What is the difference between 44 and 48 khz?

[tominator]

Hi
I want to encode movies to WMV with 1308 kbit video and 192 kbit audio.

My problem is I can chose from 44 or 48 khz and I dont know what khz is and henceforth neither the difference when using one instead of the other.

Could someone please explain?

[setarip_old]

Without getting into the details of frequency/kiloHertz (kHz), suffice it to say, all other things being equal, the higher the frequency, the higher the quality of the audio...

[tominator]

Without getting into the details of frequency/kiloHertz (kHz), suffice it to say, all other things being equal, the higher the frequency, the higher the quality of the audio...

Okay. But wouldnt higher quality mean more bitrate needed for audio? How can I chose from either 44 or 48kHz without making the file larger?

[setarip_old]

All other things being equal, the file will be larger using 48kHz.

To compensate, either you or the program may choose to lower the audio sampling bitrate...

[ursamtl]

Hi
I want to encode movies to WMV with 1308 kbit video and 192 kbit audio.

My problem is I can chose from 44 or 48 khz and I dont know what khz is and henceforth neither the difference when using one instead of the other.

Could someone please explain?

The answer is really quite simple. If you're creating a DVD, you have to convert to 48kHz because the DVD standard calls for that. Use a high quality sample rate converter such as SSRC (free commandline program, just Google it) or r8brain (www.voxengo.com, look for the free version).

In contrast, CD audio and most MP3 players, etc. need the audio at 44.1kHz.

[tominator]

The answer is really quite simple. If you're creating a DVD, you have to convert to 48kHz because the DVD standard calls for that. Use a high quality sample rate converter such as SSRC (free commandline program, just Google it) or r8brain (www.voxengo.com, look for the free version).

In contrast, CD audio and most MP3 players, etc. need the audio at 44.1kHz.

Thanks a lot. You wouldnt happen to know about a program which can take an ac3-file and change the volume level to a specified db-level?

[ursamtl]

Well I believe Foobar2000 will play an AC3 file but as for setting to a specified dB level, I don't know. It will process Replaygain quite nicely. I don't know if that will work with n AC3 file. You'd probably have to decompress it, make the change then recompress it, something that would be bad because you'd be applying lossless compression twice!

[tominator]

The answer is really quite simple. If you're creating a DVD, you have to convert to 48kHz because the DVD standard calls for that. Use a high quality sample rate converter such as SSRC (free commandline program, just Google it) or r8brain (www.voxengo.com, look for the free version).

In contrast, CD audio and most MP3 players, etc. need the audio at 44.1kHz.

Sorry about all these questions. Just want to get everything straigt.

You say 44 kHz is standard for mp3s and 48 kHz for DVDs. My audio is 192 kbit/s (same as most mp3s). Would that mean I should be using 44 instead of 48?

Thanks for all the help. Its great.

[mitsubishi]

You're confusing sample rate with bitrate there.

The sample rate is a bit like the frame rate for audio, 44khz is 44 thousand samples per second, ie the analog sound wave is sampled every 1/44000 seconds so the wave can be reconstructed from these points. The bitrate is the amount of data allowed to be used to store the compressed samples.

[Lokean]

You wouldnt happen to know about a program which can take an ac3-file and change the volume level to a specified db-level?

Try BeLight

[tominator]

Try BeLight

Thanks. Tried it. But under AC3 I can chose from endians Motorola and Intel. What is the difference?

[miztadux]

My problem is I can chose from 44 or 48 khz and I dont know what khz
44/48 hkz is the sample rate value.
It represents how many values are stored per seconds.
It's somewhat analogous to the frame rate of a movie: 24 fps means 24 frames per seconds; 44khz means 44000 samples per second.
It's a really important factor for the quality of an audio capture as it limits the highest frequency that can be reproduced (Shannon's theorem says it's half the sample rate: 44khs samplerate => max 22khz audio)
And, like a movie frame rate, changing this value is a complex operation (as you need to extrapolate new sample values) and thus can degrade quality.

Long story short, I think the standard practice for high quality rips is to keep it the same as the source: like this you avoid a noisy operation.
Moreover, when downsampling (ie decreasing samplerate) you first need to low-pass: to remove any frequency higher than the new maximum (your 48khz source may have 24kz frequencies which can't be reproduced correctly in a 44.1khz sample)

Anyway, if you're planing on using 192kbps wmv stereo, you'll have enough bitrate to encode 48khz.

