Nov04 dbPowerAmp = 26% as fast as previous version on identical system

**Wayne** · August 04, 2005, 10:32 AM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

Are you converting the same file or ripping the same track from a CD? If it is a file what format is it? Are you running dMC with the same priority on both machines?

I don't think Spoon has ever sent emails informing registered users of upgrades. It has never been a problem for myself as I am a regular visitor to this forum keeping up to date with developments and occasionally providing advice to other users.

Wayne

**xoas** · August 04, 2005, 10:59 AM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

You may need to re-enable some of your advanced settings (ripping priority and rip to ram). I notice in this thread (http://forum.dbpoweramp.com/showthre...low+conversion) that these advanced settings do not automatically carry over when you convert to a different version.

You may also want to try the current dMC beta to see if that works any better for you.

Best wishes,
Bill

**Spoon** · August 04, 2005, 07:29 PM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

The only way you can check speed is not to use a CD but first create a wave file, also make sure no DSP effects are being used (such as volume normalize).

Try converting to wave files on both versions, I don't see why Blade (why not using lame?) is slowe between versions when blade has not been touched in years.

**GeorgeButel** · August 04, 2005, 09:25 PM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

I am ripping cda files with dMC Audio CD Input, and using lame, not blade as I mis-wrote. I am not using 2 different machines, but one machine with Roxio/Norton GoBack on it. I autoback the machine to the same configuration, then uninstall dB powerAMP and reinstall either version 9 or version 11 without having uninstalled the PowerPak. The program is running in normal priority. What caused me to keep using autoback and testing different configurations was that I installed the ape codec at first, knowing that it would slow the program down, but it was so much slower than realtime that I uninstalled it, then discovered that even without the ape codec, it was still running at 80% of realtime on the average. I have been using the program for years without any problem, in preference to Roxio, Steinberg's get it on CD, or Nero (Nero doesn't like directcd so I either have Nero or Roxio on a given configuration). If I uninstall v11 then reinstall v9, v9 runs normally, about 3.84 x realtime, so if something is changing that I am unaware of, it only occurs when going from v9 to v11, not the other way around. The ripping options are set at very high, constant bit rate, and the priority comes up as normal in both versions. I do not find the rip to ram setting anywhere in v9; the powerpak page gives only "use these settings for next conversion" and the temp path. The only option that I had to reset was that the bitrate had gone to variable from constant, but the quality was still set on very high and 320 and 44100. (Yes, I am braindead, but removeable storage is getting dirtcheap, about 50 cents a gb if you wait for the deals with the rebates,)

I remember seeing the rip to ram option, so I will install v11 again and check the speed, and if it still less than realtime, perhaps I should have uninstalled the powerPAK also, then reinstalled it after reinstalling the converter. What I will do next week is uninstall both dB and the powerpak, then create an autoback point, and then install v11 and the powerPAK, then try measuring the conversion of wav files and cda files, then autoback and install v9 and the powerPAK, and measure the same files.

I would try the new beta (since I have GoBack with autoback enabled, the machine is perfect for testing a beta; no matter what happens I can revert it), but when I click on the link, it only takes me to the beta forum.

**xoas** · August 04, 2005, 09:53 PM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

This link should hook you up with the latest dMC R11.5 Release Candidate2 beta:

http://www.dbpoweramp.com/beta/dMCr11.5rc2.exe

Best wishes,
Bill

**LtData** · August 04, 2005, 10:33 PM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

dMC r9 might have an earlier version of the PowerPack that is included in r11, but I don't know. Also, the dMC beta is very stable. I have been using the dMC r11.5 betas since Spoon first released them, as they are normally stable and include bugfixes.
The "Rip to RAM" setting is in the Audio CD Input window, click on the Rocket Icon.
Also, interestingly enough, CDs rip faster when you set the priority to "Below Normal."

**GeorgeButel** · August 05, 2005, 03:27 AM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

Thanks, gentlemen, for the advice. I will try the beta. I did not remember the rocketship icon in the audio cd input window because I had only gone there a few times, and I have had the program set the way I wanted it for a long time, so that setting might have been the problem. I will not have time to see until next week. The Powerpak is from a purchase in Jan of '03, so it and the installer v9 I have saved are from whatever was available then. Unfortunately, I often change the installer file names when I archive them, so I only have the date to go on.
Yours truly
George

**Spoon** · August 05, 2005, 07:50 PM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

It could be the lame encoder version has changed between versions (look in dMC configuration for the Lame versions).

**GeorgeButel** · August 09, 2005, 10:14 AM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

I uninstalled dMC and the powerpak, put in dMC v7, then went back to v9 with and without the old powerpak, then tried v11 again, then the beta, both on cda files and wav files. I only ran 2 cda files of a total of 3m14s and one wav file of 9m30s, so the numbers I got cannot be thought of as statistically valid, but they are telling me which ballpark I'm in, and the results are as before. v7 was 3.5x realtime for cda and 4.6x realtime for wav, rip method default MMC (measured with stopwatch). v9 was marginally faster, 3.8x realtime for cda and 4.8x realtime for wav, set to rip to ram and encode at same time, with no significant difference if the encode at the same time is not checked . v11, however, again was much worse: .75x realtime for cda, and 1.0x realtime for wav. There was no significant change with v11beta. The rip to ram default setting, auto, was indeed worse, .9 realtime for wav (did not test cda). Since the rip speed and the encoding speed are now shown, I can see that the problem lies in the rate of encoding, which is much worse than realtime. All tests were done at a constant bit rate of 320 kbps, quality very high, and with whichever version of Lame was included with the individual version. I can understand that a change in the encoder would change the rate of encoding, but this change seems to be too great to be the entire explanation. Maybe if the user could configure the program with only the encoders he or she actually wants, it would speed it up. Unless I can figure out something to do to speed the encoding up, I am going to have to revert to v9. I have varied everything I could think of varying.
I do notice a scripting problem with the beta--or maybe it's my windows--but the rocketship icon appears overlayed on the genre bar under the year bar, and shrinks to a sliver if I click on the genre bar. It has been this way on each installation. A separate annoyance that is part of all versions has been that once you have visited the freedb database, if you guess wrong about which of multiple choices to select, and get one that will require that each individual file have its name and/or tag edited, you cannot go back and make a different choice because either dMC will remember your choice (or is it freedb remembering via a cookie?) Another annoyance is that on a cd that is a compilation, you cannot edit the individual artist names in the dBpowerAMP window.
Thanks for the consideration.
George

