from Berlin, Germany
Yes and No 
The soundcard is normally not so important for only ITB because it's only used that you hear something (monitoring). But... a good audio card (same with monitors or a good room treatment) is important to hear all the details you need for a good mix.
The other thing is the latency. Good audio cards with good drivers allowing a lower latency and this is important if you play something with the midi keyboard. If the latency is too high it's hard to play "on time".
So for the converting stuff it's true, all is handled internally by the software. But there are also other things which I would take into account.
The soundcard is normally not so important for only ITB because it's only used that you hear something (monitoring). But... a good audio card (same with monitors or a good room treatment) is important to hear all the details you need for a good mix.
The other thing is the latency. Good audio cards with good drivers allowing a lower latency and this is important if you play something with the midi keyboard. If the latency is too high it's hard to play "on time".
So for the converting stuff it's true, all is handled internally by the software. But there are also other things which I would take into account.
| Links-
- KVRian
- 1055 posts since 3 Jul, 2006
No.manducator wrote:do I have to concern about the quality of my soundcard for converting 24 bit audio into 16 bit or converting 48 kHz samplerate into 44.1 kHz?
-
- KVRAF
- 1586 posts since 7 Jun, 2007
Yup, it's usually all handled by the software.
In Sonar, audio files which don't match the project settings will be converted as you import them into the project, and Cubase does this too. They create a proxy file to use so they don't have to convert 'on the fly' in realtime.
So this is the more likely scenario when using different bit depths and samplerates. Conversion first, create proxy file, use proxy file in playback.
Usually however, you'd set your soundcard to the desired samplerate... I usually work in 48k, an old habit from many years of video editing, but 44.1, 96, 88.2, whatever you prefer in terms of sound vs CPU efficiency.
Your DAW will possibly default to whatever your soundcard is set to, although many, like Cubase, have a project settings menu/submenu where you can set up the project at a different rate to the soundcard. Why one would want to do that still confuses me, but in Cubase the options are 'samplerate clock internal'(defined by Cubase) or 'external'(defined by the soundcard or any wordclock being fed to the soundcard).
So if no audio is going in or out the box, except the audio to your speakers or headphones, it's not a big issue. Higher samplerates will reproduce 'non-linear' synth waveforms better, like synced osc's etc, with lower aliasing, but anything 'normal' or sample-based like guitars, piano, voice or drums will be hard to definitively hear the diffs in quality on playback based on sample-rate alone.
Many samplebanks are provided in 24 bit, 44.1k only, and the output of the sample player eg. Kontakt, would be converted in realtime.
The bit depth should ideally match between soundcard and DAW. If your project is set to 24bit, with the soundcard set to 16bit, you're losing the benefit of the 24bit dynamic range on playback (depends on the style of music, how hot you mix or push the master bus limiter, but a very expressive classical piano piece might benefit from 24 bit playback, the same applies for a very dynamic drum track). The realtime down-conversion, if possible, will probably be handled by the soundcard, although not necessarily in particularly good quality, having to dither in realtime. Again depending on the type of soundcard.
I hope this answers your questions!
I'd rather just match my soundcard and project settings, as this eliminates most conversion worries right from the start
In Sonar, audio files which don't match the project settings will be converted as you import them into the project, and Cubase does this too. They create a proxy file to use so they don't have to convert 'on the fly' in realtime.
So this is the more likely scenario when using different bit depths and samplerates. Conversion first, create proxy file, use proxy file in playback.
Usually however, you'd set your soundcard to the desired samplerate... I usually work in 48k, an old habit from many years of video editing, but 44.1, 96, 88.2, whatever you prefer in terms of sound vs CPU efficiency.
Your DAW will possibly default to whatever your soundcard is set to, although many, like Cubase, have a project settings menu/submenu where you can set up the project at a different rate to the soundcard. Why one would want to do that still confuses me, but in Cubase the options are 'samplerate clock internal'(defined by Cubase) or 'external'(defined by the soundcard or any wordclock being fed to the soundcard).
So if no audio is going in or out the box, except the audio to your speakers or headphones, it's not a big issue. Higher samplerates will reproduce 'non-linear' synth waveforms better, like synced osc's etc, with lower aliasing, but anything 'normal' or sample-based like guitars, piano, voice or drums will be hard to definitively hear the diffs in quality on playback based on sample-rate alone.
Many samplebanks are provided in 24 bit, 44.1k only, and the output of the sample player eg. Kontakt, would be converted in realtime.
The bit depth should ideally match between soundcard and DAW. If your project is set to 24bit, with the soundcard set to 16bit, you're losing the benefit of the 24bit dynamic range on playback (depends on the style of music, how hot you mix or push the master bus limiter, but a very expressive classical piano piece might benefit from 24 bit playback, the same applies for a very dynamic drum track). The realtime down-conversion, if possible, will probably be handled by the soundcard, although not necessarily in particularly good quality, having to dither in realtime. Again depending on the type of soundcard.