Thanks. Tried it. But under AC3 I can chose from endians Motorola and Intel. What is the difference?
BeSweet will allow you to re-encode your AC3, meaning a full cycle of AC3 ---decoding---> wav ---encoding----> AC3, if you just want to change the gain I don't think it's a good solution.
And to encode AC3 in besweet, you'd better use an updated GUI like BeLight which supports the higher-quality Aften (http://forum.doom9.org/showthread.php?t=113074) AC3 encoder.

[tominator]

Long story short, I think the standard practice for high quality rips is to keep it the same as the source: like this you avoid a noisy operation.
Moreover, when downsampling (ie decreasing samplerate) you first need to low-pass: to remove any frequency higher than the new maximum (your 48khz source may have 24kz frequencies which can't be reproduced correctly in a 44.1khz sample)

Anyway, if you're planing on using 192kbps wmv stereo, you'll have enough bitrate to encode 48khz.


All my sources are 48kHz (all files come from retail DVDs) so I don't think Ill have to down sample anything.. or am I wrong?

My solution to my whole gain problem at the moment is this:
Im adding a volume filter which makes the audio louder and I also add a filter which normalizes the peak to -3db.

Do you think thats a good solution?

Thanks.

[Mug Funky]

foobar2k can put replaygain info into ac3 files. only problem is your authoring program will complain that the last frame is bad (because that's the tag...). shouldn't be a problem.

but please bear in mind that unless the DVDs you're ripping got n00bs to master, encode and author them, they should be at a correct, calibrated volume that it's not wise to mess with. they don't need to go louder - the playback device does :sly: you'll find the replaygain value will be around +6dB for most of the ac3s you scan, and that's pretty much right. stereo mixes might come out different though - they tend to be less dynamic and hover at an average of -20dB, with a hard limit at -10dB. this should correspond with 0 and +10 on your stereo's VU meters (if they're calibrated...).

if you have the option to decode your ac3 with dynamic range compression enabled, that should be enough for WMA encoding. bear in mind it doesn't have to be as loud as a pop CD.

[tominator]

but please bear in mind that unless the DVDs you're ripping got n00bs to master, encode and author them, they should be at a correct, calibrated volume that it's not wise to mess with. they don't need to go louder - the playback device does :sly: you'll find the replaygain value will be around +6dB for most of the ac3s you scan, and that's pretty much right. stereo mixes might come out different though - they tend to be less dynamic and hover at an average of -20dB, with a hard limit at -10dB. this should correspond with 0 and +10 on your stereo's VU meters (if they're calibrated...).

if you have the option to decode your ac3 with dynamic range compression enabled, that should be enough for WMA encoding. bear in mind it doesn't have to be as loud as a pop CD.

When watching a DVD the TV or the DVD-player will amplify the sound. My files are to be watched on a computer and when using headphones the sound is very low even if windows vol is max as well as the vol on the media player.

Therefore I boost the vol just a bit and normalize peaks to make sure it doesnt get too loud in some places.

Am I doing this wrong? Should I just let the volume be low?

If yes, could I give my users a program which will boost the volume for them?

[tebasuna51]

To backup ac3 (from DVD movies) to wma stereo 192 Kb/s to be played without the original dynamic range volume (don't mistake with 'quiet') I recommend:

- Preserve the original samplerate. Only if the output format force to a explicit samplerate (not here) must be changed.

- Decode with DRC normal (Azid-BeLight, NicAc3Source-BeHappy, ...). Avoid other type of boost if possible.

- Apply the downmix 5.1 -> 2 if needed (dplII if you want preserve any surround info, well supported at 192 Kb/s).

- Normalize at 90% (-1 dB), is enough to avoid overflows.

- And encode to WMA.

[miztadux]

All my sources are 48kHz (all files come from retail DVDs) so I don't think Ill have to down sample anything.. or am I wrong?

That's right, you don't have to downsample.
As tebasuna51 said, the only reason to downsample is to comply with some standard (for example [s]vcd is @ 44khz), but for a PC rip keep the samplerate of the source

[Mug Funky]

@ tominator:

that's fair enough to boost it if there's not enough juice in the headphone amp... it's a little odd because usually with the volume maxed on my machines it'll be just right, and with DRC it'll actually need to be turned down to halfway. my home sound chip is an AC97 and my work card is an audigy2 ZS (the sweet one with the breakout box :)). both give the same volume on large headhpones, but if it's too quiet on your machine you should still have about 10dB of headroom for boosting, so long as there's some kind of limiter going as well (competent DRC should handle this)