Windows XP PRO or XP 64 Bit for recording

[On Time Recording]

I will be building a new computer for my recording setup. Would it be better to build a 64 bit computer so I can run Windows XP 64 bit edition or should I build a 32 bit PC and run Windows XP PRO 32 bit OS ?? I want the better choice for my recording setup. Any sugestions?

[scottdemarko]

I am pretty sure that 64bit isn't compatible with Pro Tools, at least it wasn't in previous versions. I'll have to have Tech1 chime in this one, but I am pretty sure you want to stick with XP Pro for now.... perhaps later on (like a year from now look at Vista)

[tech1]

For Pro Tools: http://www.digidesign.com/index.cfm?navid=54&langid=100&categoryid=35&itemid=24211

If you're considering another system, make sure the software and interface you're using is supported in Windows XP Professional x64 (or Vista 64). Pro x64 definitely has a much more limited scope of compatibility. Also, ALL of your computer hardware (motherboard, video card, etc.) needs to have drivers that are supported in XP Pro x64. Again, check the manufacturers' websites for downloadable drivers.

As far as it being "better" or not, that is a totally subjective question. Audio files recorded into the system will still be 24-bit (due to interface limitations); it is in the mixing stage where 64-bit engines are actually used. The decision needs to made on whether the software/hardware you plan to use is supported, whether you will meet the minimum requirements, and whether you feel that 64-bit processing will allow you to make better music.

[Justin]

I'd hold off on 64-bit.

I feel that most vendors never finished their XP 64-bit support. They all had beta drivers or software out, then ceased development as soon as Vista was on the horizon.

Personally, I'd use XP Home 32-bit right now. In 6-months to a year, when it's really polished and ready for audio production, I'd upgrade to Vista Home Premium 64-bit, but only if all your hardware and software has 64-bit support. (otherwise go 32-bit)

[On Time Recording]

Well the reason I was going to build a computer for my recording setup based on a XP 64 bit was because, with windows Xp 64 bit, you can run more ram & CPU GHZ for bigger audio files, such as 24 bit 192k recording......thats why.

I want to record in 24 bit 192k

I noticed that the music recording studios that are digital are going 24-192

[dpd]

Why do you feel you need to record at 192 Khz sample rate? There is NOTHING audible in the top two octaves being sampled (24-48 Khz and 48-96 Khz). So, all you do is waste disc space, buy expensive gear and put a HUGE load on your system (e.g. CPU, RAM, plugins, etc).

Lots of good papers on the lack of benefits of 192 Khz sampling. You can make some killer recordings at 44.1 Khz but a lot of the marketeers are trying to convince you otherwise.

24 bit, on the other hand, is very useful

[OSSforever]

The sample rate is not the frequency of the sound wave. It is the frequency of a sample of the audio wave being recorded.
For instance, a 10kHz sine wave is audible, and can be recorded with a piece of equipment. Equipment A is capable of 16bit/44.1kHz recording (cd quality). This means that over the course of a second's time, 4.41 (quantity) 16-bit samples of the 10kHz sine wave is recorded as data. If Equipment B records at 24bit/192khz, the same 10kHz sine wave is sampled 19.2 times per second in a 24-bit long sample.
So, if you've never heard the difference between recording something at 24bit/96kHz and at 24bit/44.1kHz, I welcome you to try. Your ears should be good enough to discern the difference.

[DAS]

So the premise of your point is that the 4.41 samples per cycle produces better reproduction of your 10 kHz audio wave than the 19.2 samples per cycle?

If so...well...I don't know how else to say it, but that's just dead wrong. I realize it's intuitive, and in a lot of things in the world more is better, but digital audio just doesn't work like that. One only needs two samples per cycle (no more) to be able to reproduce a sine wave with all amplitude and phase information completely in tact. This is one of the basic ideas that provides the foundation for everything in digital audio.

[OSSforever]

So, how is it possible that Sweetwater can sell Apogee D/A converters that cost THOUSANDS, if all we need is two samples per cycle?


Do we all need to retake Physics and Computer Science 101? Physics tells us how waveforms work in the physical world. Computer Science tells us how we can copy that physical event and then replay it from 1's and 0's arranged on a disk.

I'm not here to be argumentative, but I think the last comment misunderstood my explanation that taking a gajillion digital samples of a physical event over the course of a second is much more accurate than five or six.

Resources: http://www.saecollege.de/reference_material/pages/Recorders.htm

[OSSforever]

A sample is a piece of data. 16 bits long, or 24 bits long.
A rate is how often we do something.

I would like to get a 24 bit long piece of information from what I'm hearing right now, the project manager's telephone conversation in the office next to mine. In order to get the most accurate representation of how his voice sounds, I have to get more than one of these samples every second(24bit/1Hz). To remember what his voice actually sounds like, I will need more information than that. The more information the better. If I take a 24 bit sample of his voice 96,000 times over the course of a second's time, I have a much better understanding of what myriad of frequencies his voice box produces when saying ng's and sss's and th's, and how whiny he can sound when something isn't going his way.

Understanding Digital / Analog Conversion

[tech1]

Very true. However, you have to take into account the upper limits of human hearing: 20 kHz. Most people can't really hear past 18k. To build on what DAS pinted out, a sample rate need only be twice the highest frequency present in order to accurately reproduce it. This theoretically puts your practical sample rate at 40 kHz to accomodate the entire realm of human hearing.

However, due to real-worls problems with aliasing, the "stuff" above 20 kHz needs to be filtered out to avoid aliasing (foldover). Hence, 44.1kHz, which allows for 2.2 kHz of "dead air" to make room for a filter. Better D/A converters do this more transparently, without any noticeable effects to the resultant recorded audio.

[DAS]

OSSForever - I mean this with all due respect. Seriously, I think you need to take a step back and do some research/reading on these issues. Many people have been confused in the same ways you appear to be. I've seen it many times before. You need to understand this subject better.

You may want to start by looking through some of the terms that have been defined in our Word For the Day archives. You might start with Nyquist Frequency.

http://www.sweetwater.com/insync/word.php?find=NyquistFrequency

There are many others throughout the archives that could be helpful.

To be clear...the number of samples taken, beyond two, doesn't (in theory) matter. More samples beyond that just makes more data, the sound doesn't change at all. Look, I know this sounds completely bogus, but when you really understand how digital audio and anti-imaging filters work it will all make sense.

If you want to read up on the subject and become an authority consider this resource. It's a good book written in plain English.

http://www.sweetwater.com/store/detail/NikaBook/

We can work through many of the issues on this forum if you don't want to read the book. I just offer it as a one-stop way to get up to speed.

[DAS]

So, how is it possible that Sweetwater can sell Apogee D/A converters that cost THOUSANDS, if all we need is two samples per cycle?


The quality of the converter will absolutely have an impact on the resulting sound quality. There is much more to it than how many samples and bits. The analog stages, clock, filters, power supply, and the converter circuitry itself all help determine the outcome.

[OSSforever]

I'm amazed at how many people are mistaken at what the difference between a digital recording and an analogue recording is.

Sweetwater PC Technical Support specialists are apparently also clueless as to why the industry keeps moving to higher digital samplerates, because apparently, we don't need anything higher than 40kHz.

I'm posting this so hopefully everyone active in this thread learns something about the way it really works. Next time a guy from Sweetwater tries to sell me a 24/192 Pro Tools HD system, he better be able to tell me the real benefit of sampling at 192kHz.

First, read this page. You will have to scroll down to the digital section, as the top part of the page deals with analog recording which doesn't concern this discussion.
http://www.saecollege.de/reference_material/pages/Recorders.htm
A reminder, Hertz is the guy who decided that we need a unit to count how many times something happens in given time period. The standard he set was the Hertz. Hertz literally means Times per Second, or x/sec, where x is an integer greater than zero. If something revolves at 60Hz, it creates a sine wave that reaches my ear. Alternating Current, as implemented by Tesla, oscillates a positive and then negative flow of electrons at 60Hz. When we put our ear up to a power line pole, we hear a low hum. That's because the 60Hz that Tesla picked is in a human's audible range. This is the hum that we get whenever we fail to properly ground our audio recording equipment. Grounding equipment to the earth (earth ground) is a way to remove this 60Hz hum from becoming audible when we pass analogue signals through amplifiers.