I hope this answers your questions!
I'd rather just match my soundcard and project settings, as this eliminates most conversion worries right from the start
-
- KVRian
- 1055 posts since 3 Jul, 2006
Do you have any material to support this statement? As far as I know, the soundcard does no digital processing of the audio (it only converts numbers into analog audio, as fast as possible).xalama qo wrote:The realtime down-conversion, if possible, will probably be handled by the soundcard, although not necessarily in particularly good quality, having to dither in realtime. Again depending on the type of soundcard.
-
- KVRAF
- Topic Starter
- 1800 posts since 10 Feb, 2007
Ok, thank you guys!
I have a hercules dj console MK2. It has asio drivers and is good enough for what I want. The only input µi have is a midi controller, so no audio transfer out of the box involved.
I have a hercules dj console MK2. It has asio drivers and is good enough for what I want. The only input µi have is a midi controller, so no audio transfer out of the box involved.
-
- KVRAF
- 5913 posts since 17 Aug, 2004 from Berlin, Germany
Yes.manducator wrote:Ok, thank you guys!
I have a hercules dj console MK2. It has asio drivers and is good enough for what I want. The only input µi have is a midi controller, so no audio transfer out of the box involved.
Perhaps people which hear the grass grow, will hear a difference between the different converting algorithm's of the DAWs. But IMO this is insignificant. Anyway, the soundcard is not involved.
| Links-
- KVRAF
- 1586 posts since 7 Jun, 2007
Hi, no I don't have any reference for that, it was an opinion, hence the 'if possible' and 'probably' disclaimers. I should have said IMO, or 'in my understanding', sorry.jackoo wrote:Do you have any material to support this statement? As far as I know, the soundcard does no digital processing of the audio (it only converts numbers into analog audio, as fast as possible).xalama qo wrote:The realtime down-conversion, if possible, will probably be handled by the soundcard, although not necessarily in particularly good quality, having to dither in realtime. Again depending on the type of soundcard.
It is very possible that I am completely wrong. Wouldn't be the first time!
Always open to learning too. You seem pretty certain in your reply, would you care to elaborate?
-
- KVRAF
- 1586 posts since 7 Jun, 2007
Just to clarify, my understanding is that if the DAW is outputting a 24 bit audio stream, but the soundcard is set to 16 bits, then it will just say no-go. Thinking of the typical scenario of trying to play a 24bit file in Windows Media Player, no-go. That's *probably* more of a codec issue.
A soundcard *I'd imagine* would pop up a dialogue box saying please change your settings as I am unable to play a 24bit file while I'm set to 16bits, or *most likely* do so at the start of the project setup.
So, hypothetically speaking, and *IIRC* some interfaces like RME have DSP dithering onboard, *I believe*, so, if I did not just dream that, which is also possible,( i tend to have quite vivid dreams), then the soundcard would down-convert from the DAWs 24bit output to a 16bit stream, *if the soundcard is set to 16bits*.
The more pertinent question is, if i did dream that, why was i dreaming of audio interfaces? And do i need to get a life?
But I'm probably way off here cos I'm thinking in terms of master bus output to whatever monitoring system.
On a track by track level, bit depth and samplerate are converted per audio clip within the DAW, and *in my experience* usually by the proxy file method.
Virtual synths output will conform to the DAW's project setting.
At least that's how I understand it. If I'm wrong please educate me, I'm all for furthering my education! (no sarcasm intended at all, no, really).
A soundcard *I'd imagine* would pop up a dialogue box saying please change your settings as I am unable to play a 24bit file while I'm set to 16bits, or *most likely* do so at the start of the project setup.
So, hypothetically speaking, and *IIRC* some interfaces like RME have DSP dithering onboard, *I believe*, so, if I did not just dream that, which is also possible,( i tend to have quite vivid dreams), then the soundcard would down-convert from the DAWs 24bit output to a 16bit stream, *if the soundcard is set to 16bits*.
The more pertinent question is, if i did dream that, why was i dreaming of audio interfaces? And do i need to get a life?
But I'm probably way off here cos I'm thinking in terms of master bus output to whatever monitoring system.
On a track by track level, bit depth and samplerate are converted per audio clip within the DAW, and *in my experience* usually by the proxy file method.
Virtual synths output will conform to the DAW's project setting.
At least that's how I understand it. If I'm wrong please educate me, I'm all for furthering my education! (no sarcasm intended at all, no, really).
