Re: [Audacity-quality] Benchmark 16-bit
A free multi-track audio editor and recorder
Brought to you by:
aosiniao
From: Gale A. <ga...@au...> - 2014-04-09 06:47:45
|
> From Steve the Fiddle <ste...@gm...> > Tue, 8 Apr 2014 13:34:21 +0100 > Subject: [Audacity-quality] Benchmark 16-bit > > On 8 April 2014 08:13, Gale Andrews <ga...@au...> wrote: > > > >> From Steve the Fiddle <ste...@gm...> > >> Wed, 2 Apr 2014 13:22:16 +0100 > >> Subject: [Audacity-quality] Benchmark 16-bit > >> I think to tool was originally intended only for developers to use as > >> a rough guide. In practice I find there is very little difference > >> between the number of 16 bit tracks that can be played compared with > >> the number of 32 bit tracks. Other factors (such as whether sample > >> rate conversion is required) have a much bigger impact. > >> > >> Steve [...] > > Perhaps at least as a relative indication it is some use, if run > > under typical conditions that you would use Audacity for. > > > > I tried shutting all programs off except e-mail client and Firefox > > (20 tabs open). Default host, devices and sample rate (44100 Hz). > > > > * Win 7 x64 dual core 6 GB RAM 2.4 GHz: result of 68.7 tracks. > > If I play without invoking resampling I can play 50 3-minute > > tracks or 80 tracks but rarely 100. > > > > * Ubuntu 13.10 1 GB RAM, 1 GHz: result of 27.3 tracks. > > Without resampling I can play 15 3-minute tracks, or > > (half the time) 20 tracks or (occasionally) 30 tracks. > > > > * OS X 10.9.2 MacMini 4 GB RAM 2.3 GHz: result of 125.6 tracks. > > Without resampling I can play 100 3-minute tracks or 125 > > tracks (almost always) or (about half the time) 140 tracks. > > > > Apple wins. > > :-) > Just wondering how you are defining "can play". > On my (32 bit Debian) system I find that as more tracks are added > Audacity becomes progressively less responsive. There is a pause > between pressing Play and playback starting and the pause gradually > gets longer as the number of tracks increases. For example, my machine > is a Core2 Duo T5800 2GHz 3GB RAM. Benchmark indicates 179 tracks. > With 128 3 minute tracks there is a delay of about 1 second before > playback starts. With 256 tracks there is a delay of about 3 seconds. > 179 tracks seems like a reasonable figure for Audacity being usable, > but I don't see a clear "cutoff" point. "Can play" for me is defined as they will play for at least 30 seconds without audio breakups once they start playing, which I guess is the most important criterion for a user. Like you I find on all three machines, the more tracks there are, the longer delay before they start to play. However on my machines if the tracks have not started to play within two seconds (which I feel is a reasonable delay for a lot of tracks), then they won't play through without interruption. So my "half the time" means that on half the occasions the tracks will start within two seconds and play, and half the time they will take a bit longer to start then can't be relied on to play flawlessly. I agree there is not a clear cutoff, but if as sounds likely you get 2 seconds delay on 179 tracks and they play without interruption then the benchmark prediction seems on a par with my results. If your 256 tracks play flawlessly after a wait then perhaps the prediction is not so good, but I've not done a comparable test on Linux using my (better) 64-bit Windows machine. If the benchmark result was rounded to a range e.g. 51.3 became "approximately 45 to 55 tracks" and clarified that it assumes there is no resampling then I think you could regard this as of some value. One more test: * Windows 7 x86 1 GB RAM, 1 GHz: result of 40.5 tracks. Without resampling I can play 40 3-minute tracks, or 50 (half the time) or 60 tracks (occasionally, same definitions). I had Nautilus file manager running when I tested that machine on Ubuntu, so had Windows Explorer running for testing on Windows 7. Gale > >> On 2 April 2014 10:08, Peter Sampson <pet...@ya...> wrote: > >> > On reviewing the "Help Menu" page in the manual > >> > this thought struck me: > >> > > >> > Why does Benchmark work with 16-bit sample format when Audacity's > >> > default sample format is 32-bit float? > >> > > >> > It does use Audacity's 44.1kHz default sample rate > >> > > >> > Is this intentional, an oversight or a bug? > >> > > >> > Peter. > >> > > >> > Peter Sampson > >> > Tel: +44 (0)1625 524 780 > >> > Mob: +44 (0)7732 278 299 |