Second, take a look at some things Sweetwater is selling:

Marketed as the crazy technology that makes SACD's, the http://www.sweetwater.com/store/detail/MR1/ has the ability to record 1-bit word length samples 2,800,000 times every second, or 2.8MHz

A different rackmount A/D converter that is very popular for quality and service is the http://www.sweetwater.com/store/detail/Rosetta200/ . It has the ability to record 24-bit word length samples 192,000 times every second, or 192kHz.

Another handheld recording device that is capable of recording 24-bit word length samples at 96,000 times every second is the http://www.sweetwater.com/store/detail/MicroTrack/


To compare the total amount of bits of data that is being recorded per second we can simply multiply the sample length by the sample frequency.
The handheld Korg unit can record 1 x 2,800,000 bits every second, or 2,800,000 bits per second of sample data.
The rackmount Apogee unit can record 24 x 192,000 bits every second, or 4,608,000 bits per second of sample data.
The handheld M-Audio unit can record 24 x 96,000 bits every second, or 2,304,000 bits per second of sample data.

For an apples to apples comparison from purely a sampling capability perspective, the M-Audio unit is a better value because it can record 2,304,000 bits per second for $300, while the Korg unit charges an extra $400 to only record 496,000 more bits per second.



Are there any open positions at Sweetwater?

[DAS]

Are you now asking why these higher sample rate units exist? It's hard for me to understand where you are going and what you really want to know. I'm trying to help.

I checked out the recording article. It has some good, basic info in it, and then has some stuff that is dead wrong and/or misleading. Feel free not to believe me and do additional research on your own. Or...we can try to break the whole subject down right here point by point. But...I think there is a lot of ground to cover. I can't really write a whole book here on the forum, but we can tackle specific issues.

What is the way it "really works" you are referring to? Maybe we can start there.

[OSSforever]

As a senior moderator, is it not your job to aid in Sweetwater's marketing efforts?

What kind of equipment does Sweetwater sell? Pro Audio.

A/D converters do one thing. They take audio signals and convert them into a format that can be put on a tape or disk that can be stored, edited, mixed with other digital signals, and then played back by a D/A converter.

The book you referred to is published by Sweetwater Press. I'm sure if I "wrap my head around it" I'll be able to draw the same conclusions you have, and understand exactly what you're trying to tell me. It seems like this book and your reference to it is simply a marketing tool.

The reference I am using is from the School of Audio Engineers, an institution formed in the early 70's in the Netherlands. The Dutch don't sell knowledge, they give it to those who seek it. Americans like to sell "knowledge" for profit. It's all about the American Dream.

I use linux to record audio. I bought an M-Audio Delta 66, a Soundcraft Spirit M12, a PreSonus Central Station, some good monitors and mics and cords, all from Sweetwater. The card was a good choice because of the linux support. When I use the ALSA Jack Server, I can choose from several different recording sample rates. When I record my acoustic guitar at 24bit/22050Hz, I hear lots of noise, or aliasing, intermingled with the high frequencies of the recording. When I record at 24bit/44100Hz, I hear the same amount of noise as when I play a CD through the computer and monitors. When I record at 24bit/96000Hz, I do not hear any noise. The best recordings I've ever heard are those that I've made with this equipment. If I get a chance to compare 24bit/192000Hz to 24bit/96000Hz on someone elses equipment, I will probably be able to hear the difference in sound quality, and would then understand the noise (aliasing) that's present in my 24bit/96kHz recordings, but I currently can not.

Using Linux allows me to record and process the digital audio at about 2 milliseconds latency, using a specialized pre-emptive kernel patch on the Athlon 64 3000+ processor using the Debian based 64Studio distribution of linux. I don't have to pay the pro-audio software tax.

Don't get me wrong, Sweetwater sells great hardware. However, the proprietary software that is sold by them is what makes them the real money, and I'd rather buy tangible things that can't be pirated.

[DAS]

As a senior moderator, is it not your job to aid in Sweetwater's marketing efforts?

Not really, but that's beside the point anyway.



The book you referred to is published by Sweetwater Press. I'm sure if I "wrap my head around it" I'll be able to draw the same conclusions you have, and understand exactly what you're trying to tell me. It seems like this book and your reference to it is simply a marketing tool.

Things aren't always as they seem. I assure you the intentions behind the book and my intentions here are simply to educate. Nothing more, and certainly nothing sinister. But, as I wrote before...feel free not to believe a thing I say, but PLEASE do some more research and reading on your own.