**LtData** · August 09, 2005, 11:31 AM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

Regarding the compilation comment, with the latest beta I believe you can edit individual artist names. With r11, however, it was a bug that it changed the names for all of the tracks, I believe.

What are your system specs, out of curiosity?

Also, did you have "Rip & Encode at the same time" checked? Are you encoded files going to your HDD or an external drive?

**xoas** · August 09, 2005, 04:37 PM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

I do notice a scripting problem with the beta--or maybe it's my windows--but the rocketship icon appears overlayed on the genre bar under the year bar, and shrinks to a sliver if I click on the genre bar. It has been this way on each installation.

I do not have such a problem. I wonder if its a Windows issue, maybe a display setting issue or whether it could be sign of some corruption somewhere along the way. What is your display resolution?

I do wonder about your process here. I believe that Power Pack, mp3 decoding and the Lame mp3 encoder are all different between v. 9 and v. 11. I wonder whether there might be some registry settings that are not getting fully cleared or reset in the switch betwen versions (although such a problem might affect the speed of both versions). You might try uninstalling v. 9 (with the Lame codec), clearing the dMC registry (there is a sticky from Razgo in the dMC forum on Uninstalling/Reinstalling issues that describes this procedure if you are unfamiliar with it), rebooting the computer, and then installing dMC v 11.

I am also unsure as to what your OS (Windows or Linux version) setup is. If you have Windows 95 or 98 (which require some additional system files to work right), maybe the problem lies there.

Best wishes,
Bill

**Spoon** · August 09, 2005, 07:02 PM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

For testing try installing the Ogg codec, then you know it is the same encoder for both versions, Release 11.5 (the beta) should be slightly faster than R9 and R7.

>you cannot go back and make a different choice because either dMC will remember your choice (or is it freedb remembering via a cookie?)

Press the Freedb button.

>Another annoyance is that on a cd that is a compilation, you cannot edit the individual artist names in the dBpowerAMP window.

Select an track and type the artist name at the top, select the next track and type a new artist

**GeorgeButel** · August 10, 2005, 08:01 AM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

I have tried it with and without the encode at the same time. I normally rip to a removeable USB HD (I have to do so with GoBack because the virtual partition it uses to "protect" the main drive fills up too fast for it be of any use); however, for the experiment I was doing, I was ripping to the C drive. I have an old oem WinMe on an 866 PIII with 524MB ram; I restore the OS about once a year (I'm waiting for Longhorn, now Vista). The gui problem is probably a windows problem; I had a similar gui oddity afterward, ran Norton's WinDoctor, and it found 3 invalid dB xcom registry entries. I will run registry first aid and check out the uninstall/reinstall forum suggested by Bill, thanks. (The resolution is 1024x768.) I did run all tests with the identical background programs (GoBack, AVG and Blackice at startup, which was giving me 88% resources free initially, then adding only a network connection, i.e., I don't allow isp software on my system.) And thanks for the freedb tips; I should have done more looking for how to info instead of putting in the artist names on the tags and filenames manually all this time.

Sorry to have bothered you with this problem, because I could live without the newest version, but I guess this is one of those "known unknowns" that is driving me crazy. Please forgive my posts if there has been any lack of protocol but 99.9% of the problems I encounter get solved with the help of google, and most of my computer time is spent on doing things (no IMing or chatrooms). Also, I live too far from SBC's central office to get dsl, so I still use dial, which makes web seaching somewhat arduous.

I used to have rambooster running all the time, until I started using dMCv7. Every time the free ram fell below a certain number, rambooster would run, and there would be a millisecond or so "gap" on the ripped file, so if anyone ever complains about gaps in their files, look for the presence of such a utility.

Yours truly,

George

**GeorgeButel** · August 10, 2005, 07:27 PM

Re: Nov04 dbPowerAmp = 26% as fast as previous version on identical system

I will try doing the tests with the Ogg codec in the future for a true comparison. I did, however, try doing a PROPER uninstallation/installation--the way I do it for major programs (like Norton, Office, etc.), i.e., cleaning and compacting the registry, deleting all the temp files, and rebooting even when not instructed to do so. I am begging for Vista to come out because reboots will no longer be necessary. I haven't been doing "proper" installations for what I think of as "little" programs. After numerous reinstallations, I am puzzled why I sometimes get a reboot necessary type of message and sometimes not. By doing the proper uninstallation/installation, I have improved the beta's speed to 1.25x realtime, still slower than R9 with the earlier codec but 2.5x faster than before.
George

Nov04 dbPowerAmp = 26% as fast as previous version on identical system

Nov04 dbPowerAmp = 26% as fast as previous version on identical system

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment