benchmarking / reports - laptops for Avenue

Just let it all out, buddy. You're among friends here.
Post Reply
User avatar
VJair
Posts: 609
Joined: Wed Jan 19, 2005 21:23
Location: Kent - UK

benchmarking / reports - laptops for Avenue

Post by VJair »

having just been through the process of getting myself a new laptop i know how daunting it can be trying to select a model, so i thought that maybe i could start a thread on it and list the models we have and the performance we get from them using a standard set of clips.

im going to make some test clips which i will make available for everyone to use so that we know the results are as fair as possible. i plan to make 3 versions ( 640x480, 1024x768, 1280x720 ) DXV is the format im going to make them in ( its the obvious choice ) but i have a quick question about the files themselves.

does a clip with less colours and less going on use less resources than a full colour clip with a lot of action? ie : should i make 2 different clips - 1 of more minimal graphics and 1 of a film sample ?

User avatar
VJair
Posts: 609
Joined: Wed Jan 19, 2005 21:23
Location: Kent - UK

Re: benchmarking / reports - laptops for Avenue

Post by VJair »

ok, so ive made some clips and run some initial tests to get an idea of how we could do this.

it seems that the contents of the file is irrelevant as i got the same frame rates with 20 layers of minimal graphics video ( text video ) as i did for 20 layers of video footage ( skate video ) - see attached images for screenshots of both ( this surprised me as i expected 2 colour imple video to use less resources than real video footage )

it also doesnt seem to make any difference if i use 1 video loaded 20 times or 20 copys of the same video - again i get the same frame rates for both

because of this i suggest that i just upload 1 clip in 3 sizes and that there will be no need to duplicate the clip for the test - you can just load the same clip multiple times.
text video
text video
text video.jpg (44.6 KiB) Viewed 3933 times
skate video
skate video
skate video.jpg (89.1 KiB) Viewed 3933 times

so, for the benchmarking i suggest we make three 20 layer composition's - one composition for each video size ( 640x480, 1024x768, 1280x720 ) we then load the relevant clip 20 times ( once on each layer ) and state what the frame rate is. we will also need to state computer type / model, cpu type / speed, graphics card type including amount of graphics memory, hard drive type and speed and the amount of system ram.

the aim will not be to boast who has the best computer, but to give a good view of what machines are capable of, which will hopefully help anyone looking to buy a machine to run Avenue :)

how does this sound? have i missed anything essential? if there are no objections or further suggestions, i will start a fresh thread for the results.

User avatar
gpvillamil
Posts: 550
Joined: Mon Apr 04, 2005 03:33
Location: San Francisco, California

Re: benchmarking / reports - laptops for Avenue

Post by gpvillamil »

That sounds good, but once you have more than 10 layers, the grid gets pushed down and you can no longer see the fps indicator. My screen is 1680x1050, and once I have 10 layers, I can no longer see the fps indicator.

I think maybe just 5 layers will be enough for comparative benchmarking.

User avatar
VJair
Posts: 609
Joined: Wed Jan 19, 2005 21:23
Location: Kent - UK

Re: benchmarking / reports - laptops for Avenue

Post by VJair »

odd, the fps indicator shows either on the preview monitor or on the output monitor if you collapse the preview one. im on a pc laptop, can someone check this is the same on a mac. ( maybe collapsing of the preview monitor should be specified in the testing spec? )

i did think of using a lower amount of layers, but my machine ( which isnt "high end" but is decent enough ) doesnt flinch until quite a few more are added, so i thought a higher number would be better to give a greater idea of the impact.

Post Reply