The reference I am using is from the School of Audio Engineers, an institution formed in the early 70's in the Netherlands. The Dutch don't sell knowledge, they give it to those who seek it. Americans like to sell "knowledge" for profit. It's all about the American Dream.

I am offering to give it free as well. Will you take it? That reference page has errors and things that are misleading. Feel free to research it all on your own.



I use linux to record audio. I bought an M-Audio Delta 66, a Soundcraft Spirit M12, a PreSonus Central Station, some good monitors and mics and cords, all from Sweetwater. The card was a good choice because of the linux support. When I use the ALSA Jack Server, I can choose from several different recording sample rates. When I record my acoustic guitar at 24bit/22050Hz, I hear lots of noise, or aliasing, intermingled with the high frequencies of the recording. When I record at 24bit/44100Hz, I hear the same amount of noise as when I play a CD through the computer and monitors. When I record at 24bit/96000Hz, I do not hear any noise. The best recordings I've ever heard are those that I've made with this equipment. If I get a chance to compare 24bit/192000Hz to 24bit/96000Hz on someone elses equipment, I will probably be able to hear the difference in sound quality, and would then understand the noise (aliasing) that's present in my 24bit/96kHz recordings, but I currently can not.

Cool. By all means record at rates that sound the best to you. There are things that can cause this, it isn't necessarily the rates in and of themselves (this discussion has taken place many times on this forum over the years so feel free to search for it). Different filter sets (for each sample rate) can be a big culprit, especially on lower cost converters.



Don't get me wrong, Sweetwater sells great hardware. However, the proprietary software that is sold by them is what makes them the real money, and I'd rather buy tangible things that can't be pirated.

I don't follow this. ??


OSSforever - I didn't intend for this to turn into a battle, and I for one don't have time for one. I am not telling you to buy Pro Tools, or Apogee, or any other specific products. I have only commented on statements you've made that expose some areas in which I think I can be of help. I have a lot of expertise in this field. I am happy to give it away for free if you want it. If not, that's okay too. I do encourage you to do some more reading if you want to learn how digital audio really works. Between now and then thanks for being a customer. We'll always try to help you any way we can.

[dpd]

OSSforever - here is a link to an excellent article (http://www.lavryengineering.com/documents/Sampling_Theory.pdf)about sampling theory

[OSSforever]

Thank you for the information.
It now makes sense that all of that extra information recorded by 192kHz machines is unnecessary.

The information that I was citing prior to now was obtained from a paper written when digital sampling was in its infancy. The assumption was that more information translates into more accuracy in the resolution of the waveform.

I kind of went off on a wild tangent about marketing and how Sweetwater sells equipment, and I extend my apologies to DAS and everyone else posting in this thread.

It still, however, does not make sense that Sweetwater sells equipment with those capabilities far exceeding what the human ear can hear yet has forum conversations on exactly why nobody in recording music for humans needs such equipment.

But what if I want to record music for my dog? Will she feel jipped if she only hears the bass track, because I wasn't paying attention to her ear requirements?

[tech1]

We have customers outside of the music industry, too. For example, biologists recording bats needs to record audio up to almost 100 kHz, thus making recording at 192 kHz entirely necessary.

[Justin]

I think a big part of it is that some manufacturers have decided to put support for this in their "high end" products. So while you may never use the high sample rate, it makes sense to get a product that supports it because it has higher quality converters or preamps in it, or just an overall better quality.

Of course with HD-DVD and BLU-Ray formats both supporting up to 192k audio, people are going to think that they need to start recording in it to keep up with the Joneses. Even if it is out of the range most people can hear. I'm sure there will be audiophiles and home theater enthusiasts raving about it...regardless of whether they could really tell the difference in a blind listening test.

[dpd]

^^^ reminds me of when microprocessors went to 16 bit busses from 8 bit. Software engineers suddenly started requiring 16 bit converters instead of 8 bit, even though most of them had no clue if they really needed the resolution.

[DAS]

Try to find products that DON'T have those capabilities these days. We sell what they make.

Now, with all of this said...and I do believe that there is no theoretical validity to higher sampling rates...there are many people out there who do claim to hear a difference. I don't believe they are all faulty or making it up. There are practical implementation issues that can cause differences in different sampling rates. There shouldn't be, but there are.