-
- KVRian
- 1055 posts since 3 Jul, 2006
First of all, I'd like to apologize if my tone didn't sound very friendly. I'm here to learn also, and I really didn't mean to be offensive in any way. It's the first time I read about realtime onboard dithering, because I didn't know it existed. I'm sorry again if I offended you.
The typical hardware soundcard that I know about (leaving the line-in and mic aside) is just multichannel digital to analog converters, analog filters and analog amplifiers.
Professional soundcards offer direct hardware support for multiple input and output sound channels, as well as higher sampling rates and offer better fidelity as compared to 'regular' soundcards. From the user's perspective, the expensive soundcard driver is simply able to forward more data (more bits at higher rates) to the DAC chips, without using more CPU cycles.
I've also heard that more expensive cards include DSP chips that include multi-channel mixers, multi-band EQ and a dynamics processor on each audio channel. These are functions that would have ordinary been implemented in software via the driver of the soundcard.
However in case of dithering and resampling, I've always thought that this is done in software mode, mostly before the soundcard's driver (let's say at DAW level), and sometimes implemented as a function in the driver (still in software, at the ASIO driver level). Of course I could be wrong, and my knowledge could be out of date. If so, I will stand corrected.
Cheers!
The typical hardware soundcard that I know about (leaving the line-in and mic aside) is just multichannel digital to analog converters, analog filters and analog amplifiers.
Professional soundcards offer direct hardware support for multiple input and output sound channels, as well as higher sampling rates and offer better fidelity as compared to 'regular' soundcards. From the user's perspective, the expensive soundcard driver is simply able to forward more data (more bits at higher rates) to the DAC chips, without using more CPU cycles.
I've also heard that more expensive cards include DSP chips that include multi-channel mixers, multi-band EQ and a dynamics processor on each audio channel. These are functions that would have ordinary been implemented in software via the driver of the soundcard.
However in case of dithering and resampling, I've always thought that this is done in software mode, mostly before the soundcard's driver (let's say at DAW level), and sometimes implemented as a function in the driver (still in software, at the ASIO driver level). Of course I could be wrong, and my knowledge could be out of date. If so, I will stand corrected.
Cheers!
-
- KVRAF
- 1586 posts since 7 Jun, 2007
Cool, no worries man. I'm no expert on anything even remotely code-based, so my description of the soundcard getting involved with any down-conversion was purely within the context of a DSP driven dithering capability within the soundcard itself...which may or may not even exist....I dunno. All hypothetical.
I just seem to remember it being mentioned in connection with some RME soundcards at some point, and my ears pricked up cos it seemed to be an unusual feature. Maybe it was Prism Sound or some other ridiculously high end DSP interface. Or an embarrassingly geeky dream...
I'd compare it to the early days of HDV when you had the option of capturing the HD video as SD if you wanted to, and as far as I understand, the down-conversion happened within the camera or tape deck before it hit the firewire cable for transfer to the compuer. The computer CPU never got involved with the down conversion on the fly...as far as I understand... I'm pretty sure if it did, my CPU back then would've given me the finger!
Time for a sig:
Disclaimer, I know nothing, I just use the stuff! hehe!
Maybe someone who's taken the time to Google all this will come and prove us both wrong!
I just seem to remember it being mentioned in connection with some RME soundcards at some point, and my ears pricked up cos it seemed to be an unusual feature. Maybe it was Prism Sound or some other ridiculously high end DSP interface. Or an embarrassingly geeky dream...
I'd compare it to the early days of HDV when you had the option of capturing the HD video as SD if you wanted to, and as far as I understand, the down-conversion happened within the camera or tape deck before it hit the firewire cable for transfer to the compuer. The computer CPU never got involved with the down conversion on the fly...as far as I understand... I'm pretty sure if it did, my CPU back then would've given me the finger!
Time for a sig:
Disclaimer, I know nothing, I just use the stuff! hehe!
Maybe someone who's taken the time to Google all this will come and prove us both wrong!
-
- KVRAF
- 1586 posts since 7 Jun, 2007
K, it's official, i need a life ...just Googled Prism Sound website, they have an SNS noise shaping technology:
"Since Orpheus also includes the full suite of the famous Prism Sound 'SNS' noise shapers, you can also reduce to 16-bits at mastering-house quality."
Not a 'told you so', but it looks like it does exist for users with budgets as large as Sting's!
http://www.prismsound.com/music_recordi ... s_home.php
...just Googled Prism Sound website, they have an SNS noise shaping technology: "Since Orpheus also includes the full suite of the famous Prism Sound 'SNS' noise shapers, you can also reduce to 16-bits at mastering-house quality."
Not a 'told you so', but it looks like it does exist for users with budgets as large as Sting's!
http://www.prismsound.com/music_recordi ... s_home.php